Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › What is the best DIY DAC up to $500
New Posts  All Forums:Forum Nav:

What is the best DIY DAC up to $500 - Page 4

post #46 of 51
Originally Posted by wakibaki View Post
If you want to be viewed as a reputable designer, you will adopt the same standards as those adopted by the the DAC manufacturers.




Oh, and you can't apply A-weighting at a single frequency. Didn't you know that?



The appropriate SNR is stated just below the 139dB.  It's not recognized as such because it doesn't have the SNR nomenclature.


Your post has merit, but it seriously lacks tact.


AMB lists both the full spectrum, and the 1KHz.  All the information is there in the screen capture in the posts above.  I don't understand this nitpick, and the force by which the nitpick is presented.  The standard measurement is noted there, and the odd one.  The SNR @ 1KHz is more similar to dynamic range @ 1KHz, but not quite.


Yes, you can specify A-weighting to a single frequency.  The A-weighting curve are dB adjustments at different frequencies.  @ 1KHz the adjustment is 0.  Measuring a 1KHz tone would be the same, regardless of weighting as all weighting scales that I know of apply 0dB correction at 1KHz.  dbA, dbB, dbC are all 0 dB adjustments at 1KHz.  It makes stating dBA for 1KHz pointless, but it's not incorrect.


My suggestion, make a DNR measurement and rename it.  Also add SNR to "noise level" so that people's eyes can "pop" to it.  Some folk see SNR, not read the qualifications of it, and go bat **** crazy.  Fully qualified SNR is better than some unqualified SNR that leaves one wondering what was actually measured.

post #47 of 51
Originally Posted by mvrk10256 View Post

That wolfson chip is the same unit the grub/pup dacs use. Maybe a $100 pupdac would be equal to the gamma. BTW Wakibaki is right, however I am not sure why they claim a higher dB. Might be a typo, or a mistake. 

Hate to say it, but this is typical of some of the discussion in this thread - lots of BS and few facts.  The Wolfson DAC chip is not even the chip that the GrubDAC uses, much less the PupDAC, which uses a TI chip - not a Wolfson at all.


The GrubDAC uses the WM8524.  It possesses a charge pump that creates a reference ground at VDC=0, which means no output coupling capacitors.  It's totally different than the Gamma's WM8741.  The pupDAC uses the PCM1794, probably the finest chip that Burr-Brown makes (owned by TI).  It is definitely a balanced DAC and the pupDAC uses an opamp convertor as the output - the OPA2836, a very capable chip with sterling specs despite its very low voltage.

post #48 of 51
Originally Posted by mvrk10256 View Post

Thanks for posting that W. I was wondering when I read the response from AMB... most people can hear 20Hz to 20kHz, what is the purpose of the 1kHz value? why is it more important/valuable than anywhere else in the spectrum? Is it because of the log scale? Or did you simply find the frequency with the best response and site that?
The 20Hz-20KHz aggregate number (-106.x dB) is listed under the graph.

The -139 dB number is just another data point, it's the ratio between a 1KHz -1dBFS sine wave signal and the noise floor. Many measurements are done with a 1KHz reference, such as THD. There is nothing "special" about it with respect to human hearing. It's just a convenient middle-frequency useful for a variety of purposes. The fact that this number is a S/N ratio, it can be called that, with the test condition listed. It is what it is, and is true. If you look at the graph, you'll see that the noise floor is exceptionally low all across the spectrum, not just at 1KHz, but I must emphasize that all these measurements are limited by the equipment used. The actual performance is probably better.\

Wakibaki, you can nitpick all you want, but there is no intention to mislead. The fact remains that the γ2 has superb performance.
post #49 of 51

OK, sorry if my last post wasn't as considered as it might have been, but I'd just been administered my day 8, 2nd. cycle adjuvant chemotherapy, and I was feeling a bit poisoned.


Look at this intersil appnote:- http://www.intersil.com/content/dam/Intersil/documents/an51/an519.pdf.


It shows how noise is always measured in a bandwidth. A spot noise is measured in a 1Hz bandwidth. There are various sources of noise in a circuit, but the most significant and commonly considered is thermal noise.


When the measurement bandwidth is reduced the RMS noise power (or voltage, or current) gets smaller and smaller. Recall differentiation. When the bandwidth is zero, the noise power is zero. 


The bandwidth of a pure sine wave is zero. Therefore the signal to noise ratio of a pure sine wave is infinite if the noise integration is performed over a zero bandwidth.


RMAA does not state the integration bandwidth over which the points for the trace are computed.


It is neither meaningful or useful to compare the -1dBFS sinewave power with a noise power obtained by integration over an unknown small bandwidth.


