Thanks for this comment. These are exactly the aspects I wanted to refer to yesterday, but I had to go.
Last edited:
Only thing is that those speculative comments have nothing to do with the document. Better read the document before commenting on it. It only explains what SINAD is but makes no attempt to claim SINAD is all there is.
The forum, from which the document originated treats it like the holy grail.
Why else would one publish these kinds of overviews, to which the disciples then eagerly respond?
And btw, I've read the whole d..n thing.
Last edited:
No it does not. The document reflects the attitude of ASR towards SINAD. Some here seem to have a knee-jerk reaction whenever somebody mentions ASR or SINAD.
Then it is better to comment on possible factual errors in that document. I could not find any.
DAC Distortion & Noise (SINAD) Higher Better.
It's the only metric by which DACs are ranked in a list.
It's the only metric by which DACs are ranked in a list.
Yes, higher SINAD means that noise & distortion are lower so having a ranking based on SINAD is technically justified. Ranking based on speculation or subjective opinion has no merit.
As the document states in many places sole reliance on SINAD should be avoided. E.g.
"SINAD rankings can be useful for identifying potentially flawed designs (very low SINAD) or comparing engineering prowess. However, they should not be the sole or primary factor in purchasing decisions"
As the document states in many places sole reliance on SINAD should be avoided. E.g.
"SINAD rankings can be useful for identifying potentially flawed designs (very low SINAD) or comparing engineering prowess. However, they should not be the sole or primary factor in purchasing decisions"
What is fascinating is your knee-jerk reactions and lack of arguments. Having a ranking based on well-established engineering goals is not treating it like the holy grail.
Some of us are more aligned with the views of Sean Olive and Earl Geddes (Also, see quote from Amir at the bottom of this post below):
Sean Olive:
Earl Geddes:
...Our conclusion; people are satisfied with THD and IMD. It’s like the story of the cop who asks a drunk under a street light what he is doing on his hands and knee’s. The drunk replies “I’m looking for my car keys.” The officer asks “Where did you loose them?” and the drunk replies “Over there by my car.” Baffled, the officer asks “Then why are you looking for them here?” to which the drunk replies, “Because the light is better.” Everyone knows that THD is meaningless, but it’s easy to do and “the light is better.”
...The bottom line here is that we know so little about how humans perceive the sound quality of an audio system... that one should question almost everything that we think we know about measuring it. From what we have found most of what is being done in this regard is naive. ... Maybe someday in the future we will be able to quantify perceived sound quality and move audio away from a marketing dominated situation to a data driven one.
http://www.gedlee.com/Papers/Comments on howard.pdf
"THD and IMD have no correlation to the perception of the distortion that they are intended to represent."
https://gedlee.azurewebsites.net/Papers/The Perception of Distortion.pdf
----------------------------------------------------------------------------------------------------------------------------------------------------------
Now, here is a quote from Amir:
...There does come a point where regardless of these, the device is transparent. That occurs at SINAD of around 115 dB. Anything else becomes a gray area.
https://www.audiosciencereview.com/...ter-review-updated-version.44457/post-1587673
Sean Olive:
Earl Geddes:
...Our conclusion; people are satisfied with THD and IMD. It’s like the story of the cop who asks a drunk under a street light what he is doing on his hands and knee’s. The drunk replies “I’m looking for my car keys.” The officer asks “Where did you loose them?” and the drunk replies “Over there by my car.” Baffled, the officer asks “Then why are you looking for them here?” to which the drunk replies, “Because the light is better.” Everyone knows that THD is meaningless, but it’s easy to do and “the light is better.”
...The bottom line here is that we know so little about how humans perceive the sound quality of an audio system... that one should question almost everything that we think we know about measuring it. From what we have found most of what is being done in this regard is naive. ... Maybe someday in the future we will be able to quantify perceived sound quality and move audio away from a marketing dominated situation to a data driven one.
http://www.gedlee.com/Papers/Comments on howard.pdf
"THD and IMD have no correlation to the perception of the distortion that they are intended to represent."
https://gedlee.azurewebsites.net/Papers/The Perception of Distortion.pdf
----------------------------------------------------------------------------------------------------------------------------------------------------------
Now, here is a quote from Amir:
...There does come a point where regardless of these, the device is transparent. That occurs at SINAD of around 115 dB. Anything else becomes a gray area.
https://www.audiosciencereview.com/...ter-review-updated-version.44457/post-1587673
Last edited:
The discussion was about a document in post #590 which does not make any of the claims in your post. Making negative comments about a document without even reading it is just trolling.
It should be obvious to most people that Amir's sacred text on SINAD exists in large part to help create an appearance of science when he makes unscientific cult claims about "transparency" being solely a function of SINAD.
