Why do modern DACs over $1,000 sound different?

Steve,

I used there USB to I2s bridge and bought its independent power supply.

http://www.ciaudio.com/products/VDC5MKII
http://www.ciaudio.com/products/TransientMKII

"Am I right to read Steve's response as suggesting that I'd have to put a device between the changer and the DAC that would reduce the jitter? If so, what's that device?"

Synchro-Mesh resampler/reclocker.

Steve N.
Empirical Audio

Steve N is right, but I want to share and support him in stating that the use of the W4S I2S input is a much better than the USB or bridge to S/PDIF.

I have enjoyed the W4S very much, but still found it lacking vs high-end vinyl until I put a CIA USB to I2S bridge between the Mac Mini and the W4S.

Holly Moley, 2D to 3D and all of a sudden I am listening to digital files one after the other. I bought the associated CIA power supply for the bridge and this eliminated the USB noise, solidified the 3D illusion (depth not only width), and I think improved the bass and timbre in some instances. I used a Harmonic Technology 1.4 HDMI cable. Also, the USB to the CIA is the W4S USB cable which was better than my previous Kimber USB cable.

Bottomline. I am grateful to Steve Stone for mentioning this I2S bridge approach for the W4S DAC-2 in a recent Absolute Sound. Fantastic.

A follow up question (or two) about jitter and how to reduce it.

After reading this thread, I have a question. I have a Sony cd changer that I use because of the convenience. It has an optical output. I am thinking about trying a DAC to learn if I can hear an improvement in. Am I right to read Steve's response as suggesting that I'd have to put a device between the changer and the DAC that would reduce the jitter? If so, what's that device? Or would I just be better off buying a cd player when I want to do some critical listening? TIA

Jim

Steve nailed it.

I have done experiments with low jitter sources (an Off-Ramp and Audiophellio). They are both below 10ps, which is very very low, so low you need specialist equipment to measure it, and they sound different.

I have done the experiment with the WFS DAC (as well as others). With the WFS the USB sounds dreadful compared to either the Audiophellio or Off-Ramp with the Off-Ramp sounding the best. In fact the Off-Ramp has slightly worse jitter than the Audiophilleo - so why does it sound better? It has a better spectrum.

In my experience the biggest determinant of sound differences is the jitter of the source. The SABER everyone touts as jitter immune simply isn't - simple as that. This does not mean DAC chip output stage etc do not make a difference - they do - but jitter seems to be the main factor.

For example SABER based DAC's are not my favorite but when fed with a low jitter source I can easily live with them - otherwise I couldn't.

Thanks
Bill

Steve nailed it.

I have done experiments with low jitter sources (an Off-Ramp and Audiophellio). They are both below 10ps, which is very very low, so low you need specialist equipment to measure it, and they sound different.

I have done the experiment with the WFS DAC (as well as others). With the WFS the USB sounds dreadful compared to either the Audiophellio or Off-Ramp with the Off-Ramp sounding the best. In fact the Off-Ramp has slightly worse jitter than the Audiophilleo - so why does it sound better? It has a better spectrum.

In my experience the biggest determinant of sound differences is the jitter of the source. The SABER everyone touts as jitter immune simply isn't - simple as that. This does not mean DAC chip output stage etc do not make a difference - they do - but jitter seems to be the main factor.

For example SABER based DAC's are not my favorite but when fed with a low jitter source I can easily live with them - otherwise I couldn't.

Thanks
Bill

You should ask this over at AVSForum ;)

The biggest difference in most DACs is the input interfaces, particularly USB. The design of these varies all over the map as does the quality of the master oscillators, which equates to jitter.

For instance, I recently heard a W4S and drove it with Off-Ramp5 to S/PDIF and I2S inputs. The I2S input is worlds above the other inputs on this DAC, making it a world-class contender. Depending on how people use the DAC, they will report entirely different experiences:

http://www.audiocircle.com/index.php?topic=112012.0

It is critical that the digital source driving the DAC have very low jitter in order to determine the real performance. A high-jitter source, such as the typical CD player will make many DACs sound a lot worse than they really are.

A low-jitter source will often make a $1000 DAC sound identical to a $8K DAC.

The next thing that sets DACs apart is the power subsystem and op-amps versus discrete transistors.

These things result in differences in harshness, smoothness and dynamics. Difficult to listen to most DACs with a lot of op-amps without fatigue IME. The smoothness of discrete stages is usually apparent.

These things if executed well can result in a nice analog sound. This is not an easy thing to accomplish however, so many DACs actually have compromised power systems that mask detail to achieve a softer sound. This is the wrong way to accomplish this IMO.

The final thing that makes DACs sound different is whether or not they use reclocking and upsampling to reduce jitter or not. Those that use ASRC upsampling will always suffer from some coloration from the upsampling IME. Those that have no reclocking or upsampling will sound worse with a high-jitter source driving them, but also better with a low-jitter source driving them. This is why I prefer DACs with no upsampling or reclocking in them. Gives you more control over the performance.

Steve N.
Empirical Audio

DACs are no different than any other audio compoent in that regard.Every thing affects the sound. Current/voltage conversion options.Circuit,op-amps(which one?) or instead choose discrete analog circuits Each part(capacitor,resistor,wire,insulation,chassis construction-material(which metal, How thick?)how about grounding solutions.Power supply,transformer choices.... you get the idea,it all matters. With so many variables how could any of them sound the same?
Regards,