Header

Objective Reviews & Commentary - An Engineer's Perspective

February 1, 2012

Feb 1 Update

BLOGGER PROBLEMS: Just a quick update. I've been away from the blog for a couple of days only to find Google Blogger is having problems displaying more than 200 comments per article. Sometimes when you click “Load More” at the bottom of the comments it just gets stuck on “Loading…” and never refreshes the page with the latest comments. I’m not sure yet what the problem is, but if you can’t see comments posted in the last few days to articles with over 200 comments, that’s probably why. If I can’t find another fix, or Google doesn’t fix it, I need to find the time to go back and “clean up” the comments to keep the total number under 200 per article.

ODA/ODAC UPDATE: Many want to know when I’ll have the next ODA/ODAC article done. I’m working on it as time allows but that hasn’t been very often lately. I’ve been severely overloaded with other obligations, many of them unexpected, so progress has been much slower than I would like. Hopefully everyone can understand this not-for-profit blog, and related projects like the ODA/ODAC, are not my highest priority in life. I do plan on completing both the ODA and especially the ODAC. It’s likely the ODAC will be available first.

odac plus o2ODAC + O2 = USB DESKTOP AUDIO JOY: The current ODAC revision is designed to fit inside the O2 if you remove the batteries. This works even with the slim B2-080 standard enclosure. It’s not a plug-and-play modification, but anyone with a soldering iron and some DIY skills shouldn’t have any problem. The O2 will need to be powered from the AC wall adapter and the ODAC is self powered from USB. It will require a rear panel modification for the mini USB jack. Hopefully someone can design (perhaps with some key dimensions from me) a Front Panel Express CAD file for the rear panel. The ultra-high quality photo at right--the best image I could get with a flagship Nikon D4 DSLR ;)--clearly shows the ODAC PCB laying on top of the O2 PCB where the batteries usually hang out. Many are using their O2 for desktop duty so hopefully this might be a worthwhile option for some of them.

NEW SCOPE: Adding to my arsenal of test equipment is a new heavily optioned Agilent oscilloscope with one option that’s very applicable to this blog and the development of the ODAC—it has I2S audio decoding and triggering. So with the new scope I’ll be able to very precisely measure things like DAC chip latency and explore bit accuracy and other parameters in more detail. I’m not aware of anyone beyond a few of the chip makers publishing such data. The new scope can correlate digital audio bitstream data against the analog DAC output with nanosecond accuracy. Stay tuned for more.

questionmarkA NOTE ON COMMENTS & QUESTIONS: Increasingly I’m receiving a lot of public and private comments that boil down to something like: “Would the UltraDAC XL or the UberDac 4000 be a better match for my TurboAmp and Mk II WonderPhones?” While I’ve tried to answer most such questions, the volume is becoming overwhelming. Often the questions have little to do with the topic of the article they’re posted under or even anything I’ve reviewed on the entire blog. Paid advice columnists typically only reply to a few such comments per week or per month but I’m getting several per day. Regrettably, I just don’t have the time to answer them all. I will still try to answer questions directly related to a specific article and those with broad appeal. So please don’t take offense if you don’t receive a reply to a question. I wish I could answer everyone but it’s turning into nearly a full time job and I have to pay the bills somehow.

186 comments:

  1. Looks great, NwAvGuy! We'll do a matching ODAC rear-panel for the O2 as soon as the design is ready.

    --JDS Labs

    ReplyDelete
    Replies
    1. Anyway you can make this easier on us non-handy types will be welcomed!

      Delete
    2. Awesome! I'll be buying one, right away.

      Delete
  2. Thank you for the update!

    I'm one of those that waiting for the o2 with a DAC!

    Will there be an updated o2 scheme that will fit the dac perfect or could I just buy the o2 now and add the dac later?

    ReplyDelete
    Replies
    1. I'm not planning on updating the O2 PCB or packaging because that's what the ODA is for. Once you take the batteries out of the O2 it's not longer a portable amp. So for those who don't already have an O2, they're better off waiting for the ODA. The main idea with keeping the ODAC small enough to fit in the existing O2 enclosure is to let people retrofit their existing O2 being used for desktop use to become a USB headphone DAC if they want.

      So, if you want a portable battery powered amp, by all means get the existing O2. If you want mainly a desktop amp, I would suggest waiting for the ODA.

      Delete
    2. Roughly, how long will the wait for the ODA be?

      Delete
  3. Wow: was not expecting this. Complete signal chain transparency up to your headphones never came in a smaller box!

    ReplyDelete
    Replies
    1. Me too, not expecting it will be compatible with the O2..

      Do let me know if you need help on the back plates.

      Delete
  4. Looks awesome! BTW, great that you got a D4, I'm sure there are many other professional photographers that haven't got theirs yet!

    ReplyDelete
  5. Ah, every new post gets me giddy with anticipation. I want this silly upgrade bug to end--I just bought the HD650 for a decent price and will buy the ODAC as soon as it's available.

    I just need to get a black case for it and I'm good, heh.

    ReplyDelete
  6. Woah! I was second-guessing my impatience with ODA, but I guess this will make the theoretical update easy :) Thanks NwAvGuy! This is amazing. While the majority of the amp/DAC industry is going for heavy multiple-specialized-units space-wasters, the technical awesomeness per m³ is going off the charts with the O2 :)

    ReplyDelete
  7. No need to push yourself. I've just discovered this goldmine and from all I've seen your articles are definitely worth waiting for.

    That said I'm personally really looking forward to seeing how the Clip Zip holds up to the Clip+. There are some small board changes but in basic tests it measures practically the same. It will be interesting to see if that stays true when scrutinized.


    While on the topic the Fuze+ does indeed seem to have met with a worse fate. Even in RMAA some less welcome changes become apparent:
    http://www.hifi-forum.de/viewthread-125-6615.html

    Low frequency response seems fine even under load but in higher frequencies it is positively wobbly. Any idea what kind of design flaw would manifest itself this way?

    ReplyDelete
    Replies
    1. High frequency response problems usually indicate poor DAC filtering either in the digital domain, analog domain, or both. In a purely analog headphone amp it can also indicate poor high frequency stability compensation with a roll off indicating too much compensation and a peak or ripple indicating too little.

      I haven't tested the Fuze+ but it's likely suffering from one or more of the above. Battery powered devices are often optimized for low power consumption over audio performance which makes the Clip+ all the more amazing. With the Fuze+ Sandisk probably either was trying to save power and/or money causing them to switch to a different chip from the original Fuze.

      Delete
  8. The new scope sounds pretty cool. You're probably equipped to do just about anything related to audio electronics now...

    ReplyDelete
  9. What would be the benefit of waiting for ODAC? Rear inputs, internal power supply, more
    power, better quality op amps? If the DAC fits into the O2 enclosure, why reinvent the wheel?

    ReplyDelete
    Replies
    1. See the ODA/ODAC article. The ODA has several benefits for desktop use over the O2.

      Delete
    2. can you confirm if the oda+c includes pre-out? i would really like to use the odac as source for my home setup.

      thank you for all the time and patience you are putting into this projects.

      Delete
    3. I can't 100% confirm there will be pre-out, but that's the current plan.

      Delete
  10. Just can't wait... What would be the price of the board with the parts?

    ReplyDelete
    Replies
    1. The ODA raw board will be similar in price to the O2 with batteries and the AC wall transformer. The enclosure and panels, however, will be more expensive.

      The ODAC board will be sold assembled for under $100. See the ODA/ODAC article and comments for more.

      Delete
  11. This is so exciting NwAvGuy! I can't believe you made it so that I can just drop the ODAC into my current O2.. that is SO awesome. I can't wait!

    How long do you think it will be until the ODAC board is released? I'm literally on the edge of my seat right now.

    ReplyDelete
  12. That's good news.
    Which Dac IC are you considering?

    ReplyDelete
  13. Wow!! If it will be as simple as modding my existing O2, then sign me up!!

    ReplyDelete
  14. Thanks for the update. Many people are waiting patiently for your design for ODA and ODAC :D

    ReplyDelete
  15. This silly thing called 'real life'..you need to get rid of it ;)

    but anyway, the O2 with casing is small enough but to think the ODAC will fit inside that..WOW!

    ReplyDelete
  16. That is a great idea! fit the ODAC into the O2, this would make me so happy!!!

    Don't stress about this blog, I am honestly amazed at your output... Myself - I can hardly find time to read your articles, can't imagine the work that has been put into writing them!

    Anyway keep up the good work, we will patiently await for things to come out when they are ready.

    ReplyDelete
  17. I assume that installing the ODAC in the O2 renders the line input jack unusable, turning the O2/ODAC combo into a dedicated USB headphone DAC. Correct?

    How about with the ODA/ODAC combo? Will there an ability to switch between line input and DAC?

    If not, no problem: I can go with a two-box approach, keeping the ODAC in its own enclosure, so that I can us the ODA with my line level sources.

    ReplyDelete
    Replies
    1. You are correct about the O2 unless a DIYer wants to add their own switch which they could certainly do and probably even squeeze in somewhere with the B2-080 small enclosure. The ODA should be more flexible.

      Delete
  18. Thanks for the offer. A few others have made similar suggestions but this blog is sort of philanthropy for me. The best thing everyone can do is help get the word out that you can't believe much of what you read on commercial sites like Head-Fi that have a strong interest in promoting many audio myths for their own profit.

    ReplyDelete
  19. Hi NwAvGuy,
    Have you though about setting up your own forum? I realize that there are plenty of audio forums, but except for some pro audio gear ones, too many audiophile forums, if not all, are filled with snakeoil voodoo and like you said, heavily commercialized (Head-Fi even forbids ABX discussions).
    Donations can be limited if you want, just to cover server/hosting fees. Plus, it will help keep the blog articles on topic while everything else, like offtopic requests, can go in the forum for everyone else to contribute/discuss.

    ReplyDelete
  20. Thanks to Google, this blog is hosted for free. I have several friends who hold significant positions within Google and have to admire how much Google offers for free. This blog has similar traffic stats with InnerFidelity yet is entirely without advertising and entirely free. I think that's pretty amazing in this day and age of corporate profits dominating over most everything else.

    As for hosting my own forum, I've thought about it, but it's an even bigger time commitment to keep up with a forum than this blog. I would have to find some trusted moderators willing to donate their time who could babysit a forum 7/24 as that's pretty much what's required.

    The problem with popular forums is the "for profit" guys will try, in all sorts of devious and creative ways, to leverage it for their own financial gain. It's challenging to try and moderate the commercial interests without excessively censoring the non-commercial members as the former often poses as the latter.

    ReplyDelete
  21. I have no idea how you could possibly answer all questions without starting to charge a consulting rate that would include testing each person for their own particular preferences.

    I, for one, don't blame you for not answering every question, and I don't know how you could possibly keep up.

    Then again, when somebody doesn't like your answer, we all know what we'll hear then.

    ReplyDelete
  22. NwAvGuy, looks terrific so far, I can hardly wait to implement it! I have my O2 boards populated now. One board is for battery use and the other two for AC. I have two questions for you. First, what would be the harm in wiring the analog out of the ODAC to the P2 input position so you can use either input, so long as both are NOT simultaneously used? Second, does the ODAC have sufficient output voltage to drive dual O2 boards? Reason, I found a Hammond case large enough to hold 2 O2 boards. I figure an 18VAC/1A (Jameco ADU180100) ps should be fine for dual boards. My plan is to mount the pots, out's & LED to a nicer front panel and move the input, gain switches and ac to the rear panel. Cheers!

