Benchmark DAC1 now available with USB

[audioengr]

Quote:

Originally Posted by EliasGwinn /img/forum/go_quote.gif Steve, What method did you use to determine the jitter attenuation? Thanks, Elias

In this case, I relied on my ears. Like I said "audible". It is not interesting to me to have a lot of measurements if they dont correlate to what I can hear. I let my ears be the final judge.

You can throw stones at me all day on this, but after all the smoke clears, it is what is audible that matters IMO.

Steve N.

 

[EliasGwinn]

Quote:

Originally Posted by audioengr /img/forum/go_quote.gif In this case, I relied on my ears. Like I said "audible". It is not interesting to me to have a lot of measurements if they dont correlate to what I can hear. I let my ears be the final judge. You can throw stones at me all day on this, but after all the smoke clears, it is what is audible that matters IMO. Steve N.

Steve,

Don't worry, I don't enjoy stone-throwing matches either. As much as I am capable, I try to limit my discussion to constructive discussion, not destructive.

I agree that no audio equipment can be judged on measurements alone. However, I do have doubts as to how reliably one can distinguish jitter performance well below -100 dBu just by listening. What source material were you listening to?

Thanks,
Elias

 

[EliasGwinn]

Quote:

Originally Posted by music_man /img/forum/go_quote.gif does the dac1 sound any different if you feed it a 24/192khz signal than if you feed it a 16/44.1. signal. i can't tell if i can really hear a difference. maybe because it just converts everything into 110khz anyways? music_man

It depends on the source material. If something was recorded at 44.1 kHz/16-bits, then sample-rate converted to 192 kHz/24-bits, then there will be no difference at all. There are no bandwidth advantages because when the audio was recorded, it was low-passed at 22 kHz. Upsampling to 192 kHz will not add any more audio information above 22 kHz (except distortion). Also, the increase in word-length will not add any resolution because there will be no new information in the newly added 8 LSB's.

On the other hand, if something was recorded at 192 kHz/24-bits then down-sampled to 44.1 kHz/16-bit, then it will sound different. This is because the bandwidth will be reduced to 22 kHz and the noise floor will increase to 16-bit levels and resolution will decrease to 16-bit levels.

I hope I answered your question. If not, please follow up.

Thanks,
Elias

 

[EliasGwinn]

Quote:

Originally Posted by Crowbar /img/forum/go_quote.gif If I remember correctly, the DAC1 uses AD1853. That has internal oversampling (called interpolation filter in the datasheet). Are you bypassing that when feeding the 110 kHz in, or does it get 4x or whatever by the DAC chip's internal filter?

The filter in the D-A chip is frequency shifted to 55 kHz to not interfere with the Nyquist frequencies (see above post concerning why the DAC1 re-samples to 110 kHz).

This, effectively, replaces the D-A filter with the SRC filter which is a much higher quality filter.

Thanks,
Elias

 

[EliasGwinn]

I would like to expound on why 192 kHz conversion is not a good idea.

The filters in a D-A chip ideally pass all audio below the Nyquist frequency and block (filter) all audio above the Nyquist frequency. In reality, however, there will be some audio below the Nyquist that is being filtered some, and there will be some audio above the Nyquist which is not filtered enough. The latter is very dangerous because those frequencies will be aliased and cause distortion.

Here's the problem with 192 kHz: the filter used for 192 kHz is of far less quality. This is true of ALL D/A chips on the market. What happens is this: the filter cut-off becomes less defined, causing audio below Nyquist to be attenuated. And, more importantly, AUDIO ABOVE NYQUIST IS NOT FULLY ATTENUATED!! The filter does not do its job as well at 192 kHz.

Another 'real-life' limitation to these filters is amplitude ripple. This means that the audio below Nyquists will have ripples in amplitude across the frequency spectrum. This is equivalent to inaccurate frequency response. This is also something that happens in all D-to-A chips, some more so then others.

