DAC related "sound signature"

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have always wondered whether the choice of one DAC over another has a significant impact on the sound character of a music player. To make this more useful for the less technical educated (like me) let’s discuss companies and not specific chips. So... choosing a player or DAC sporting Burr Brown vs a Wolfson or a Cyrus chip means that it will tend to sound in a particular way, or the “character” is linked to the rest of the design? I try to learn here and make more educated decisions in the future. If you think you can help, I would really appreciate the input!!!
 
I have always wondered whether the choice of one DAC over another has a significant impact on the sound character of a music player. To make this more useful for the less technical educated (like me) let’s discuss companies and not specific chips. So... choosing a player or DAC sporting Burr Brown vs a Wolfson or a Cyrus chip means that it will tend to sound in a particular way, or the “character” is linked to the rest of the design? I try to learn here and make more educated decisions in the future. If you think you can help, I would really appreciate the input!!!

My experience has been that there are just too many implementations variables to simply and reliably say this or that company, or this or that chip, has a consistant sound signature, in and of itself, that always comes through. I know, many do swear by the sound signature of the TDA1541, and Audio Note feels that the AD1865 has the best sound. Others feel that there is a distinctive and common (unpleasant) character to the sound of all sigma-delta type converters.

All of which indicates to me that maybe half of the resultant sound of any DAC is due to the surrounding implementation variables. Such variables include, the i/v circuit, the output amplifier/buffer, the power supply regulation, jitter suppression or lack thereof, even the critical selection of key passive components. As always, in my experience, the only way to reliably select a musically satisfying component or DIY circuit is to audition it within the context of the rest of your system.
 
Last edited:
Thank you for the reply Ken. I have met people swearing by Wolfson and others calling it "too digital". Others say the TDA1541A is the best chip ever. I do have an Arcam Alpha 5 sporting two of those and it is easily beaten by the humble Beresford Caiman that uses a low end Wolfson chip. Personally I THINK I like the sound of Wolfson equiped players, but I have not heard every player out there... That did not keep me from ordering a Burr Brown equiped player a couple of days ago though. I really wonder if anyone has performed some dependable tests regarding that aspect... but even if someone put all the effort needed, the resulds would be irrelevant some months later since DAC chips keep on improving.
 
I have a DAC that takes modules for testing various chips. ecp audio :: Walnut DAC

I, too, like the Wolfson chips a lot. They sound great, and are very easy to work with. While they sound better than most I out chips with either passive or OpAmp IVs, I do think a properly implemented IV with most current out chips sounds (and measures) better. Fwiw, of the ones I have tried, I find BB chips a little darker with an emphasis on bass, and Analog chips a little brighter. The Wolfsons seem to emphasize a little more midrange. But, this is from a pretty limited sampling so may not be representative.
 
I have a DAC that takes modules for testing various chips. ecp audio :: Walnut DAC

I, too, like the Wolfson chips a lot. They sound great, and are very easy to work with. While they sound better than most I out chips with either passive or OpAmp IVs, I do think a properly implemented IV with most current out chips sounds (and measures) better. Fwiw, of the ones I have tried, I find BB chips a little darker with an emphasis on bass, and Analog chips a little brighter. The Wolfsons seem to emphasize a little more midrange. But, this is from a pretty limited sampling so may not be representative.

This is very interesting!!!

Thanx for sharing your findings!

:cheers:
 
Thinking about it, it would be quite reasonable for a chip maker to somehow "mark" its product line with a discrete sound signature. Otherwise their sales would only rely on absolute technical features - pointless in this market - or price, which is not wise. Customers have to chose you for something.

Another thing that makes sense is that there is a vague consistency in reviews I' ve read over the last weeks concerning cd players with BB DACs. They tend to be described as portraying an "anologue-ish" sound. It must be what you refer to as "darker".
 