    ReplyDelete
    Replies
    1. If you cut the shorting traces on the switch contacts for P2, and wire the ODAC to the switch pins (pins 3 and 4--see the schematic and PCB layouts in the PDF), then you can use the line input. But you'll have to unplug your source to listen to USB. You can't just wire the two in parallel because your line level source will "fight" with the output of the ODAC even if the ODAC is not powered up.

      I haven't tested it, but the ODAC should drive the approximately 5K load of dual O2 boards. If you want to be safe, replace the 10K input resistors of each board (R14 & R20) with approximately 20K resistors to provide a similar 10K total load.

      Yes that transformer should be fine although the regulators will run fairly warm. If you can get the WAU16-1000 that would be an even better choice (16V 1A).

      You may also want to wait for the ODA as that would save you some work and offers other benefits. You could use one of your O2 boards as the second output and power it from the ODA.

      Delete
    2. Do you mean the P1 inputs? or do I have an older version of the PDF? P2 looks like the output? I'm a little confused..

      Delete
    3. Actually neither is correct. I was just using "P2" as that was in the post I was replying to. The correct designation is J2==the 3.5mm input jack.

      Delete
    4. Yes, my bad, I used the wrong designation. I hurriedly looked at the HO and not the INPUT. Sorry for any confusion. :)

      Delete
  23. Greetings, Mr. NwAvGuy!

    Please don't think I glossed over your comments above decrying posts irrelevant to the subjects at hand. I simply cannot resist, nor do I wish to resist, congratulating and thanking you for designing the O2. Today, I received mine from JDS Labs. I realize much critical listening is ahead in order to familiarize myself with the amp's qualities. However, even after only an hour or so listening to highly compressed pop files using my 4G Touch as a source (and a FiiO L3 LOD to connect accompanied by ER-4Ps), I am staggered. The amp is literally absolutely silent in terms of S/N ratio, and the sense of "cruising" through music of extended dynamic range and "openness" is entirely new to me.

    I hope the world takes notice of your accomplishment. It and you represent quality on a planet where little of such is to be found.

    Jonathan S.

    ReplyDelete
  24. Hi NwAvGuy,

    I'm from Germany and got interested in HiFi since my father bought a Apogee Duet 2. And now the funny thing: He also bought some cheap KSC 75 headphones and we both compared them with a AKG K701. There wasn't a big difference at all but the price. That was the point when I lost my trust towards the market. I bought those KSC 75 too and am now searching for good DAC/AMP solution. I already read a lot about Fiio products, about the Behringer UCA202 and thought of maybe combining the DAC of the UCA202 with a different AMP as the headphone out is so bad. However, then I found your blog and was fascinated. You cleared so many things up and I saw your O2. Sadly (only for me actually :D), it's a DIY-project, so I'm not able to "have" them as I can't build them.

    So my question is, will there be a store or something that you recommend that will sell assembled ODACs in the future?

    ReplyDelete
    Replies
    1. JDS Labs in the USA has already indicated they plan to supply the ODA assembled, the ODAC boards will only be available assembled (and it's likely at least one vendor will sell a complete ODA/DOAC), and there may be other sources in Europe as well such as Epiphany in the UK.

      Delete
  25. I was wondering if you had time to write an article reviewing Denon AH-D2000s.

    I am currently on Sennheiser HD 555s with foam mod, but I need closed headphones. I realize "closed" doesn't mean literally "closed" but the Sennheisers can practically be used as speakers if I put them on my desk. I'm moving into a dorm so I need to respect roommates.

    I know you own AH-D2000s and I would really appreciate any thoughts you had on it, as well as its successor the AH-D5000s.

    I know little to nothing about amps and dacs, but from what I gather reading your posts here, the D2000s don't really require amps or dacs anyway. Anyway, this is the sort of stuff I'm curious about.

    To give some background info I mainly listen to jazz and hip hop and my library is primarily flacs. I don't know if that helps but I might as well put it.

    ReplyDelete
  26. I was wondering if you had time to write an article reviewing Denon AH-D2000s.

    I am currently on Sennheiser HD 555s with foam mod, but I need closed headphones. I realize "closed" doesn't mean literally "closed" but the Sennheisers can practically be used as speakers if I put them on my desk. I'm moving into a dorm so I need to respect roommates.

    I know you own AH-D2000s and I would really appreciate any thoughts you had on it, as well as its successor the AH-D5000s.

    I know little to nothing about amps and dacs, but from what I gather reading your posts here, the D2000s don't really require amps or dacs anyway. Anyway, this is the sort of stuff I'm curious about.

    To give some background info I mainly listen to jazz and hip hop and my library is primarily flacs. I don't know if that helps but I might as well put it.

    ReplyDelete
    Replies
    1. I'm not nwavguy, but I have a lot of headphone experience.

      Firstly, don't get rid of your Senns, there will be plenty of times where it is appropriate to use open headphones. A lot of people aren't bothered by them at all. I used my Grados all the time.. including at night.

      Secondly, the D2000 is nice, a bit heavy on the bass, but it doesn't get out of hand and sounds generally great. It doesn't need an amp, by any means and should play well with your hip hop.. I think you might prefer your Senns for jazz, but the denons will do s respectful job.

      Delete
    2. If you dont want to buy an amp do not get the D2000. Low impedance aside, they need a lot of current. People underamp the denons all of the time and wonder why they don't sound good.

      Delete
    3. @Anon, I have to disagree. The D2000 is quite sensitive and does not need a lot of current. What it does need, and likely the reason for people thinking it needs a lot of current, is a very low output imepedance from the source. Many sources, including iPods and a lot of PC's, have an output impedance that's way too high for the Denons. So people buy a beefy amp, they sound better, and they think it's because the Denon's need more power, but that's not it.

      The lowly Sansa Clip+, for $29, drives the D2000 wonderfully well because it has a very low output impedance. See my impedance articles back in Feb of last year.

      Delete
  27. Closed does mean closed. Just use your Sennheisers and disrespect your roomate, that's what I do. :P

    ReplyDelete
  28. Hi NwAVGuy,
    Remember it's a marathon, not a race. The last thing anyone wants is for for you to get burned out and give up on the blog because you're trying to do too much. I think most here feel the blog is a great resource for information and that the updates are worth the wait.

    Regarding compensation I respect your stance, however considering all you do, and the time commitment, I don't think anyone would be bothered much if you used something like Google Adsense; no sponsor bias, as you don't control the advertisers, and no one is obligated to click on anything for you to get something out of it; give it some thought.

    I also think pre-outs would be awesome on the ODA :)

    ReplyDelete
    Replies
    1. I will second the vote for preamp outputs on the ODA and a switch to mute them (If I want to mute the headphones, I can always unplug them, but I'd not be opposed to a switch for them, also).

      Again, thanks for everything. I continue to enjoy my O2 and have convinced a couple of coworkers to get them.

      Delete
  29. Couple questions.

    Will the ODA have more gain options than the O2?

    Also, it appears that my 600 ohm DT880s need a gain of 3.1. How can I tell what would be the best gain setting to allow the most headroom without clipping? I found a calculator online that allows me to put in my headphone and source specs to determine the ideal gain for listening at 115 dB. Is there something similar to determine whether the amp will clip?

    ReplyDelete
    Replies
    1. Please see my All About Gain and More Power articles. If you're talking about someone else's amp (not the O2 or ODA) it can be difficult to figure out where it will clip as many provide very weak or "optimistic" specs.

      The DT880-600 needs over 5 Vrms to play reasonably loudly with all types of music. That's 14 volts peak-to-peak a most popular portable amps, and even many home sources, cannot manage that much output.

      The O2 and ODA, however, both output over 5 Vrms and work fine with the DT880-600. How much gain you need comes down to the source you're using. Again, see the two articles mentioned above.

      Delete
    2. Oh, I am referring to your amp. Sorry for not being more clear. My source outputs 2.16 Vrms (6.7 dBV), so the gain needed is actually 3.2.

      I just want to have everything straight when about clipping/headroom and compatibility with any future headphone or source upgrades. I mean, I wouldn't have a problem opening the amp up and swapping the resistors if I really had to, but I'd rather not.

      That's why I'm curious about the gain structure within the ODA. If it's better than what's in the O2, I'd rather just wait for it to come out than buy an O2 right now.

      Delete
    3. As I've already said in the ODA article/comments, and elsewhere, the ODA will have more flexible gain options than the O2. I've already documented input clipping vs output clipping of the O2 and there has been lots of other discussion on the topic.

      The O2, and especially ODA, are more flexible and able to drive a wider variety of headphones than most headphone amps on the market at anywhere close to their price.

      Delete
    4. "Oh, I am referring to your amp. Sorry for not being more clear. My source outputs 2.16 Vrms (6.7 dBV), so the gain needed is actually 3.2."

      A gain of 3.2 would not clip with a 0 dBFS sine wave, but it would be close to the clipping level. By the way, what source are you using (its output impedance can slightly reduce the voltage) ? The 2.16 Vrms value looks familiar from the Stereophile Xonar STX review, but other DACs may have the same output level.

      Delete
  30. What is your opinion of folks requesting and setting the gain of the o2 at one and three rather than the specd design? Does this impact sound quality in a negative way? I ask because most of the headphones I am considering are either IEMs, or very easy to drive cans (HM5).

    ReplyDelete
    Replies
    1. 1X and 3X (or 2.5X) is fine if it suits your needs. As I've said many times, you want the lowest gain that gets the job done.

      Delete
  31. I just have to say, nice B&W photo there. Reminds me much of an SEM.

    ReplyDelete
  32. @NwAvGuy, there is one feature I'd like to request on future headphone amp designs: a crossfeed feature. The basic idea is simple:

    When you are listening to actual sounds in the environment around you, you hear the sounds with both ears. When you play back a stereo recording with speakers, you also hear the sounds with both ears. But when you play back a stereo recording into headphones, you hear just the left track in the left ear, and likewise just the right track in the right ear. (Unless it's a binaural recording. http://en.wikipedia.org/wiki/Binaural_recording)

    The crossfeed would work by delaying the left channel and gaining it down, and mixing it into the signal sent to the right ear; and likewise delaying and gaining down the right channel and mixing it into the signal sent to the left ear. The delay should be roughly the delay caused by the speed of sound for the width of an average human head.

    If you want to consult an expert on what the delay and gain should be, @AudioSkeptic would be a good one. But I think a delay of 0.3 milliseconds or so is about right.

    As for the gain, you can use a negative or positive gain of about 0.3 for the cross channel, and 0.8 for the main channel. Like so:

    L output == 0.8 * L + 0.3 * (R delayed by 0.3 msec)
    R output == 0.8 * R + 0.3 * (L delayed by 0.3 msec)

    Well, those numbers assume a limiter or something to keep you from hard clipping with any possible input signal. If you didn't want to have a limiter you might need to use 0.2 for the crossfeed gain.

    This is perhaps not strictly essential, because some music players have a built-in crossfeed feature that can be enabled. But it would be great to have it in the actual hardware so that any source can have crossfeed applied. (I listen to a lot of music from Internet streaming, and the streaming player doesn't offer a crossfeed.)