The problem with 192 kHz: the ripples in amplitude become much more exaggerated when filtering 192 kHz signals. Consequently, the frequency response is much less accurate and distortion goes up.

This is why most all converter designers and recording engineers don't recommend 192 kHz. The chip technology has not provided the means to effectively convert 192 kHz without these problems.

So, the trade-off for the extra analog bandwidth is an increase in aliasing and frequency-response distortion. Although some people may 'enjoy' listening to 192 kHz more then lower rates, it is not as accurate, objectively speaking.

Thanks,
Elias

 

[LAMark]

Quote:

Originally Posted by audioengr /img/forum/go_quote.gif I suppose it is vague. I like to keep as much as possible the mods I do a trade secret. As I mentioned before, I discovered that the BUF634 is a difficult device to feed in terms of power delivery, and the topology that is recommended by TI tends to aggravate this IMO. I actually had a mod for the volume/headphone for a long time, but I wasn't happy with the level of improvement and could not recommend it to my customers. I only had a couple of orders for this. Then, I revisited it a year later (mostly as a result of the Head-Fest) and developed some improvements that are significant enough that I now recommend this mod. The improvements consist of: 1) rewiring of some traces 2) some redesign of the circuitry around the volume pot 3) replacing all of the op-amps 4) power delivery improvements for the op-amps 5) power delivery improvements for the output drivers There is barely enough space to accomplish all of this, which makes the rework difficult, and a bit expensive. I do not recommend this headphone section mod stand-alone. There is a subset of the Turbomod that should be combined with this. I demonstrated this mod at the Head-Fest in San Jose and Ray Samuels commented that it was quite good. It is not as good as a Raptor though. I want one of those. BTW, a really nice Head-fier brought his grace amp over to compare both the head-amp in it and the DAC to my modded DAC-1 about 2 weeks ago. I will see if he is willing to report back. He also brought me some awesome new music. Steve N.

I am that Head-fier. I visited Steve in central Oregon, lugging along my Grace m902 head-amp and my Senn HD650's (w/Equinox cable). Actually I came to listen to Steve's other offerings, not really thinking about doing a direct comparison between the Grace and the modified DAC-1. But there it was...so what's an audiophile to do?-we made several comparisons of course. The configuration as I recall was using Steve's laptop as the source going out through USB (what else?) to an Offramp, then through I2S to the modded DAC-1. Then we took the output of the DAC-1 to the Grace so we could compare just the head-amp sections. We also had a configuration to compare the DAC-1 DAC+Headamp to the Grace DAC+Headamp, a comparison I have made before with a stock DAC-1 and smugly concluded the Grace to be superior. In this case, though there was a decided advantage to the modified DAC-1 in both comparisons, especially in the highs where the Grace had previously excelled, or so I had thought. Basically the configurations with the DAC-1 were superior by almost any measure-staging, resolution, you name it. None of the graininess I had noticed with the stock DAC-1. Even in the head-amp comparison, where I was pretty sure the Grace would hold its own, it didn't. Oh well. Now to save my pennies. Anybody wanna buy a nice headamp?

 

[Crowbar]

Seriously guys, I keep hearing claims of listening results contradicting measurements, but these are equally invalid unless you guys do those tests with proper blind methodology.

As for the measurement people, I recommend more attention be paid to perceptually weighted metrics in order to have meaningful numbers. An excellent example is this paper:
Hollier, M.P., Hawksford, M.O., and Guard, D.R., "Error Activity and Error Entropy as a Measure of Psychoacoustic Significance in the Perceptual Domain", Vision, Image and Signal Processing, IEE Proceedings, Vol.141 No.3, June 1994, pp:203-208.
Also relevant is Hawksford, M.O., "System Measurement and Identification Using Pseudorandom Filtered Noise and Music Sequences", Journal of the AES, Vol.53 No.4, April 2005, pp. 275-296.

 

[puntloos]

Hey Elias,

First off thanks for taking the time to thoroughly answer, again, it is very much appreciated.