That is not necesarly true (that the manufacturers to mark their cips"). In order to "add a signature" their chips would have to be not linear, to depart from the ideal representation of the audio. NONE of them do that, they all keep the resolution, linearity, distortion as tight as possible.
The difference could be made by:
1. The OpAmps at the I/V stage (or internal OpAmps for the ones with voltage output). Using faster and low-noise/low-distortion ones aleviate the differences here.
2. Power supply. Rolling OpAmps and DACs in various DIY settings require proper decoupling of the power rails (to ground). Some chips are more "forgiving" in this aspect and lousy made PCB's can fav the cips that require less decopling (slower OpAmps for example). The sound will "change" on those bad-desinged PCB's - but only in worse compared with a proper designed PCB.
3. Source jitter. Some DAC's are more or less sensitive to that and in the DIY the sources are not always the best quality. For example the Sabre DAC's are better on handling the jitter, so a crappy digital signal will sound better on them then on other DAC's. But that doesn't mean nothing for somebody that has a decent transport or a de-jitter ASRC/DSP in the digital path.
4. OS/digital filter and interpolation stage. Here there is a slight difference between manufacturers and always a dedicated filter DSP can achieve better results. "Feeding" the DAC with a signal that was either externally up-sampled and interpolated OR higher native samplerate will minimize the difference. At 88.2/96kHz samplerate or from DSD, the differences made by internal filters are practically zero.

Of course there will be difference between lower cost DAC's and higher end ones. But that is not because a "special signature", it is because the DAC performance itself. A cheap DAC cip will sound worse than a high-end one, no matter who makes it.

I see a lot of "experiments" on DIY world and not all of them are technically sound. For example using passive I/V conversion will give very different results on different DAC's (due to internal sensitivity to value of output load) - but all of the pasiive will sound worse than a good OpAmp. Using the recomended OpAmp will minimize that variation. But a lot of DIY botch the critical decoupling of the fast OpAmps and then they say that "OpAmps sound bad".
I saw examples of diodes (made from transitors) series with rails, capacitor between rails instead rails-ground, low-ESR electrolitic capacitors with long traces...
Swapping OpAmps in a lousy designed PCB will lead to wild variations in sound. But they all will sound bad (in different degrees) compared with what is possible on a good PCB.
 
Last edited:
There are too many ways to implement a DAC chip within an audio system. I don't think vendor is important. A 1-bit DAC will always sound different compared to multi-bit ones regardless who the vendor is, for example.

Important "signature" of a DAC can be seen easily from its specification sheet. One critical/important specification imo is the DNR. Whatever the brand, and whatever you do to perfect the power supply, a 100dB (of DNR) DAC will NEVER sound like a 120dB-DNR DAC.

Most Wolfson DACs are expensive (but not all!). Their higher performance DACs have at least 120dB DNR. My "reference" DAC has only 114dB (CS4362), but I think it is sufficient :cheers:
 
Then again, the very same DAC chip on the very same implementation can sound (and measure) quite different just by tampering with it's internal filter settings.

Now put into the picture DAC chips from different manufacturers (who often follow different means to the same end and who don't necessarily have the same know-how - some, for example, excel at apodizing filter design), different generations and types of chips from the same manufacturer and, last but not least, different principles of operation (Delta-Sigma, R-2R ladder etc.) and you get the idea....
 
Sonic wrote:
"For example using passive I/V conversion will give very different results on different DAC's (due to internal sensitivity to value of output load) - but all of the pasiive will sound worse than a good OpAmp. Using the recomended OpAmp will minimize that variation."

Hogwash. Central, don't you believe a word of that. Sonic is taking what is his own listening preference and making blanket claims about what will be the preference of others. All passive (meaning, a simple resistor from the current DAC output pin to ground) i/v implementations may sound worse than a "good" op-amp to sonic, but they may not all sound worse to you. In fact, you might find that you have a great preference for passive i/v, depending on the specific implementation. Not all, or even most, current output converter (DAC) chips are suitable for passive i/v, but some most certainly are. In fact, some of the most acclaimed commercial conversion stages utilize passive i/v, such as those from Audio Note, Ypsilon (I believe), BAT, and others.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.