    Ideally there should be a switch to enable/disable this feature, because if you ever want to listen to a binaural recording you don't want to add the artificial crossfeed on top of the naturally recorded crossfeed.

    I live and work in Kirkland, WA and I could give you a demo of crossfeed if you like. And I would be very happy to answer any questions.

    This is of course a request and not a demand! I don't know if this would be difficult or easy for you. Easiest would perhaps be if you could do it in the digital domain on the DAC card, but I don't know if you will have custom firmware on that card. If this would be difficult, then never mind.

    Thanks for your blog and for designing the O2.

    ReplyDelete
    Replies
    1. You wrote: "The crossfeed would work by delaying the left channel and gaining it down, and mixing it into the signal sent to the right ear; and likewise delaying and gaining down the right channel and mixing it into the signal sent to the left ear. The delay should be roughly the delay caused by the speed of sound for the width of an average human head."

      Sound does not travel through the human head from one ear to the other (in mine, there's something in the way). Nor do listeners typically have half of the band on one side of their head and the other half of the band on the other side of their head. More typically, the band is in front of the listeners, so the delay for off-center instruments is described by a triangle with the points being the sound source and each of the listener's ears.

      Unless you are listening to early Beatles stereo recordings, most multi-channel studio recordings have been mixed with the instruments and vocalists panned into their desired location, not 100% on the left or 100% on the right. That means that the delay would not be a constant. For a performer directly in front of the listener, the sound would arrive simultaneously at both ears. One that is a few degrees off of center would have only the tiniest difference. One way off to the side would have the greatest delay. So, unless you had access to the multi-track recording, where you could individualize delays based on panning of each track, you couldn't properly delay the sound (no such thing as a one-size-fits-all).

      You are also ignoring the effect that the pinna has on the frequency spectrum based on it's angle relative to the sound source. A patient with 100% deafness in one hear can still get directional cues from that. Until that can be modeled real-time, you can't expect headphones to truly sound "natural." In fact, cross-feeding a signal which has not been frequency-shaped to mimic the effect of the pinna might be even less natural sounding.

      Delete
    2. You need to compare it to speakers, not any "original" that may or may not even exist. The vast majority of 2 channel audio content is mixed to be played over stereo speakers and not headphones. Without electrical crosstalk between the channels headphones can never achieve the acoustic crosstalk that you naturally get from a pair of speakers.

      Its obviously not as good a binaural recording, a proper DSP, or something that clones a whole listening room like the Smyth Realiser but a simple Linkwitz crossfeed is so much better than nothing.

      Delete
    3. @Fred M., in the first place I urged NwAvGuy to talk to a real expert such as AudioSkeptic. I am not a real expert (nor do I play one on TV).

      That said, the simple crossfeed that I described is one that I have implemented in software, and it sounds good to me. I have tried it with both a positive gain on the crossfeed and a negative gain. The negative gain seems to make the soundstage wider, but in subjective listening I preferred the positive gain; but I haven't tested scientifically so take this with about a thousand grams of salt. But I like it, and the only reason I don't use it all the time for my headphone listening is that my software only plays wave files.

      I am aware that it is possible to do all sorts of processing to make this effect more complicated. I even know how to write the software to do the heavy lifting.

      I suggest that the ideal crossfeed should involve carefully measuring the HRTF of the user's head, so that we can better simulate what would be going on in the real world if we were listening with speakers. In the absence of measuring the user's head, you would have to try for a universal HRTF, and I would worry about whether it would sound equally good for everyone.

      On the other hand, you can buy products from Headroom that have some sort of crossfeed feature, and they view it as a major selling point; I submit to you that whatever they are doing is likely to be of similar simplicity to what I proposed. (Their least expensive amplifiers are probably doing all the processing in the analog domain, rather than running the input through a ADC -> DSP -> DAC chain. Actually I would bet that all of their amps do all the crossfeed processing in analog.)

      On the gripping hand, I would prefer the ODAC have a simple effect similar to the one I suggested, rather than no crossfeed effect at all.

      Delete
    4. It seems to me that folks are missing the point here. First and foremost, simple crossfeed reduces the abnormal sensation one gets by hearing something in only one ear. A delay of .1 to .2 ms seems to be enough for most (but there is an argument for making the delay variable and person-dependent), based on people's actual response to crossfeeds. As this is not talking about a full Atal-Schroeder cross-cancellation, I don't think HRTF's are as germane as you suggest.

      This is not a component in a full virtualization system, it is a system to substantially reduce discomfort from mixes that have too much left/right isolation.

      Consider: When you play back your theoretical panned recording in two speakers, you DO get a 'crossfeed' that includes HRTF's from the two speakers from the basic direct-signal acoustics, so your objection to center-panned and non-pingpong stereo panning through a crossfeed is simply misguided. To this day, most recordings are still intended for loudspeakers (never mind how most are consumed), and still presume that the interaural mixing including HRTF's is going to happen. In fact, at low frequencies, the poor HRTF approximation of a panpot in speakers actually provides something near-ish to a real interaural time difference, albiet an effect that fails terribly at frequencies above 2kHz.

      So, alternatively, I could argue that a simple Schroeder crossfeed with no frequency shaping is a zeroth order approximation to the opposite-side HRTF simulating the loudspeaker setup, not the original mix. Since the mix was intended for loudspeakers, we have to assume that's the right thing to do.

      Adding a bit of frequency selectivity (and having different signs at low and high frequencies in the crossmix as appropriate via non-minimum-phase filters) is a better simulation of the actual system you're listening to in speakers, of course.

      If you notice what steveha has suggested here, you see the number .3 millisecond, not .9 millisecond, which would be more in line with a side-position HRTF delay (that being, of course, a mix of at least 3 delays (around the front, over the top, around the back) along with the appropriate diffusion effects and pinna shadowing.

      Finally, of course, you've leaped to a full virtualization system, which is not what was suggested, rather, what was suggested was a zeroth-order simulation of the HRTF effects of two speakers at ISO left and right listening positions.

      It's not perfect, but some people very clearly prefer this kind of processing in headphones, and it's very, very simple.

      Delete
    5. ^ This.

      Thank you for the more technical elaboration on my point audioskeptic.

      Delete
    6. Stereo mixes already employ panning so that one hears an instrument in both ears -- equally if the instrument is panned to the center in the mix. If you include a delayed cross-feed, you hear not only what was intentionally mixed (panned) to that ear, but also a very slightly delayed version of it at a lower level. At a frequency around 1700hz, you'd get the effects of phase cancellation with your 0.3ms delay.

      If you are going to presume that the engineers mixed the recordings for speakers, since that is what most listeners have, shouldn't you assume that the EQ was done for typical mid-fi speakers? Wouldn't it follow that the ODA should have a switch to roll-off the bass and high frequency response to mimic something like a pair of $300 Bose 301 bookshelf speakers from Best Buy?

      I want to see an ODA with the purist approach that the O2 employs.

      I know that there are arguments in favor of balance controls, tone controls, subsonic filters, high frequency filters, channel cross-feeding, Fletcher-Munson curve loudness compensation, parametric equalization, compression, expansion, digital echo, and countless other forms of signal processing, but each has a cost (parts, labor, S/N ratio, THD, complexity, etc.). And every person who wants one of those features will make the same arguments that I've seen here: It's inexpensive, simple, other products have it, and lots of people like it. Listen to all of those people, and you end up designing the next Pioneer AV receiver.

      I don't want to pay for any of it or have switches and pots to control it.

      If you want crossfeeding, there are many options available to you. You can rip CDs and make crossfed versions of them. You can use digital playback software that does crossfeeding. You can build a summing amp that does the processing. And the beauty of that is that it doesn't affect any of us who don't want cross-feeding.

      Delete
    7. @Fred M., once again you are taking a simple request and trying to make it seem complicated. You also sound kind of angry about it. Sorry if so; this was never about annoying you or anyone.

      I'm glad I raised the request here, because there has been some worthwhile discussion, and I plan to try out that Foobar2000 plugin.

      Perhaps the ODAC should be designed with provision for an optional effects card, and a simple crossfeed effect could be made as an optional plugin card. As long as I can get it, and there is a control on the outside of the box to enable/disable it, I'm completely happy; I don't need to push it onto you also.

      @NwAvGuy, my offer stands: if you would like a chance to listen to what this sort of crossfeed sounds like, I'll give you a demo of my software. Then you can make your mind up about whether it is worth the additional time/effort for you to change the ODAC design to make a crossfeed possible.

      Delete
    8. OK... I think the various views on crossfeed have been at least reasonably well represented here (for those just now reading the comments there's more below). I'm aware there's a minority who enjoy crossfeed and, for various reasons, would prefer it be implemented in hardware in a headphone amp. It's a very subjective thing and somewhat goes against the O2/ODA principal of the amp simply disappearing and not modifying the signal. I'll think about ways it might be optionally implemented

      Delete
    9. @steveha: While the request is simple, the ramifications are complicated. That's the nature of psychoacoustics.

      I'm not angry, just opinionated.

      I'm a purist. I use the "bypass" outputs on my Adcom preamp. They bypass the tone controls and loudness contour circuitry. I don't have any signal processors: No equalizers, BBE Sonic Maximizers, or SRS (Sound Retrieval System) units. While I do have an AV receiver for video use, I don't use its "CINEMA DSP 3D (17 DSP programs) with Adaptive DSP Level," the "Adaptive DRC (Dynamic Range Control)", the "Compressed Music Enhancer," and so forth.

      I have no problem, whatsoever, with features added through software. If the driver for the DAC does crossfeed and makes you happy, great. Where I have issue is when a minority-requested feature gets implemented in analog hardware. It increases the (volunteer) work for NwAvGuy and the cost for every person who wants to build, buy, or sell an ODA (switch, pot, hole in front panel, lettering on front panel, etc.). It may also decrease performance. possibly adding noise or distortion) even when not in use depending on the implementation.

      Delete
    10. If interested in crosfeed processing, give a try to the 112dB Redline Monitor plugin - so far IMHO it's the best emulation of speakers via headphones... See http://112db.com/redline/monitor/

      Maybe someone wants to reverse engineer that algorithm? :-D

      Delete
  33. NwAvGuy:

    First of all, congratulations for all the content on this website and the stuff you have designed! It's truly an amazing resource and quite an achievement :)

    Some time ago I bought an Audio-GD DAC19. I searched your blog and it seems like you don't like their approach in products such as the NFB-12.

    However, I was wondering if the DAC19 might be better given that it uses the "legendary" Burr-Brown PCM1704UK chips? Or should I sell it and get an ODAC later this year?

    Thanks!

    ReplyDelete
    Replies
    1. Given what I know it's hard to trust anything Audio-GD sells. If there were an online audio dictionary, and they defined the phrase "snake oil", they would probably refer you to Audio-GD as an example.

      The ODAC will have published results that demonstrate its performance and audio transparency. Unless Audio-GD can publish similar measurements, and they can be confirmed by someone unbiased and independent, I would be suspicious of their products. Their website is full of conflicting and mythical information. From everything I know, including the reviews by Samuel Groner, Audio-GD's products are poorly engineered.