Quote:

Originally Posted by EliasGwinn /img/forum/go_quote.gif Puntloos, When you listed possible DAC's you might buy, you mentioned that the Bel Canto has the same UltraLock system as the DAC1. This is not true. They call their system UltraClock (which sounds strikingly similar to UltraLock...perhaps just coincidence ). More importantly, however, it is not the same circuitry.

Hah, yeah I guess my brain just connected the dots the wrong way huh, basically both you and Bel Canto are 'extremely proud of the anti jitter' and both were called 'ultra....ock' .. Additionally, I seem to remember some review or some forum post that said the two used the same technology.
Quote:

I am not familiar with the effectiveness of their jitter reduction technology, but I can assure you that the DAC1's jitter reduction technology is much more then a plug-and-play solution. Their are several design considerations that require meticulous engineering to achieve. You also mentioned that it has a better input stage then the DAC1. Can you elaborate on this a little bit?

Well, admittedly this was my conclusion after reading what basically amounts to the 'marketing version' of the specs. Bel Canto claim complete galvanic isolation of the inputs and shielded input transformers. While I think these are 'good things' in general, maybe these are also totally trivial and you simply chose to not mention this in your specs since 'who wouldn't do it that way!'.

Quote:

Also, please see my previous post concerning 192 kHz.

Done.. thank you for your explaination.. I am afraid I am out of my depth though. While my 'cowpoopie detector' is not ringing with you, I really cant gauge how your design choices would rank against the choices the belcanto, aqvox or lavry engineers made. (aqvox and belcanto sample at 192khz with the Burr Brown 1796 and 1792 DACs respectively.). It would be interesting to hear you guys discuss this subject matter, although it'd probably be like an ant watching giants fight and trying not to get stepped on.

I do have one additional question though. I am trying to arrange to get a Bel Canto, an Aqvox and a benchmark in one room. Do you have any suggestions on how, or what to test? For example is there some way you would reccommend someone with limited pro resources to create a bad jitter situation to test a dac's resilience?

 

[Crowbar]

I know these were meant for Elias but I'll chime in as well

Quote:

Originally Posted by puntloos /img/forum/go_quote.gif Bel Canto claim complete galvanic isolation of the inputs and shielded input transformers.

Input transformers are common on pro-audio DACs, and I see them mentioned in many receiver and ASRC chip datasheets. Nothing special about that. I found in my DIY that it helps if your source has a lot of noise (in my case the sound card S/PDIF out had lots of HF crap), and also can potentially prevent some ground loops. But in most cases I doubt it's necessary. You can always add an input transformer to any DAC that doesn't have one if you really feel the need, just make sure you pick the right one. scientificonversion.com make some nice S/PDIF transformers for example, but make sure you get a 1:1 and have proper termination.

Quote:

I am trying to arrange to get a Bel Canto, an Aqvox and a benchmark in one room. Do you have any suggestions on how, or what to test? For example is there some way you would reccommend someone with limited pro resources to create a bad jitter situation to test a dac's resilience?

This is a great opportunity to get a friend and do some blind testing! I couldn't recommend this more! Another great test you can do is, if you have a high quality ADC, maybe one on a good musician's sound card, I suggest you email Hawksford referencing the papers from my previous post and ask him to send you his MATLAB code. This is definitely better than the stuff people would normally do at home such as with RMAA.

 

[audioengr]

Quote:

Originally Posted by EliasGwinn /img/forum/go_quote.gif Steve, Don't worry, I don't enjoy stone-throwing matches either. As much as I am capable, I try to limit my discussion to constructive discussion, not destructive. I agree that no audio equipment can be judged on measurements alone. However, I do have doubts as to how reliably one can distinguish jitter performance well below -100 dBu just by listening. What source material were you listening to? Thanks, Elias

Real music of course. I have a collection of excellent tracks that friends/colleagues turned me onto over the years who attend CES and drop in to see me every year. It's the best of the best. Very dynamic, extended and some superb vocals, piano, percussion, pretty much everything that pushes a system to be great. It requires many different tracks to do this type of evaluation. One or two is not sufficient.