We all know what the SINAD rankings are really about, and it isn't real science; its cargo cult science.
Regarding the discussion in this thread, I can only hope that people here in the thread looking for an excellent dac at low cost are not being fooled into believing in cult claims.
We all know what the SINAD rankings are really about, and it isn't real science; its cargo cult science.
Regarding the discussion in this thread, I can only hope that people here in the thread looking for an excellent dac at low cost are not being fooled into believing in cult claims.
Last edited:
There are many people that don't. It is based on noise and harmonic distortion. In my opinion, the noise is the more important factor.
There is anectodal evidence that having a very low SINAD does indeed improve music presentation. See some of the reviews of the Topping LA90 amplifier, e.g., https://www.theabsolutesound.com/articles/topping-la90-power-amplifier/
Terry,
Its necessary to go back and consider that SINAD was developed and mostly researched in the days of analog amplifiers, which are mostly only weakly nonlinear. Fourier Transforms are considered reasonable approximations for analysis of actual behavior of such systems. However, even for such systems, Sean Olive and Earl Geddes make valid points. How humans tend to interpret FFTs remains problematic.
Later with the development of digital audio, particularly sigma-delta technology, and with the development of class-D amplifiers, now we are dealing with highly nonlinear systems. Things like correlated noise which are not necessarily reflected in SINAD measurements may make a device look good in SINAD measurements yet sound awful in reality. What it means is that when SINAD is good enough, the the biggest remaining problems are from things other than SINAD.
In addition, SINAD measurements were mostly studied and developed outside the context of stereo reproduction imaging accuracy. What SINAD measures is what ESS called PSS (periodic steady state) distortion, and stationary noise. Again, that's not all there is in strongly nonlinear systems. And its not all that's needed to assure good stereo reproduction either.
Mark
Its necessary to go back and consider that SINAD was developed and mostly researched in the days of analog amplifiers, which are mostly only weakly nonlinear. Fourier Transforms are considered reasonable approximations for analysis of actual behavior of such systems. However, even for such systems, Sean Olive and Earl Geddes make valid points. How humans tend to interpret FFTs remains problematic.
Later with the development of digital audio, particularly sigma-delta technology, and with the development of class-D amplifiers, now we are dealing with highly nonlinear systems. Things like correlated noise which are not necessarily reflected in SINAD measurements may make a device look good in SINAD measurements yet sound awful in reality. What it means is that when SINAD is good enough, the the biggest remaining problems are from things other than SINAD.
In addition, SINAD measurements were mostly studied and developed outside the context of stereo reproduction imaging accuracy. What SINAD measures is what ESS called PSS (periodic steady state) distortion, and stationary noise. Again, that's not all there is in strongly nonlinear systems. And its not all that's needed to assure good stereo reproduction either.
Mark
Last edited:
The only remaining "problem" in audio is recording technique, the loudspeaker driver and a loudspeaker system that go along with the recording. This not to say that every dac and amp is perfect.
//
//
People looking for a low cost Chinese DAC have no chance of finding any useful information on this thread. There seems to be a tacit agreement on this forum NOT to discuss Chinese boards available e.g. on AliExpress, despite there being several that could be usefully evaluated. All discussion of such boards is being avoided in principle and in practice.
As JeanPaul said in post #80 a long way back
"There are 0 real DIY audio players. What is available are PCBs. No one objects. Practically everyone buys one ready made and possibly a diyaudio.com external DIY project DAC or ChiFi DAC is added. That is what digital audio is to the majority. Of that majority again a majority is not able to master the complexity to design and build a digital source or a DAC. It is that simple. It is a non-discussion really. Build what you can and want to build and otherwise just don’t."
The ChiFi boards are useful for many because they don't have complex surface mount equipment and skills. The DIY bit can be adding power supplies cases etc. We could have evaluated several such boards in 636 posts but this hasn't happened. I completely understand that those using superior boards by Miro and others are sitting out this discussion, but surely somebody has something to say that isn't completely off topic?
But what is there to say? Unless you are building simply for the sake of building, it is much cheaper to buy a finished product.
I spent a lot of time trying to see if there was a practical way to make Chinese boards into good dacs. There is a very long thread on ES9038Q2M where I did the most work. This was not long after such boards first started becoming available. The hope many people had was that there would be some easy things that could be improved to make the dacs sound better, more like the ESS datasheet specs would seem to imply the chips should be able to do. It turned out that the board designers knew exactly what they were doing. They didn't make any stupid mistakes in the design that would make it easy to replace a few parts and have a much better dac. What they did do was make sure that everything about the boards with the exception of the ESS chip was a cheap as possible. The dac board itself was only two layers, the output stage was voltage mode, it was all junk. I extensively modified one in pretty much every possible way except to replace the whole PCB with a new one of my own design. After a whole lot of work, I learned a lot about those dacs, and had made the cheap Chinese dac sound maybe slightly better than my Benchmark DAC3 in one way, but not in all ways.