      Delete
    2. Thanks for the info. I tend to favor truth over blissful ignorance :)

      So let's say I sell my Audio-GD DAC-19 for ~$600.
      What would be the best "objective" replacement among the following?
      1) Benchmark DAC1 USB, 2) Centrance DacPort LX, 3) ODAC

      Delete
    3. I guess it depends on your priorities. The DacPort is USB powered which might be an advantage for some. If you want multiple inputs and more of a desktop DAC, I would suggest the Centrance DacMini. If cost is little object, I would get a Benchmark DAC1 (or perhaps a Grace). If you can wait, and especially if you want a high quality integrated headphone amp, I would suggest the ODAC/ODA as a way to save some real moneyh.

      Delete
    4. I'm currently running the following chain: PC -> Audio-GD DAC-19 -> Pioneer Elite A35-R speaker amp -> HiFiMAN HE-6

      Would I still be better off with an ODAC/ODA than with a standalone DAC + relatively cheap speaker amp? I've read that the HE-6 is so demanding that it's best paired with a speaker amp. However, I think you said the HE-6 was one of the edge cases that the O2 design allows for.

      Or who knows, perhaps it's better to sell the HE-6 and use the more efficient HE-500 with an ODAC/ODA :P

      I hope these questions don't overwhelm you, as I'm well aware that you don't have the time to give personal advice to every music lover out there :P I hope that this can genuinely contribute to the conversation. For example, by making it clear whether an edge case like the HE-6 can be handled by the O2/ODA.

      Thanks!

      Delete
  34. Hey Pal, I received my O2 kit form europe at lunchtime today and finished the build a few hours later. Wow, great sound, I ma really happy with it.

    I went with the 2 x and 5 x gain resister sets, which is ample for both an iPod as a source or my new favourite, the Colorfly C4. I want to thank you for engineering on our behalf. Very cool and can't wait for the ODAC.

    Cheers, Pal

    ReplyDelete
  35. IMHO the most convincing crossfeed implementation I've come across is this one: http://www.hydrogenaudio.org/forums/index.php?showtopic=90764

    The author provides some info on the implementation and will likely provide more if requested.

    I have question on a different topic, what's a good design for a headphone distribution amp? I'm looking at this Jaycar kit right now: http://www.jaycarelectronics.co.uk/productView.asp?ID=KC5417&keywords=kc5417&form=KEYWORD Here's the circuit: http://i32.tinypic.com/2rmun1c.jpg

    Cheers!

    ReplyDelete
    Replies
    1. xnor's crossfeed implementation is my favorite as well.

      I believe controls like crossfeed, EQ, etc. are best done in software anyway, but I guess others like them in hardware. Maybe it's more important for portable systems, which may not have software that supports such controls. For DIY amps with through-hole parts where size is a constraint, I doubt it's worth putting in though, as it's not something everybody would want.

      Delete
    2. I do agree that its better to do in software but that's not always practical. My main listening setup uses the TB Isone VST in foobar but I can only use that with with content played through foobar or some other media player which supports VST plugins.

      To my knowledge you can't use something like that globally so all audio played from your PC is processed. (At least not without too much latency and/or dropouts like I got when I tried running software loopbacks through a DAW.) That makes good DSPs like that no good for internet radio, youtube, streaming movies, etc, etc so I still need a hardware implementation. At the moment I use my UHA-4 and its crossfeed circuit with that kind of content.

      I'm a huge fan of crossfeed but I also agree that it probably shouldn't be added to main design since not everyone needs it. If I listen to anything but recent rock/pop type stuff that doesn't have much stereo separation to begin with then I tend to get headaches without using crossfeed. I need it pretty bad. Some symphonies will give me headache in 5 minutes or so.

      Because of that, I'm going to add a crossfeed circuit to my ODA since space won't really be an issue but its not going to replace TB Isone when I'm able to use it.

      Delete
    3. @mikeaj, my use case for wanting it in hardware is that I am not always listening to Foobar2000.

      I did say there should be a switch to disable it; if you have a crossfeed in software and you want to use it, you could disable the one in the headphone amp.

      And I don't insist the crossfeed be retrofitted to the O2; I personally would be content to just have it on the ODAC. (But Headroom has their crossfeed feature on all their products, including the least expensive portable ones.)

      Delete
    4. Rockbox compatible devices can use a crossfeed implementation that's very similar to xnors crossfeed: http://www.rockbox.org/tracker/task/11577 (the very first one posted)the downside is that you need to compile your own build, but that's not too hard if you follow the wiki. It's much more realistic than Meier type crossfeed which would certainly be a waste of board space IMHO.

      Delete
  36. Quote from the MORE POWER section: "TESTING WITH FLIM & THE BB’s: Next I repeated the above test but this time using the same Flim & the BB’s – New America track shown earlier." Wow, never met anyone else that listens to Flim &The BB's! I have all their CD's, aren't they wonderful? I always equated them with "happy music". Too bad they don't record any longer, such a shame. I have all of Billy Barber's stuff as well, "Shades of Gray" being a favorite. As I am re-reading everything again from the beginning, I ran across this tidbit.

    I received the AC adapters from Jameco (ADU160100Z5401) 16VAC/1A and finally listened to my 3 boards, sans case. After a few days of listening to jazz, classical and dance music I have some observations.

    There is absolutely no hiss, even on my SF3's and TF10P's. There is an effortless quality that doesn't sound edgy, strained or dynamically compressed. The better the recording, the better the O2 sounds. Or more accurately, you can hear the real recording as it is with no coloration from the amp. It really does simply 'disappear'. Differences I do note are from other devices lacking in some way. When compared to my iPod Touch 3g (which get adequately loud) the bass has more authority and control. Gone are the little 'chuffs' or 'flapping' noises on some deep bass notes from being underdamped. There never seems to be a loss of control, even at loud volumes. The mids and highs are there without graininess or viel. If there is 'air' in the recording, it comes through the O2. Nothing is exaggerated or diminished. When compared to the HO output of my PreSonus 16.4.2 mixer, the differences are much more dramatic. Again, because of the deficiencies of the mixer. Connected to the 2Track Out RCA connectors (+18dBu, 100Ω) and swapping jacks, it is quite easy to hear an SQ difference between the O2 and mixer HO. Even makes the much used Digidesigns (now Avid) MBox Pro (33Ω HOZ) sound better. Now, I can hear what my recordings and mix-downs really sound like. I am wholeheartedly recommending the O2 (and upcoming ODA/ODAC) to all of my pro-sound friends. I just know some are going to want to use as a battery powered body pack or use on mic stands to power their personal monitor headphones/IEM's.

    Awesome job NwAvGuy, cheers!

    ReplyDelete
  37. More about the crossfeed idea:

    The perfect crossfeed feature would be in the headphone amp, and would only affect the headphone jack. Any other outputs (for connecting to other audio gear, or perhaps for connecting self-powered computer speakers) should just get the passthrough signal.

    And, while it may be easiest to do it in the digital domain, ideally it shouldn't just work on DAC content; ideally it should work on analog input as well.

    Of course this is a feature on Headroom amplifiers. Do you have any plans to review anything made by Headroom? I've been tempted to buy the Micro DAC and amp, after reading some favorable reviews.

    http://www.dansdata.com/microstack.htm

    ReplyDelete
    Replies
    1. I've heard bad things from good friends about the micro dac.

      Delete
  38. The wait is killing me :( I might just go with O2 and buy the ODA+ODAC later.

    ReplyDelete
    Replies
    1. I have made up my mind: I'm going to buy an HRT Streamer II and an O2, and put those on my desk until I have an ODAC.

      Delete
    2. I already did. The O2 lives in my office at work and the ODA+ODAC will live at home.

      Delete
    3. I was going to buy this tomorrow but I have no skills in fitting in the ODAC once it comes out. Any one knows when ODAC+O2 or ODAC+ODA is coming. I really need a good amp and DAC. I dont mind O2 lone I can use that with my hm 601 or ipod rockbox, if the wait is too long I can just go with the O2 .

      Delete
  39. Because I want to be able to use my soundcard's DSP abilities in game, I currently connect my NFB-5 over optical to it. Unless the ODAC gains optical at some point, I'll need to use either a trick where I set the ODAC to output the same audio as the X-fi (induces lag so I'd prefer not to) or I could connect it via analog out from my card and use the USB input for music. The second option seems better as the X-fi is acceptable for general audio, but I'd like to make sure I don't feed the ODA too strong a signal, do you think that the X-fi would have enough output to damage the ODA?

    Also, once I get it finished, would you be at all interested in testing my NFB-5?

    ReplyDelete
    Replies
    1. If you've got a multi-meter and a signal generator program you can get a good idea of its output by measuring a 60hz sine wave at full volume.

      Delete
    2. well I guess the O2 is designed to fill a different gap in the market, and not one for gamers because sound cards already fill that niche. Eg, Stereophile gave a raving review on the Essence STX. And I doubt NwAvGuy will want to cover the same ground as Stereophile, which is a well respected publication (you just have to read between the lines a bit, as they sidestep a lot of obvious issues to their target audience wide).

      Delete
    3. come to think of it however, USB makes it incompatible with almost anything that's not a PC, so what is the "void"? Hopefully an optical extension board will come out in due time, whether by NwAvGuy or someone else.

      Delete
  40. Any chance you will add a bass boost (only subtle mind, not too overpowering) on the ODA? Also, ETA on the new desktop version? I'm literally brimming with excitement!

    ReplyDelete
  41. @ NwAvGuy: Whilst you're at it, could you make it dance the tango?

    Seriously @ the"Can you make it do X" people, the phrase about not looking a gift horse in the mouth springs to mind :D
    It's even sillier directly following a blog post talking about how much work NwAvGuy has *aside* from giving people on the internet free amp/DAC designs with thousands of words of documentation!

    ReplyDelete
    Replies
    1. I was going to say it, and then you beat me. Although, nwavguy does encourage suggestions, if enough people ask for something, he tries to make it happen.

      Delete
    2. Hi Willakan
      Why were you banned from Head-Fi?
      It's a shame
      Take care

      Delete
    3. I'll be unbanned in due course. I expressed my strong belief that a certain individual who believes firmly in audio woo and consistently insinuates that everyone who does not is retarded should drown in his own excrement.

      In hindsight, perhaps a little OTT...

      Delete
    4. I suspect trying to promote objective audio information didn't help your cause with the H-F admins much either. Such information goes directly against their primary business model--hence the ban on discussing blind listening, etc. Many of their sponsors sell products that have no objective benefits whatsoever compared against lower cost options.

      Delete
  42. NwAvGuy,
    I, for one, do NOT want any extraneous circuits of any sort on my ODA/ODAC. I want the most pristine sound you can humanly manage within your predetermined constraints and objective goals. This means absolutely no bass EQ, mid EQ, treble EQ, parametric EQ, shelving control(s), Aural Exciter®, Big Bottom®, Sonic Maximizer®, SRS®, DSP shenanigans, cross-feed (of any flavor) and especially no FOTM crapola. As a compromise, perhaps a quad RCA jack could be used as an input and loop-out connector (like an insert jack on a mixer) for adding any outboard gear a user may wish to use. As for me, I will be happy and very thankful to be in receipt of of your best free-to-the world open-source DIY desktop headphone amplifier with your co-developed 96kHz/24bit USB DAC add-on board that also fits inside the O2. Yes, I will thank you and be very happy, indeed. Cheers and best regards!

