Actually, that would be very unsatisfactory for me. I know I am biased and hopeful of good experimental outcomes (although I continue to work on countering biases). I am so fortunate to have some skilled listeners who will give me their honest opinions: "something about this doesn't sound right," "its not as realistic as you had it before," "its harder to listen to now," "The space isn't as good as it was." Its not that I take those comments as factual, but they do get me checking things. In the end everyone benefits from honest feedback. One problem I have is that one skilled listener really doesn't care about my hobby. I have to coax him to find a time to listen. He has his own hobby too, which is computer graphics. He has to coax me to look at his graphics work. We criticize and or compliment each other while describing observations (say, colors, textures, light reflection effects, lighting illumination effects, shadow contrast (or maybe gamma), etc. for graphics), all the while neither of us careing much about the other person's hobby. We have other things in common, however, but just not hobbies.
Last edited:
If you would want to hear my opinion on this... I am in favor of the middle ground on this topic. Listen tests and experiences are important (as is overall enjoyment of listening to music) and measurements add data to improve designs or compare devices. I would expect at least decent standard measurements from good audio equipment, but not only that. And ideally, new measurement methods would be invented on clear observations that cannot be explained by current measurements...
Last edited:
One of the problems with new measurements is that some of them get harder and more expensive to do. Regarding new measurements that could be run on AP hardware, it kind of looks like AP customers are not asking for more measurements, at least not new measurements that very many customers would be willing to pay extra money for. OTOH, AP's problem is that their machines are already around SOA as audio spectrum analyzers are concerned. Like any company, AP needs a continuing flow of money from customers. They need to find things customers are willing to pay more money for, or else they need to change the remuneration model of the system to subscription-based. If you can't buy AP software outright, only subscribe to it, then that solves their problem of having a reliable income stream. IOW, AP is not solely motivated by offering more and more measurements that can be done using their hardware.
Not even the freeware REW is adding every potential measurement they could. For example, they could be doing volume envelope analysis using Hilbert transforms, but the they appear to have no interest in doing that. Someone else would need to lead the way first. If new soundcard-based measurements were to take off and prove useful then more software designers might start to take interest. As any product, free or not, becomes mature enough, there is likely going to be some inertia to leading the way like a small startup might be able to do.
Not even the freeware REW is adding every potential measurement they could. For example, they could be doing volume envelope analysis using Hilbert transforms, but the they appear to have no interest in doing that. Someone else would need to lead the way first. If new soundcard-based measurements were to take off and prove useful then more software designers might start to take interest. As any product, free or not, becomes mature enough, there is likely going to be some inertia to leading the way like a small startup might be able to do.
Last edited:
The measurements taken are superficial (continuous signal linearity).🙄
The biggest differences are in intermodulation, audio is 100% intermodulation, unless you are a fan of continuous 1kHz signals.
The biggest differences are in intermodulation, audio is 100% intermodulation, unless you are a fan of continuous 1kHz signals.
There are also multi-tone test signals that can be used for IMD. That's not really so much the problem. Its that such test signals are what ESS described as PSS (Periodic Steady State). ESS went on to try to explain that there are some effects in dacs that can be audible but which are not PSS. They said that PSS measurements will not show the problems. However, they did not provide what some people would consider to be "proof" of the claims. Also, ESS probably did not describe every effect they know about. More likely they only talked about effects where they think they have a competitive advantage. Personally, I sure would like to know what they kept secret.
ESS is the measure of DAC knowledge? I thought it was a commercial company producing and selling ICs.
OK, listening was not good enough. Measurements are not good enough. Typical buyers of DACs don’t know jack of DACs or how to judge them. Manufacturers of chips don’t know their stuff well enough. Gear producing companies that use the chips don’t know how to implement stuff right. No it are a few hobbyists at diyaudio.com that know all there is to know.
OK, where is the product?
OK, listening was not good enough. Measurements are not good enough. Typical buyers of DACs don’t know jack of DACs or how to judge them. Manufacturers of chips don’t know their stuff well enough. Gear producing companies that use the chips don’t know how to implement stuff right. No it are a few hobbyists at diyaudio.com that know all there is to know.
OK, where is the product?
Last edited:
Some of the things they talked about that one time are IMHO pretty much like some things @MarcelvdG has talked about (and which he has sometimes explained in more detail). However Marcel doesn't talk much about audibility of such things. But the fact that such things exist might go a long way to explaining audible differences between dacs that measure well in terms of standard AP measurements.
I would also add that because ESS is a seller of dac chips, some people have rejected or dismissed what they said without even seriously thinking about what was said. If so, that might go a long way to teaching ESS never to try telling some more of the truth than they usually do since then. Better for them to go along with cheap chi-fi dacs that measure well. More chips sold that way than by telling the truth about it.
Last edited:
This paper is interesting, as it indicates the median hearing detection threshold for test participants was down well over 100dB @ 24kHz in table 1. So low, it’s not actually specified in the table, only indicated by extrapolation. In other words, it’s well below a 16-bit quantization noise floor (dither aside), and as good as the full suppression level of most silicon interpolation filters. Which indicates that a DAC interpolation filter, if I’m interpreting this right, is not needed at all for 44.1kHz and higher sample rate digital audio, for full reconstruction. Makes me want to experiment some more with NOS DACs.
Last edited:
Yeah, ESS must know more than enough about DACs.
And nobody in this thread said there are no vendors who actually know how to make great DACs. Though they are not necessarily the highest in SINAD ranking...
Just the general tone that is observed. Apparently the reverse pyramid of knowledge prevents that manufacturers listen to academic hobbyists with all their papers and wisdom. These are bearers of “the truth” in the DAC religion.
Where has this "theory" been presented? AFAIK Amir (or other experts at ASR) do not claim that SINAD measurement tells all about audibility. Their view about transparency is based on so far non-existing evidence to the contrary. As you said yourself all it takes is for somebody to show that audible difference between 2 well measuring dacs exists. This however requires objective, properly controlled listening test. So instead of hand-waving why not perform such test yourself.
So it is so easy, lets look at DACs SINAD score table on ASR, choose the cheepest DAC from "excellent" side. Profit. Why to pay more 😎
Last edited:
Or you could buy whatever pleases your ears. And accept the fact that thousands of people have bought those dacs with "excellent" SINAD and are happy with the sound and your opinion about the sound is not better than theirs.
I did not read the paper. But if it didn't start with a clear description and verification of the used sound exciters and its clinical performance, I dont give a rats butt about it...
So mics with very low self noise, a very good transducer, clean amps etc...
Clearly, the equipment used to perform the study need to be a great deal better then what the results show. I think it could be very easy to have sidebands (high AND low) coming out of the sound emitter and reveal its output in all sorts of ways...
Is there such a declaration?
//
Only problem is that from time to time the DAC will make Cheep Cheep Cheep sounds... 😛
- Status
- Not open for further replies.