The A- and other weighting curves are meaningful only when applied in relation to 1kHz. As noted by holland, the weighting at 1kHz is 0dB. The curve represents a correction factor to be applied to measurements at other frequencies with reference to a perceived loudness or measured volume at 1kHz, and therefore implies a perceived or measured level at 2 or more frequencies, 1kHz, and 1 or more others. It's meaningless to try to apply it at 1kHz, and it's true application is when performing an integration over the audio bandwidth.



Edited by wakibaki - 8/31/13 at 12:53pm
post #50 of 51
Wakibaki, sorry to hear about your health woes. Whatever it is, I hope you will feel better soon.

You are correct about a sine wave theoretically having "zero bandwidth", but the noise doesn't. If we're interested in the distance between the signal amplitude and the noise floor, then it makes sense to look at noise level at the same frequency as the signal. As you can see in the noise spectrum graph, the noise floor is basically flat across the spectrum with a little flare-up toward lower frequencies. If I use a noise bandwidth of 1Hz (or some other arbitrarily small but non-zero bandwidth), centered at 1KHz, I'd get the 139dB figure, which is all that number is (the distance between the signal and the noise at 1KHz). Nothing more and nothing less. Perhaps this is not a common parameter found in spec sheets, but it's a textual representation of what's on the graph at 1KHz. It's just another data point, and the conditions under which it is derived is clearly specified. And as I said repeatedly, the conventional S/N ratio is also shown (albeit all of these are limited by the test equipment).

Whether any of these numbers is "meaningful" or not is a matter of opinion. I think the distance between the signal and the noise floor is something interesting, obviously you don't.

You know, if I didn't list that number, this discussion would never had happened. If anything, I think I err on the side of providing too much information. Most other commercial audio gear just list a single S/N ratio number (if at all), with no qualifications about how the number was derived. What's the signal level used and what bandwidth was it integrated over? This makes S/N ratio (and for that matter, almost any spec) meaningless to compare between brands because their testing parameters (not to mention methodology, equipment, etc.) are different. There were various organizations that tried to standardize these things but I am not sure how successful the effort has been in the real world.

Oh and yes, I am aware that applying A-weighting at 1KHz does nothing (the correction factor there is 0dB). But I list A-weighting to be consistent with the other noise figures found in the test results. It doesn't change the number anyway.

At any rate, if enough of you think it's better to delete information, then I'd be happy to remove that item from the specs. Don't pollute this thread, send me PMs.

By the way, I now own an Audio Precision System Two 2322 Dual Domain audio analyzer with many option filters to do future testing with. I don't have time to re-do all previous tests with the new equipment (and in some cases I no longer have the gear I tested), but look for future projects to include test results measured on the new setup.
Edited by amb - 9/2/13 at 2:12am
post #51 of 51
Originally Posted by amb View Post

Wakibaki, sorry to hear about your health woes. Whatever it is, I hope you will feel better soon.

By the way, I now own an Audio Precision System Two 2322 Dual Domain audio analyzer with many option filters to do future testing with. I don't have time to re-do all previous tests with the new equipment (and in some cases I no longer have the gear I tested), but look for future projects to include test results measured on the new setup.


Thanks for your good wishes regarding my health. Whatever my problems, I try to keep my day-to-day difficulties from impinging on my posts here, but sometimes, I suspect, without complete success. Just a couple of days ago I mentioned to my wife that I had started to experience some quite severe pain over the last half-hour, only to have her tell me that she wondered what the problem was, as I had become very snappy with her. This was at the end of a very good day and the last person I want to hurt is my wife, without whose support I would have found it difficult to survive this long.


I am in a much better position than many other lung cancer sufferers, because despite having had 2 different types of tumour removed over 3 years, I am still told that I have a good chance of a complete recovery.


I am delighted to hear that you have acquired an Audio Precision test set, this is the best reply to any criticism levelled at designers for the DIY market. Such a development can only raise the standards throughout. My only problem is in making my own testgear competetive.


I believe that you took a hit to your reputation in a series of disagreements with another advocate of objective test, I thought at the time that this was very unfortunate; one, because it is understandable that anyone under attack should become defensive and perhaps react with a degree of hostility rather than pure rationality, and two, because of all the DIY designers extant at the time, you were probably the most committed to testing and providing test results for your products. The fact that the test equipment, methodology and software being used probably came in for some deserved criticism was something that many designers had failed to take note of, and there is some excuse for us when we simply follow the practices of our peers.


I look forward to seeing the results of your design work, now supported with a professional standard of test gear.


Best wishes,



New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › What is the best DIY DAC up to $500