    ReplyDelete
    Replies
    1. We are of like mind with regards to the purist approach -- and even both rely on Flim & The BBs ("Tricycle", in my case) for testing. Ever heard "Live From Studio A" with Johnny Frigo with Bucky and John Pizzarelli? I highly recommend it as one of the finest live recordings you will ever hear. Also great for headphone testing.

      I love my O2 and am looking forward to the ODA and ODAC. I just want to again say 'thanks' to NwAvGuy for his generous contribution of time and expertise.

      Delete
    2. Yes, I have loved Flim since I found them in the early/mid '80's. In fact, as a little history, they were one of the very first to ever record straight to digital using a big Mitsubishi multi-track recorder. They were the first to record non-classical music to the CD format. Their music was digital-ready even before digital was ready! The Frigo/Pizzarelli "Live From Studio A" is available from HDtracks for only $2.49/trk or album for $17.98 all tracks are in 96kHz/24bit FLAC. Chesky Records is well known for producing exceptional recordings along with Sheffield Lab, DMP & GRP to list a few of my favorites. If you find any of those labeled CD's buy them, no matter what is on them, you will not be disappointed!

      I have built 3 O2's so far and each is exceptional. Besides the on/off transient (which my iTouch 3g is just as loud) I am hard pressed to come up with any real sonic deficiency. The only thing I wonder is if dynamics would/could be improved by going up to 30VAC rails like the DAC1. If there is any dynamic compression in the O2, it is very slight IMO. In my honest opinion, the O2 may be as good as it gets before getting silly with a cost-no-object design. Then, you will need very high end phones to resolve any additional detail plus factor in most people's non-perfect hearing and untrained ear. Where do you draw the line? I think the O2 has laid down the gauntlet.

      Delete
    3. Flim & the BB's was one of the first to really show off the dynamic range of CDs before the loudness wars started to take hold.

      The DAC1 uses +/- 18 Vdc rails but it also uses much more robust (and expensive) output buffers capable of handing the increased dissipation from the higher rails. It can manage a bit over 10 Vrms to the O2's 7 Vrms which is about 3 dB more output.

      If you're using a headphone that's on the edge of the O2's voltage capabilities (i.e. DT880-600) AND you like to listen to wide dynamic range recordings loud, you may get some slight benefit from that extra 3dB. But, for everyone else, the O2 is unlikely to clip unless the listener is really into hearing damage.

      Truth be told, I suspect most headphones will be compressing their acoustic output long before the O2 approaches clipping. That's likely true even with the DT880-600. It's very true with my DT770. I've measured the output at high levels and they start compressing (with dramatically higher distortion) long before you reach their maximum rated input. Driving them even harder from any source won't help much.

      Delete
  43. NwAvGuy, I was just wondering about the 24/96 specification for the ODAC. AFAIK there's very little benefit of 24/96 over 16/44 for playback (correct me if I'm wrong), so is it just a low/no cost option to make it more marginally more compatible? Have you thought about keeping it 16/44 to keep costs down?

    ReplyDelete
    Replies
    1. This has been discussed in a fair amount of detail in several places on this blog. If the PC volume (player and operating system) is left at 100% you are largely correct. One blind study showed an extremely slight, but detectable, benefit to 24/96 but it's hard to argue it's very significant. But for anyone that wants to turn the volume down in software, a 24 bit USB DAC (assuming it offers >16 ENOB) can have an audible benefit because of the higher bit resolution and lower noise floor.

      A secondary benefit is 24/96 recordings are often mastered with more care and a wider dynamic range and sound better than their 16/44 counterparts--this is also true for SACD two channel recordings. A 24/96 DAC allows playing such recordings in their native format rather than downsampling them with the potentially audible artifacts that can happen with re-sampling (especially when it's done on the fly by the operating system).

      Delete
  44. NwAvGuy,

    Would it be possible to build the ODAC as a standalone unit with it's own case to complement the existing O2 amp? Alike to how the algorhytm solo works?

    ReplyDelete
    Replies
    1. Sure. Although the plan is for the ODAC to fit into a smaller enclosure than the O2. But a DIYer could find a way to secure the board in the B2-080 enclosure the O2 uses if you want them to be the same size.

      I assume you know the ODAC will also fit inside the O2 if you're willing to give up the batteries. Portable headphone DACs are an interesting challenge. USB power works reasonably well to run a DAC but has some significant limitations for operating a headphone amp. Even trying to get +/- 9 volt power rails out of 5 volt 2.5 watt USB power budget with enough power left over to run the USB interface and DAC is difficult.

      It's impossible to power the O2 and get the same performance it delivers on battery from USB power. Nobody can design a USB powered headphone amp that can come close to the O2's performance.

      The above is partly why the FiiO E10 has barely enough power for the HD600/650/etc and also why it has odd clipping behavior into low impedance loads. It's also partly why the Nuforce uDAC-2 has some of its limitations.

      So for those who want to use power/voltage hungry cans for portable use, you need battery or AC power for respectable results. The O2 paired with an external ODAC will be a good solution. You can also, of course, use other USB powered DACs as has been discussed.

      Delete
    2. Have you considered using the USB just to charge the batteries of a portable DAC/Headphone amp. The batteries could provide the current for the +/- rails. When the unit was plugged in but switched off, the batteries would be getting charged with the full capability of the USB port.

      Depending on the design trade-offs (low-current vs. higher performance), when it was reproducing quieter passages, the USB might be able to provide enough power to provide some charge current.

      The 500mA limit on USB is going away. Modern MacBooks can supply 2100ma. Some Asus motherboards can supply well over the 500mA. USB 3.0 has a minimum of 900mA. And there's always the old trick used by external 2.5" hard drive manufacturers: A double-headed USB cable to suck power out of two ports at once, giving a full amp at 5V.

      Delete
    3. Yes, that's certainly possible although it's still a battery solution. The best way to do it is with a fairly beefy single cell Li-ion (3.7V) battery running an isolated DC-DC converter with dual (bipolar) outputs for +/- 9 or 12 volts. Unfortunately, as I've discussed many times, Li-ion batteries are dangerous in the wrong hands, cause lots of serious fires (even in some commercial products), and in fact are not even supposed to be sold directly to the public except as OEM replacement batteries for things like cell phones, cameras, etc. They're only sold in raw cell form through eBay vendors and small "gray market" dealers that source cells of questionable quality from China where there's little concern over lawsuits.

      The USB standards over 500mA are frankly a mess. USB 3.0 helps quite a bit, but for now, that rules out 95+% of computers. The USB power management chip vendors are having a hard time coming up with reliable ways to "auto negotiate" current capability from a USB 2.0 port for any value over 500 mA. If it's a dumb wall charger that's fairly easy, but if it's a hub or PC, it's not so easy as going over the limit will generally shut down the port and it stays shut down on most PCs.

      Apple's "over 500mA spec" is also a moving target and arguably proprietary. The iPad 1 is different than the iPad 2 for example. Many phones that are capable of using over 500 mA also follow different standards.

      The bottom line is I don't want the liability of burning someone's house down with a Li-Ion solution gone wrong. While it's possible to use NiMh, that has another set of problems--especially if they're bipolar dual batteries as in the O2.

      Finally, any USB solution will be surface mount as all the suitable USB power management chips are surface mount. Many are in ultra-tiny IC packages intended for use in cell phones, pocket cameras, etc. Many don't even have exposed solder pads. That goes against one of the primary design goals of the O2/ODA. So it would likely require a commercially assembled board.

      Ultimately, it would be fairly easy for a commercial manufacture like FiiO to offer such a product. But it's not something I'm comfortable doing as an open source DIY design.

      Delete
    4. Li-ion is all fun and games until the battery explodes. They are super picky about their charging and I agree about keeping it out of this kind of design for safety reasons, especially because the O2 is designed to be beginner friendly.

      No matter how badly you mess up the O2 I don't think you could get more than a painful but fairly harmless jolt or maybe make the batteries leak some electrolyte on your carpet. Poor treatment of NiMh usually just gives you shorted battery life. OTOH if you mess up a Li-ion charging circuit just a little it could burn down your house. Trying to build your own charging circuit for the battery because the commercial offering are tiny surface mount chips amounts to Russian Roulette unless 00you've got some fairly precise testing gear and can figure out all the failure mode you should test against as well.

      On that note there's a new Li-ion version of the Mini3 that's supposed to have longer battery life. That would probably be last on my list of things to change about the design but whatever.

      This also makes me wonder what kind of insurance Bottlehead must have to sell those kits with bare semi-HV transformers that plug directly into the wall.

      Delete
    5. As usual, a thoughtful and coherent response.

      I've got some Cree LED flashlights that used the 18650 Li-Ion cells and the cells been rock-solid in commercial chargers. But your points are well-taken with regards to DIY efforts where a solder short and a hungry-for-blood lawyer could be disastrous.

      Come to think of it, mine are "UltraFire" brand, so I might not have much of a leg to stand on in a court case should they burn down my house. ;-)

      Many of the cells being sold now are name-brands like Tenergy and Panasonic with Amazon even starting to fulfill a lot of the orders. The no-name cells are not so common there.

      "Still a battery solution" is kind of like "still a capacitor solution." If they allow a DAC and headphone amp to run off of USB, I'm fine with that. On the other hand, I'm not holding my breath for FiiO, Behringer, or any of the others to produce a top-drawer product any time soon. I'll get an ODA/ODAC and live with the wall wart.

      Delete
    6. Thanks for the added comments. I'm not aware of any Panasonic Li-Ion generic cells being sold by any "legit" USA dealers (as in authorized by Panasonic). Digi-Key is a huge authorized distributor for Panasonic batteries and they don't sell a single Li-Ion battery.

      Another common fraud are all the "A123" cells being sold on eBay and by no-name companies. If you call A123 in the USA they'll tell you they're 99% fake "knock off" batteries made in China not by A123. They only sell to true OEMs and they're prohibited from reselling the batteries.

      Amazon is a bit tricky as they're just a credit card processor for a huge number of vendors many being really tiny. Their contracts with those suppliers release Amazon from any liability with respect to the products being sold, etc. If you do your homework on Tenergy you'll find they're really just the USA distribution "front" for a Chinese company. There are no deep pockets in the USA to go after if your house burns down.

      Delete
    7. The Amazon search I did was for "Prime" eligible purchases - ones where Amazon is doing the warehousing and order fulfillment. They ship 18650 Li-Ion cells for multiple vendors. Now whether that's wise from a legal standpoint is another matter, and there don't seem to be any warnings or attempts to qualify buyers.

      I bet that you are right that the Panasonic cells are gray market. "LED Wholesalers," "Orbtronic," and other such names pop up as the sellers.

      I'll stick with my UltraFire cells. With a name like that, they must be safe.

      Delete
    8. On the topic of wall warts, what do you think of aftermarket 12V power supplies? I know audiophile ones are all a waste of money but what about ones like Pyramid for about $25 that can take 5A?

      Delete
  45. Question: Would any DAC below 200 dollars really be significantly better than a modern pc DAC? I want to get the O2 but do not want to wait for the complete ODAC/Desktop amp to be completed :P

    Assuming the O2 can be retrofitted with ODAC, I wouldn't be bothered in this case

    ReplyDelete
    Replies
    1. Yes, easily. Look at my review of the latest MacBook Air for an example. Even the FiiO E10 beats it in many ways (and it's way under $200). And a lot of modern PCs are worse than the MacBook.