I use a Toshiba laptop with Foobar 0.8.3 and SRC upsampling it to 24/96 for all tracks. Much more detailed and dynamic this way. DAC's are driven with my Off-Ramp I2S and Pace-Car reclocker or my Off-Ramp Turbo 2 (S/PDIF coax output). Both of these USB converters are clocked with the excellent Superclock4.

Steve N.

 

[puntloos]

Quote:

Originally Posted by Crowbar /img/forum/go_quote.gif I know these were meant for Elias but I'll chime in as well

Always welcome! Hehe as long as Elias will at least say he agrees or adds his own $0.02 - no offense of course but I surmise he knows his DAC better than you

Quote:

Input transformers are common on pro-audio DACs, and I see them mentioned in many receiver and ASRC chip datasheets. Nothing special about that. I found in my DIY that it helps if your source has a lot of noise (in my case the sound card S/PDIF out had lots of HF crap), and also can potentially prevent some ground loops. But in most cases I doubt it's necessary. You can always add an input transformer to any DAC that doesn't have one if you really feel the need, just make sure you pick the right one. scientificonversion.com make some nice S/PDIF transformers for example, but make sure you get a 1:1 and have proper termination.

Well the point bel canto made was that their input stage is separated 'completely' from the output stage, more with each having their own power source etc. As mentioned, complete galvanic separation.

Your point about adding your own transformer is, i suppose, valid, especially for me since Ive actually built some circuit boards etc (with etching etc) in my studies, but Ive always been hesitant about tampering with a $2.5K piece of gear. As many modders will agree - stuff can be improved on a vanilla device, but if it has certain features from the get-go, thats definately a plus.

Quote:

This is a great opportunity to get a friend and do some blind testing! I couldn't recommend this more! Another great test you can do is, if you have a high quality ADC, maybe one on a good musician's sound card, I suggest you email Hawksford referencing the papers from my previous post and ask him to send you his MATLAB code. This is definitely better than the stuff people would normally do at home such as with RMAA.

I plan to, I have a 24/96 recording card and I have a coupla audiophile friends who plan to bring their own gear. One SACD player and one 'audiophile CD player with built-in DAC). My base set (Quad 989's with a Bel Canto Evo 4 gen2) should suffice as the fixed part of the experiment.

 

[Crowbar]

Quote:

Originally Posted by puntloos /img/forum/go_quote.gif no offense of course but I surmise he knows his DAC better than you

No offense, but your questions weren't DAC1 specific. One was about galvanic isolation, and the other about listening tests.

 

[puntloos]

Quote:

Originally Posted by Crowbar /img/forum/go_quote.gif No offense, but your questions weren't DAC1 specific. One was about galvanic isolation, and the other about listening tests.

Even less offense

but we were actually comparing the input stages of DACs.. Elias asked why I thought the DAC3 had a better input stage, I listed a few features of the DAC3 (transformers, galvanic isolation, and separately powered stages) and you said that 'many' pro dacs contain such features. While Im certain you're right, we (or at least I) still don't know if this applies to the DAC1

The open question still is if the DAC1's input stage has the features listed, and more generally is better or equal to the DAC3's .. which probably depends partially on input specs of the DAC3 none of us are privy too, but Elias might still be the best equipped to give a good guess..

 

[Bootleg]

.

 

[tonygeno]

I have hooked up the Dac1 USB to a Mac Mini via USB. The device shows up in the Sound System Preference. When listening to iTunes all is well and I can listen to the music through my stereo but when I listen to music over Safari (Rhapsody, for instance), the sound defaults to the computer speakers, even though I have chosen the Benchmark as my sound device. If I switch to the optical connection on the Benchmark, the music then plays on my stereo. Any ideas as to why?