Later I did a lot of work with dacs from Andrea Mori's, The Well Audio. Andrea would send me prototypes that I had to pay for if I didn't want to return them, apparently a problem with Italian law. So, I paid for everything. I did a lot work on Andrea's DSD dacs and some work with the R2R dac. I reported back to Andrea with what I found. Sometimes what I found was new to him, and in a few cases he already knew there was a problem. The last big set of problems I found had to do with nonlinear caps and with ferrites on his boards. Some of that is briefly described in my Clock Board thread. I have helped some people fix those problems in their own Andrea dacs via PM. Thus, I can't quite recommend those boards either. (Very sadly, Andrea's long term illness may prevent any more progress with that line of board products. My best wishes to him.)
Also, sometimes I help people with their Iancanada stacks. I did some work in that area some time ago which exposed problems with radiated EMI/RFI coupling. Today I like some of that stuff, but not all of it.
Also did work with a fair amount of Twisted Pear circuitry. Another set of problems to fix, IME and IMHO.
Anyway, what I tell people today is that if you want a low cost dac that is good for the money, buy a low cost Topping or similar that fits your budget. In the end it will cost less than modifying a Chinese dac to be as good as its possible to make it to be. The only reason I tell people to consider buying a low cost Chinese dac then trying to fix it up to sound better is if you want to learn about dacs. However, it turns out that isn't what most people really want. What most people want is a bargain at low cost. Cheapest way to get there is with the best Topping you can find. If you don't like the sound of those, then there are some low cost dac projects here in the forum from Miro and from Abraxalito.
What I also tell people is that if you want a really good dac, one better than you can buy from Topping, Benchmark, or many other companies, then I can tell you how to do that too but its a whole lot of SMD soldering involved and its not all that low cost. Plus, I'm still not satisfied with output stages available to most people. That part is still a work in progress. For myself I am lucky to have an output stage that is proprietary and not of my own design but its not something I can share.
Last edited:
"I spent a lot of time trying to see if there was a practical way to make Chinese boards into good dacs. There is a very long thread on ES9038Q2M where I did the most work."
I'm sorry to hear about your experiences with that unit. But it's just one out of a huge number of available Chinese boards. As for Topping, I had one of their well reviewed DACs for a very, very brief time before selling it on. I listen to classical, opera and jazz and as a professional musician I prioritise the tonality of familiar acoustic instruments above all else. The Topping was awful, just one model in the range but so far away from the tonality I expect of a good quality piece of audio that I closed the door to the range. Maybe some SD DACs are musically satisfying, but I joined the ranks of the multibit constructors and went back to the classic chips like PCM56, AD1865 etc. just like Miro and his designs and many others. Here I did find musically satisfying sounds with excellent tonality on acoustic instruments.
As I noted, I bought two Chinese PCM56 boards and modified them and they both work perfectly with no issues. They also sound significantly better than the European mid priced DACs I had before. The parts they contain (which many here complain of) must be doing their job properly. So yes, I'm a believer and I was hoping to find an evaluation of similar Chinese boards on this thread. If you search on "DACs" on AliExpress there are a large number to choose from. So there is plenty of opportunity to turn this thread into something useful.
I'm sorry to hear about your experiences with that unit. But it's just one out of a huge number of available Chinese boards. As for Topping, I had one of their well reviewed DACs for a very, very brief time before selling it on. I listen to classical, opera and jazz and as a professional musician I prioritise the tonality of familiar acoustic instruments above all else. The Topping was awful, just one model in the range but so far away from the tonality I expect of a good quality piece of audio that I closed the door to the range. Maybe some SD DACs are musically satisfying, but I joined the ranks of the multibit constructors and went back to the classic chips like PCM56, AD1865 etc. just like Miro and his designs and many others. Here I did find musically satisfying sounds with excellent tonality on acoustic instruments.
As I noted, I bought two Chinese PCM56 boards and modified them and they both work perfectly with no issues. They also sound significantly better than the European mid priced DACs I had before. The parts they contain (which many here complain of) must be doing their job properly. So yes, I'm a believer and I was hoping to find an evaluation of similar Chinese boards on this thread. If you search on "DACs" on AliExpress there are a large number to choose from. So there is plenty of opportunity to turn this thread into something useful.