      And yes it's very likely the O2 can be retrofitted with the ODAC (if you want to give up battery operation).

      Delete
  46. If you're really interested in addling lithium ion/polymer batteries to the O2 somehow, I think you're going about it the wrong way. Never build what you can buy. Rather than integrating bare cells into the design, a better option might be to take a page out of AMB's book and design it to accept a commonly available packaged battery. Take camera batteries, for example. The batteries are small and widely available, and they feature another thing I think you should look out, external charging control circuits. Rather than building them into the device, they are part of the adapter. I believe that Dyson (the vacuum company) used a design similar to this with their latest handheld models, the vacuum itself includes only the low voltage cut off, the charge control circuitry appears to be in the wall wart.

    http://www.dyson.com/store/partsDetails.asp?part=ACC-DC30_31-CHARGER&product=DC35

    That's where the charging status light is at least.

    ReplyDelete
  47. Like a number of others, I've been chomping on the bit waiting for the ODA/ODAC to become available, so have rather jumped the gun and ordered an O2 to drive my modified Fostex T50RP headphones, at least for the meantime. (I only need a desktop unit). I will use my Fiio E10 as a dac only. Question: with your having done a lot of listening and testing with the Fiio E10, do you think that the ODAC - when it becomes available - would make a worthwhile "upgrade" sonically to complement the O2, compared to the E10 as dac only?

    Thanks!

    ReplyDelete
    Replies
    1. That's a good question. The ODAC easily outperforms the E10 in measurements--especially the dynamic range and noise floor in 24 bit mode. How much that might be audible depends if you plan to control the volume on the PC side or always leave it at 100% and use only the O2's volume control.

      In terms of other sonic benefits that will have to wait for more blind testing and might come down to the source material (music), listener, and how revealing their headphones are. My goal is to have the ODAC/ODA be difficult or impossible to tell apart from the Benchmark DAC1 Pre in blind tests with any track, any listener, and most any headphone. The E10 can't do that.

      Delete
  48. NwAvGuy,

    Thanks for all your hard work. Would a headphone such as a Grado be worthy of an amp such as the O2?....or would you say that this phone should just be used from the headphone out from a source such as an ipod or clip+?

    Would you be able to tell a difference?...or would it just be able to make the headphones go louder?

    Thanks!

    ReplyDelete
    Replies
    1. That's hard to answer. Grado makes headphones from $69 to $1000+. The cheaper ones tend to be relatively easy to drive. The Clip+, in particular, is a decent match for the reasonably priced Grados. The iPods have a higher output impedance than is ideal for the 32 ohm impedance of the Grados so they're not as well matched.

      You might notice better sound using the O2 with an iPod, but you're less likely to notice an obvious difference with a Clip+.

      Delete
    2. Thanks

      I was thinking of a RS1i using my Laptop with lossless and dac'd and amp'd maybe through the ODA/C. Would this be a good set up?

      Delete
    3. Perhaps someone else might want to comment on the RS1i as I'm not that familiar with them. My personal opinion is the least expensive Grado's are a great buy if you're OK with the comfort but as you move up the line they become less appealing for the price--largely because they're still mostly the same clunky design as the cheapest ones. But headphones are highly subjective and everyone has different priorities.

      But, regardless, the O2 and ODA can drive any Grado made with 100% transparency and lots of headroom left over.

      Delete
    4. Thanks!!! Appreciate it.

      Delete
    5. I use the o2 with magnums. Amps don't make much of a difference (so long as they are transparent) good dacs make a difference, of course. the way I see it, you can't get a cheaper transparent amp than what nwavguy offers, so yes. It is worth it just for the cleanliness, will it transform your RS1's? No.

      Delete
  49. Can you give us an update on the possible release date on this?
    Thanks

    ReplyDelete
  50. I just received my O2 from JDSLabs. I can say immediately the sound is far less grainy and more enjoyable to listen to then the Schiit Asgard it replaced. Gone is the low level hum (which I assumed was coming through the AC power, though isn't present on the O2 even when plugged in.) I just wanted to say, thanks for saving me some money, and thank you for your contribution to my hobby :) I can't wait for the ODA, and please, please let it have a pre-amp out (that isn't volume controlled by the volume pot!) :D

    ReplyDelete
    Replies
    1. For me ( and anyone who has, or is interested in active speakers) having an option to control the volume on the preamp RCA outs is critical. I think this option makes a DAC far more versatile.

      Delete
  51. The O2 over at Head-Fi is getting quite fun. It started off with the whole inevitably "measurements are silly" furore, then petered out for a bit: now people are taking turns in a thread that sits firmly at the top of the amplification subform to say how great it sounds.

    To make things even more fun, someone is about to post a giant thread comparing the V200 and the O2: his conclusion is the O2 is the more neutral :D

    I think he's even done some blind tests!

    Regardless of what sounds like what, the fireworks should be pretty!

    On a completely different note, Epiphany Acoustics is apparently not grounding the O2 circuit board to the case in their products. Might want to call them up on it: don't want anyone sullying the O2's good name after it starts humming with their DAC!

    ReplyDelete
    Replies
    1. Thanks for sharing both those things. I haven't had the time to bother with Head-Fi but that's encouraging. That was my hope all along--to demonstrate how you don't need high-end components or esoteric designs to get really good sound.

      Unless Violectric messed up the V200 in some way, I would expect the O2 to sound the same with any headphone that's within the O2's output capability in blind testing. Considering the V200 is nearly ten times the price, that's an eye opening comparison.

      I suspect many from the Schiit/Bottlehead/etc. crowd will either turn against all amps that measure well--even high-end ones like the V200--or perhaps start to wonder if they've been going down the wrong road. It's encouraging to read comments like the one from a_recording above finding the O2 sounds better than the Asgard. The tide seems to be turning.

      Delete
    2. Hi NwAvGuy. Thanks for the response! I also post with the same handle on Head-Fi. I'm really encouraged by the O2. One thing I did this morning was snip the high gain resistors off to get a 1x gain, so that I could use the O2 from a line level source off battery power. (I don't really have any gear that needs the 6.5x gain, while I do have a LOT of low impedance headphones like the 24ohm Sony Z1000 or a number of IEMs). John from JDSLabs is now detailing the gain customisation options on the JDSLabs website. I'm curious, given that in your cmoy amp review demonstrates that the performance measures better with a 1x gain (despite the setting not being particularly useful in many cases) have you ever done any measurements with the O2 and a 1x gain? Would you expect even better numbers?

      Delete
    3. I explain the difference more in the O2 design articles, but the output stage of the O2 always runs at 1X gain. Most of the distortion in a headphone amp happens in the output stage. So a single stage Cmoy performs much worse at higher gains because there's less feedback available to correct errors at the headphone output.

      The O2 doesn't have that problem as it always has 100% of the feedback available to the output stage regardless of what gain the input stage is configured for. Plus, in a Cmoy, the noise gets much worse as you increase the gain because the single stage always runs "wide open" to the headphones. But the O2, with the volume control between the stages, doesn't have that problem either. The O2's noise (what little it has) is always in proportion to the listening level.

      So, in summary, the measurements won't improve much from 2.5X down to 1X gain with the O2. But that's because the O2's performance is largely independent of the gain. It always performs very close to the 1X gain ideal up to about 7X gain.

      Delete
    4. Thanks for the reply. :] I'll just enjoy this amazing little box then. Do you happen to know the line out voltage of a device like an iPhone 4?

      Delete
    5. I believe the iPhone 4's line out dock connector (LOD) is a relatively low 0.5 Vrms at 0 dBFS (max output) just like the iPod Touch but I haven't personally measured the phone.

      Delete
  52. As far as "review sites" go, Headfonia mentioned it: apparently it's great for the price, but lacks "dynamics," which only come when you spend more (Price is directly proportional to quality, right?)

    But fear not, for a mere $800 you can get the REAL DEAL desktop-simulating small amp, the Triad Audio L3.

    Best of all, it uses this revolutionary three channel, virtual ground design based design! Would you like some IMD with that snake oil?

    As you say, some people never learn...

    ReplyDelete
    Replies
    1. The Triad website provides many clues their products are not designed to any sort of valid objective criteria. It's full of snake oil and misleading and/or erroneous hype.

      If Headfonia said the O2 sounds as good as most anything at any price what does that do to the entire headphone amp business model and many Headfonia advertisers? They would be cutting off their own head if they didn't somehow justify the need to spend more. Sadly most all sites, even InnerFidelity, are in a similar situation.

      If someone wants to have some fun, challenge Headfonia to a public blind listening test with the O2 against something like the AMB beta 22, Headamp GS-1, or Violectric V200. After all, if we're supposed to trust their "golden ears" isn't it fair if they demonstrate their reviewing abilities?

      Delete
    2. Well said. If they did say that then there really is no point in "high-end" audio is there? Headfonia like many other review sites never rate a lower priced product over a higher one

      Delete
  53. Also, Zombie_X of ZX Amateur Cables claims that he tried to take up your DBT challenge with his $1K Auditor and you apparently stopped responding?

    Anything in that? His wording is kinda unclear...

    ReplyDelete
    Replies
    1. NOBODY has contacted me with a request to participate in any of the blind challenges I've put on the table. Tell "Zombie_X" to start a thread on diyAudio or ABI, someone let me know it's there, and I'll gladly respond in an open public forum to any blind challenge requests.

      Delete
    2. Zombie_X has responded at Head-Fi: http://www.head-fi.org/t/429371/the-beyerdynamic-dt880-600-ohm-appreciation-discussion-thread/1890#post_8465451

      Delete
    3. Someone please suggest "Zombie_X" respond somewhere I'm not banned. I would suggest ABI or DiyAudio for example.

      Delete
    4. I am not sure if the ban allows you to read Head-Fi, but here is the complete reply from Zombie_X, he apparently does not have the intention of joining ABI or DiyAudio:

      "That's hilarious man, I spoke to him via e-mail about the Auditor and the challenge and he stopped responding. I even posted on his blog about it once and I think he deleted it. In the e-mail he said no one has ever challenged him to a test because they knew better, so I replied saying I'd take him up on the offer and no responses since. I sent him three additional e-mails with no replies. I'd find it hard to believe if he says he never got them.

      I no longer have any copies of said e-mail as that was almost a year ago. I wish I had saved them to post here.

      I don't know what his game is here but I wouldn't state stuff like this if it wasn't true. I have a reputation here and I am not going to try and damage myself on purpose. I remember he started stuff here in the past (With the Shiit gear) and got people upset and such. I'm not lying here at all, that's what he told me and I'm sticking to it. I know what me and him talked about and that's that.

      I won't join diyAudio or ABI for this, I have no need to join those places at all. I won't join to dispute this as it's a petty thing to even argue. I'll let it go and walk away.

      Anyways this thread has been a tiny bit derailed so I suggest we get back to the topic. I kind of derailed it with my bit about NwAvGuy but that's that guys."

      Delete
    5. I can safely say the above is NOT true. It's pure fabrication. It's rather telling he doesn't want to discuss it anywhere I'm not banned and doesn't have the alleged emails I supposedly sent him. If he was at all serious about the challenge what's so difficult about having an open discussion about it somewhere I can reply? He calls it "petty" yet seems bent on making a big deal out of it. I'll gladly discuss blind challenges on any suitably impartial forum.

      Delete
    6. Yes it looks very suss that he insists it is the truth yet does not have any evidence to support his claim. In this case it is his word against NwAvGuy; I tend to believe NwAvGuy, but even if you don't, there is no reason to accept his word at face value.

      Like NwAvGuy says, if he really wants to challenge then it should be done in an open forum where anyone can see the responses and proceedings!

      Delete
  54. Man, damn silly thing called 'real life' getting in the way..we want ODAC :( :p
    As an Audio editor on a forum, I only ever recommend two amps for full-size headphones (if they need it, otherwise I'll say they don't need it) that being the O2 and the cheaper Fiio E9. There's no point recommending anything else.

    Anyway, we want ODAC :P

    ReplyDelete
  55. While waiting for the ODAC (take your time) I ordered the Bifrost to A/B test it against my Titanium HD. I cannot find a single difference. The Bifrost will be going back under their 15 day return policy and I will patiently await the ODAC for my listening room.

    ReplyDelete
  56. I have a question about capacitors.

    I know that you didn't talk much about them, but how do (or can) caps affect the sound? Is there really an audible difference between, say, an Elna Cerafine and an Elna SILMIC? Or maybe between a Jensen and an Auricap?

    Thank you!

    ReplyDelete
    Replies
    1. No. In the O2 Design article I talk about caps. Blind tests have been done. Beyond using the right type of cap for the applications, blind tests have shown there's zero difference. You can also measure the signal across a capacitor in circuit playing music. No signal means, by definition, the cap isn't changing the sound.

      Delete
    2. From a purely electrical point of view, would it be safe to replace tantalum capacitors with electrolytics? The tant caps in question are 20+ years old, about time I replace them.

      Delete
    3. I would replace tantalums with tantalums. They have very different high frequency characteristics than electrolytics and typically much lower ESR. Especially in regulator circuits and switching power supplies tantalums, or the newer replacements (solid polymer, etc.), may be required for stability.

      Also, I'm not an expert on cap lifetimes, but to my knowledge tantalums tend to outlive most products they're installed in. Electrolytics degrade much faster. Unless the tantalums are used in a very high temp, and/or very high ripple curren,t application they probably don't need replacement even at 20+ years.

      Delete
  57. NwAvGuy, I'd like to ask what your thoughts are about tube amplifiers and if you have any recommendations. When reading your blog, I've only seen you bring up the Benchmark DAC1, Voilectric models and the Gilmore GS1 which are all solid state amplifiers.

    ReplyDelete
    Replies
    1. I talk about tube amps in the first O2 Amp article and they get at least brief mention in several other articles. If you want accuracy and transparency tubes are not the best way to get it. The O2 will likely outperform any tube amp, at any price, in terms of noise, distortion, frequency response, phase shift, etc. The O2 delivers the music as the artist and engineer intended while tube amps often alter it in audible ways.

      In my experience, those who like tube amps enjoy listening to their amplifier rather than music, are caught up in the retro/nostalgia, or enjoy the endless tweaking possible with many tube amps. It's sort of like comparing an old British sports car to a 2012 Porsche. The Porsche puts the old car to shame by virtually every objective criteria and by a huge margin. It's simply a much better car that's faster, handles better, quieter, and even gets better fuel economy.

      Tube amps are very much like old British sports cars. They have charm and character, but their performance is rather modest even when they're at their best. And, like old sports cars, they're also more fragile and often require more maintenance.

      Delete
    2. Gotta love the car analogies. =P

      Delete
    3. I personally never understood the modern need for thermionic valve components, especially in portable applications. A few have argued that valves are easier to replace than IC chips, but I seriously wonder how many new valves will be made in the next 10-15 years.

      Delete
    4. Not to mention ICs, under normal use, never require replacement. Tubes (valves), on the other hand, all degrade even sitting on the shelf. The whole "NOS" (New Old Stock) issue with tubes is largely snake oil. The metal components in tubes, and especially the special coatings on them that are critical to the tube's performance, slowly give up molecules to the vacuum inside the tube even sitting on a shelf un-powered. And the glass envelope itself can slowly let outside air inside (especially around the pins). And the above happens much more rapidly when the tube is powered up and hot. So, by their nature, tubes tend to require replacement while ICs do not.

      Delete
    5. One case I've heard in favor of tubes is that if you want to make an amp for electrostatic headphones, tubes are the easier way to do it. No idea if it was true or not, or if that was just in comparison to using discrete components.

      Delete
    6. To directly drive electrostatics you're almost forced to use either discrete components or tubes. To my knowledge there are no reasonably priced suitable high voltage IC amplifiers. So yes, there's some truth to using tubes electrostatics--especially if it's an OTL (transformerless) design.

      Delete
    7. Hm, many years ago to drive a digital loudspeaker with (4-bit, somewhere around 1980), I found some op amps that would take 400vdc supply (+-200. They weren't cheap.

      For an electrostatic I'd use a vfet, I think, with source feedback, and current drive to the headphone resistor. Just don't use a carbon resistor with swings of 400 volts, you WILL see the nonlinearity in the carbon resistor if you do.

      Delete
    8. Hmm.

      Considering that most vendors provide no warranty for their NOS valves, might not be worth the investment. On the other hand, they sure make rather pretty decorative pieces.
      While we're on the subject of electrostats, what is the practical application for such devices, or are they just limited to audiophile use?

      Thanks!

      Delete
  58. Hi NwAvGuy, I have been reading all your comprehensive reviews regarding budget DACs and I would like to thank you for that.

    Just like some of the readers here, some of us are curious about DACs that are currently in the market right now while waiting for the ODAC(and that won't break the bank!)

    When you recommended the FiiO E10 in one of your replies, do you think its the best choice for a macbook as compared to the other DACs(X-Fi Go, U3, UCA202...) you have reviewed?

    I currently own a DT770 and ATH-M50 if that matters. Thanks again.

    ReplyDelete
    Replies
    1. If your headphones sound fine now on you MacBook, and play loud enough, you might not need an external headphone DAC. If there's an obvious problem with your MacBook's headphone output, the E10 has the lowest output imepedance (a good thing) combined with the most output of the budget DACs I've tested. It's also the only one that makes at least an effort at 24 bit support.

      So if you don't want to wait for the ODAC, and the E10 is in your price range, that would be my choice. But if you're getting good sound now, don't expect any dramatic differences.

      Delete
  59. NwAvGuy, I just wanted to say keep up the good work. My faith in Tyll has been rather shaken in the last few weeks, with him getting progressively more aggressive in the comments on his articles, and his comments about recabling coming back to the fore. I'm also never convinced by the way he tries to interpret the measurements he takes of various headphones.

    Anyway, that's by-the-by. I was actually wondering if you expect the ODAC to work with the new Raspberry Pi computer, via USB - I guess that's mostly down to the operating system and what devices it supports. My main worry though was that since the Pi is powered by USB itself, it would not be able to power the ODAC too?

    ReplyDelete
    Replies
    1. The ODAC's current consumption is fairly modest but it could be a problem with the Raspberry. The ODAC needs about 125 mA nominal and about 150 mA worst case. A normal USB current budget is 500 mA. I'm not familiar enough with the Raspberry to know what its current requirements are but I would expect it to be near or over 500 mA peak at 100% CPU usage with HDMI and ethernet in use. Depending on the power source, that might not leave much for the ODAC.

      It would be interesting if someone gets the ODAC working with the Raspberry as it would be an excellent inexpensive platform for an audiophile-grade, yet relatively inexpensive, network music player.

      Delete
    2. "My faith in Tyll has been rather shaken in the last few weeks, [...]"

      I think this is just the sad market reality for any consumer audio journalist trying to make a living.

      Delete
    3. The most successful car enthusiast TV show ever, by a wide margin, is Top Gear. It is, in fact, the huge BBC's most watched show with 350 million viewers per week. While it's funny and entertaining many believe the reason it's so successful is because it operates under the BBC group that does not accept advertising.

      Top Gear is free to say almost anything they want about any car without fear of losing ad revenue. And they do. Jeremy Clarkson will drive a new car with a six figure price tag and pronounce it "complete rubbish". It's refreshing to watch a show where they have the freedom to say what they want--hence the immense popularity. The show is the envy of many other automotive journalists--not because they cover cars particularly well, but because of their unique freedom.

      Tyll Hertsen, and nearly all other audio journalists, don't have that freedom. Stereophile and Source Media pay the bills. The bulk of their revenue is from advertising the very products their journalists write about. I think Tyll is a good guy I'm guessing he also wants to stay employed.

      Delete
    4. Great to see that I'm not the only one who thought that these two devices belong together. It would be great to see the Raspberry Pi and the ODA/ODAC combined in a single enclosure (though I imagine that the ODA/ODAC may need some kind of shielding from the Pi ?).

      Given the nature of the OS, there is certainly some interesting scope for a one box audiophile music server; an open source Olive O3HD anyone ?

      Delete
    5. Within reason, once the ODAC and ODA are hopefully fully released and able to stand on their own, I'll happily contribute towards a hardware solution (if one is even required) if the software and user interface details get sorted out. It should be possible to run both from a single power supply to help further keep the cost down and keep it simple.

      Delete
    6. "and his comments about recabling coming back to the fore."

      Perhaps the Cable Mafia made him an offer he can't refuse.

      Delete
    7. You could always use a powered USB hub.

      Delete
  60. Thanks for the reply - using the Pi as a network music player was my exact thought, and it seems to be in the spirit of your work. In either case, I'll pick one up when I can do (demand seems to be very very high for them at the moment), and worst comes to the worst, I'm out £22.

    ReplyDelete
  61. I posted this question in the Head-Fi forum. . but I figure it is better to ask here...

    Is there any way the ODA version of the amp could have the option of being built with a variable gain control (as in a pot) instead of just a switch? I would think that would give it even more versatility, all the way from super sensitive IEMs at 1x, all the way to extremely low sensitivity headphones runing from a low voltage source at the full 12X (though I doubt there are many situations where that much gain would really be necessary in a desktop version. . . not likely to be using an ipod with a desktop amp, but you never know). With the desktop version it could be on the back of the amp by the inputs and out of the way. I don't really know how that would affect performance like noise, etc. I am also sure it is not a simple process and could increase the cost considerably, but I figured it wouldn't hurt to ask.

    ReplyDelete
    Replies
    1. If the gains can be chosen appropriately, having two or three gain options will usually best suit nearly everyone's needs. In the rare instance someone has a very wide variety of sources, and a very wide variety of headphones with wildly different sensitivities, it's possible more gain options, or variable gain, could be of slight benefit.

      If the headphone amp has an impressive signal to noise ratio like the ODA and O2, and it uses a high quality volume control that doesn't exhibit channel balance errors until extremely low volumes, there's no need to precisely match the gain. Such an amp can still perform in a 100% transparent way even if the optimal gain is off by several dB.

      Variable gain circuits have a number of disadvantages. Potentiometers--especially dual stereo versions--have far more weaknesses compared to fixed 1% metal film resistors. They have to be located in the most critical part of any amplifier--the feedback loop. Using pots to set the gain of an amplifier creates many potential problems. And using a pair of single pots requires the user have instrumentation to precisely balance the channels. Another option are voltage controlled amplifiers (VCA's) but they have their own significant disadvantages and often degrade performance.

      There are some very good reasons why even very high-end headphone amp manufactures use selectable fixed gains rather than variable gain. The short version is: The overall performance of a fixed gain headphone amp will nearly always exceed that of a variable gain amp.

      The ODA will offer more gain options than the O2. I'm confident in saying a variable gain amp will not offer any benefits over the ODA. In fact, even a much more expensive variable gain amp is likely to perform worse.

      Delete
  62. More gain options? Woot! I was pondering whether to ask whether that would be a possibility, but felt that I might come across as a bit of a hypocrite after my "NwAvGuy has lots to do" speech!

    ReplyDelete
    Replies
    1. No worries, suggestions are always welcome even if I don't agree with some of them ;) With more PCB real estate to work with on the ODA some features are relatively easy and inexpensive to implement.

      Delete
  63. Greetings, NwAvGuy!

    I've had the privilege of enjoying the O2 (acquired from JDS Labs - an organization indeed worthy of distributing your designs in the flesh) for a few weeks. I admit to giggling like a little kid whenever I occasionally come across a glowing TAS or Stereophile ad or review of some multi-thousand dollar headphone amplifier's obviously divine attributes.

    Whomever is credited with the axiom "the devil is in the details" was clearly aligned with your assertion that implementation is THE primary determinant of quality, if quality is to be judged on the basis of real world performance.

    God, I love, am awed and so grateful for quality and its proprietors. You and Mr. Seaber are members of that small club.

    I hope you don't mind a suggestion by an obvious audio amateur (me). After listening for a dozen or so hours to the O2 (and iPod Touch 4G) through Etymotics Research ER-4Ps ("ruthlessly accurate" (your words) to the point of almost being "cutting" rather than soothing, whatever that means), a physician friend of mine lent to me a pair of Grado SR-60s (the original iteration, not the current "i" version).

    I do not feel qualified to attempt to describe my reaction to these inexpensive 'phones (just look at my clumsy use of adjectives above) other than to state I was moved emotionally in a most positive direction.

    Given your audio acumen and experience and love of music beautifully and intricately reproduced, perhaps you may find a listen to these 'phones engaging.

    If the newer "i" (improved) version is better - wow. I fantasize owning a pair of Grado SR-225i (top of the "pedestrian" line of Grado headphones), but I fear that, based upon my reaction to the 60s, I would die of ecstasy overdose.

    Carry on!

    Jonathan

    ReplyDelete
    Replies
    1. I'm glad you like your O2 and JDS met your expectations. I've said many times the low-end Grados are something of a bargain especially if your anatomy finds the fit comfortable for long term listening. I've only heard a handful of Grados up to around the price of the HD650 and, in my opinion, the low-end SR60/80/etc are the best deal. But headphones are highly subjective.

      Delete
    2. Just wanna say as someone who went from SR80i to 225i to Magnums and has heard many in-between including a few different magnums set ups (cocobolo, limba, aluminum) and the allesandro MS-PRO.. The best sound you can get, will bet out of a magnum, and if you do it right, you can build them for about the price of a 225i.. Just wanted to toss that out there :)

      The jump from sr60/80 to 225 isn't that great, its more of a different flavor, with some more detail, same with the jump from 225i to ms pro, in fact, i favored the 225i's in wood... neither "upgrade" is jaw dropping, the jump from 225i to magnum, is certainly jaw dropping.

      Delete
  64. I am currently sourcing parts to build an O2, and one of the hardest items to find, and in my opinion the least elegant part of the O2 system is the AC wall transformer. The only one I can easily source with my countries plug (NZ/AU) has a bunch of adaptors for the output lead, which is rather fiddly as a solution.

    I was wondering if your plans for the ODA power supply include an internal transformer. Sourcing a local power cord is easier than sourcing a local adaptor, and would be cleaner from an ergonomic/aesthetic perspective.
    Am I right in thinking this could even be one part on the BOM that works for both 220-240VAC and 110-120VAC? If not could it still be easy enough to have 2 or more options? My amateur research gives me 553-FS16-150 from Mouser as a potential candidate, though the output capacity is a little more than I imagine is needed.

    I'd also love to see what sort of case you have in mind for the ODA. The size of it will probably tie in with the transformer decision, but I'm also interested in the aesthetics and want to start thinking about what I'd like to do with it. Things like volume knobs, feet, and panel design. Once you release all the physical specs I could even do some CG renders to help visualise prospective panel designs and other choices.

    I may also have a go at designing a custom enclosure to be CNC cut from clear acrylic plastic. If I do this and it looks good in a render I will of course send you the files, and probably get one made myself if it is not too expensive. After building the amp myself I would be proud to display the guts of it in a nice colour scheme.

    Oh and thanks for what you do. As a species, the sharing of knowledge as a public resource is our greatest strength, and helps all people increase their quality of experience, and ultimately their quality of life. Shared experience also helps to unite us as world citizens. With the O2, I feel you have made a significant contribution to the public knowledge of humanity, and I look forward to following your ongoing projects for as long as you keep working on them.

    ReplyDelete
    Replies
    1. Thanks for the encouragement. I'm aware the wall transformer has been an ongoing problem in many countries outside North America (although there are several decent options for most of Europe and the UK). While it may not be elegant from the view of finding one in some countries, it's very elegant from a safety, size, and cost vs performance perspective. The O2's power supply is somewhat novel among headphone amps and neatly solves the issues of shock hazards, safety agency approval, needing a chassis ground (which frequently create ground loops with other equipment), etc. I haven't decided yet on several ODA details with the power supply being one of them. Something like what you suggest is certainly possible but had its own disadvantages.

      Delete
    2. Thanks for the feedback, I'm sure you will come up up with a good solution. I also imagine the power supply could be an area that is more accessible to customisation by less experienced engineers, compared to other parts of the design. Since the desktop form factor is less restrictive I figure we might as well take advantage of that flexibility.

      Delete
  65. Stumbled on this completely by accident and am very curious. I've never listend to music on anything other than crap headphones or earbuds. If I were to build or purchase an O2, what headphones would you recommend to be able to really enjoy it? I'm thinking in the <= $130 category. If that is not enough $$ it would be nice to know that as well.

    Also, what is the expected input to the O2? Is it line level? Or do you just pass the output from, for example, an iPod right to it? If you use the analog output do you turn the volume on the iPod all of the way up?

    Sorry for the silly questions, I just had my interested peaked and have no real background in audio.

    ReplyDelete
    Replies
    1. The beauty of the O2 is it's well suited for nearly any headphone you might even remotely consider. Unlike a lot of headphone amps and sources, there's no need to avoid certain popular headphones or worry about a "good match". That said, headphones are very subjective. I would suggest spending some time at Headroom and InnerFidelity. Both have a lot of good resources and guides.

      Yes the O2 has a line level input and it also works well driven directly from an iPod, phone, etc. It has two gain modes that can, if necessary, be optimized for a portable source and a home line level source. But most are very happy with the default gain values.

      Delete
    2. My advice would be to figure out what headphones you want and then what is suitable for running that. A lot of headphones, particularly those under $130 (and many under $130 plus the cost of an O2), will be plenty fine out of an iPod. Pretty much any of these are big improvements over low-end junk headphones and earbuds, just maybe in different ways.

      The headphones generally make a whole lot more difference in the sound than whatever is doing the D/A or amplification, so don't feel like you should be spending half of the budget on an amplifier and half on headphones.

      O2 is a great investment if you want to try all sorts of (mostly expensive and over your budget) headphones in the future, since it can run most all of them properly. Otherwise, from a practical point of view, it's not exactly the most important thing.

      Delete
    3. Hey,

      I'm new to high end headphones too...I just purchased Sennheiser HD600s after reading MANY reviews. I also purchased the FIIO E17 as a USB DAC and FIIO E9 as an amplifier after reading NwAvGuy's review of the FIIO E7...WOW, the sound is incredible, I chose the HD600s over the HD650s because most reviewers said they gave a more accurate representation of the music without boosting any frequencies. I'm sure if you had the HD600s and the O2 it would be even better.

      Also, NwAvGuy THANK YOU so much for your detailed reviews, as an engineer I truly appreciated your in-depth articles when making an informed decision regarding my equipment. Like previous reviewer @Scog eloquently said: "I feel you have made a significant contribution to the public knowledge of humanity, and I look forward to following your ongoing projects for as long as you keep working on them."

      Thanks

      Delete
    4. You can take a line level out of your iPod/iPhone through the 30 pin connector which people have indicated gives a better signal than the headphone outlet.

      At least, that's my plan once I get all the bits together.

      Delete
    5. Mikeaj makes a very good point above that I've mentioned often and should have included in my first reply. If you want headphones mainly for use with a portable player, I agree with choosing headphones that perform at least reasonably well without an amp. My More Power and Headphone Output Impedance articles can help answer questions about what headphones work well with what sources.

      But it's a lot easier to just carry around say an iPod rather than an iPod and an amplifier. My Beyer DT770 Pro 80s, for example, sound fairly good driven from my iPod Touch 3G but it's on the edge of not having enough output for some kinds of music. But I still mostly use the DT770s without the O2 simply for convenience.

      Delete
    6. From the comments it sounds like in the price range I was looking at for headphones the O2 is unnecessary, but if use my total budget on headphones I would end up with a pair that would benefit from the O2.

      I guess I should have expected that. I've always encoded all of my music at very high bitrates for playback on my home stereo. Felt like I was missing out on the go, just had no idea what a good pair of headphones would be without getting taken for $$. Since so many of the reviews read like the review for a modern art painting :-(

      I guess, I'll just have to read more of the reviews and find a place to try them out.

      Second quick question, with the ODAC + 02 combination do people plan on using that to drive headphones at a PC or to drive a nice pair of speakers?

      Delete
    7. There are really wide differences in headphones so take your time and do try to hear some of them (or at least buy from a source with a hassle-free return policy). One person's "exciting" is another's "too boomy and bright".

      There's also physical comfort, noise isolation, and durability to consider. Some easily fall off your head while others feel like a head vice and crush your ears and/or give you a headache. Some block very little outside sound and let others hear the music you're listening to. Some are fairly rugged and others are fragile and don't survive long on say a daily bus commute.

      It's hard to say how people will use the ODAC. You don't need an O2 if you only want to drive speakers. The ODA is likely to have preamp outputs which means it will be optimized to drive powered speakers (or an amp, receiver, etc.) as well as headphones.

      Delete
  66. Excellent post. I want to thank you for this informative read. Keep up your great work.

    ReplyDelete

PLEASE NOTE! If you see "LOAD MORE..." OR "LOADING..." just above this paragraph, you need to click that link to load the newest comments. It may take several seconds. When all the comments are displayed the Load More/Loading link disappears.

COMMENT APPROVAL: Comments may take anywhere from a few minutes to a few days to be approved. You might want to save a copy (highlight Ctrl-C) of your comments before submitting in case something goes wrong. The approval process is necessary to prevent obscenity, etc. Please keep comments on topic. Comments promoting commercial interests will not be approved and may be reported as SPAM.