I said before that I haven' read it because it is behind a paywall, and I asked you if you had specific concerns.
Here is the problem I have: I have a number a DACs here, and they all sound different. Some are a few years old, and I thought maybe the newer DACs don't have all the same problems. So, I ordered a DACMagic+ DAC which is highly rated in it's price range, about $350. This is it right here: Amazon.com: Cambridge - DACMagic Plus (Silver): Home Audio & Theater
The day it came with great hope I opened the box took it out and hooked it up. It took less than 10 seconds to hear it was not up the the accuracy of a 10-year old Benchmark DAC-1. It was smeared and phasey sounding, a little muffled. I thought about for a little while if there was any possible use I might have for a DAC like that, and I couldn't think of anywhere I would want to use it, so I sent it back. It sounded similar to the DAC in my old Crane Song HEDD which I use as an nice ADC, but I never could stand the sound of the DAC in it. Eventually, I replaced the DAC-1 with a DAC-3. I sold the DAC-1 at a very good price to friend in a band who does some recording. He has tracked and mixed 3 or 4 albums, and is pretty decent at it. When he got the DAC-1 he set it aside. He really only bought it because I offered it a good price and he figured he could maybe sell it for profit if he didn't like it. He had been using a Lynx-2 card up to that point which has very nice specs. A couple months later I got an email from him saying he just got around to to hooking up the DAC-1. He said, "Wow! It sounds so much better!" but, of course after doing as much recording as he has he has learned how to listen fairly well. I could go on and on with experiences, blind testing, how they sound and so on. The bottom line is they sound different, including blind. And some of them have rather peculiar sounds different than amps or speakers. So, when someone says "All DACs are proven to sound the same by four guys, each of whom represents 25% of the Earth's human population, thereby proving that no human can hear any difference between DACs," I can hardly believe it. It's CRAZY that someone would think 4 guys can prove that by their failure. ABSOLUTELY CRAZY!!! Okay, maybe your are starting to understand. I am in Northern California, if you were around here I would invite you over and within a few minutes you would be saying, "well, who would have thought...I mean how could they?...Well, I guess they do sound different." Because you could hear for yourself, and do it blind.
Of course, I would probably have to coach you a little. Did that with my brother, played one DAC then the other. He said they sound the same. I said okay listen to the cymbals in this song right now and think about how real cymbals sound in person. I want you to try to remember how these recorded cymbals sound in each of three different frequency ranges and pay attention to if any of them sound like bursts of noise rather than cymbals when you listen carefully, okay. Then I switch DACs again, "how do these cymbals sound in comparison?" I ask. "They sound more clear, Hmm, I guess the DACs do sound different." But, he says how much does this one cost compared to that one? The clear one is $2,000 and the dirty one is $200. He said, it's not worth it to me for the additional clarity.
Of course, hearing a small difference in cymbals is only a small part of the differences between the DACs, but it's the easiest thing to notice first. And it wouldn't be worth it to him to spend the time to learn how to listen to all the differences. He doesn't care, and will forget the whole thing by the time I'm out the door. But, he could hear something. By the way, the dirty DAC was a cheap, old Tascam USB outboard recording interface. The clear DAC was the awful-to-me Crane Song HEDD.
Last edited:
No claims made in that post. Just a list of how to characterize fully an electrical waveform. If you have more ideas or more possible sources of distortion please list them, I'm all ears.
Yeah.. I and an entire industry. Please educate yourself about why sinewave testing is the standard before muddying the water with "that's only your opinion" debate tricks.
Earl Geddes said he uses real music for testing because test waveforms are not close enough to the complexities of how music interacts with nonlinearities.
Yes, you can come up with numbers, but you can't tell how they correlate to human perception. So they aren't useful, according to Geddes, who researches human hearing.
Please let's keep things in context: sinewaves aren't completely useful to correlate distortion with audibility. Because our ears+brain might react differently to music compared to sinewaves. That's a given.
However, when it comes to the DAC itself, it's quite different and if we just do measurements then sinewaves testing is sufficient to establish the transfer function of a device.
iow: sinewaves aren't good enough for establishing audibility metrics. But they're good enough to know what kind of distortions a particular DAC will produce.
However, when it comes to the DAC itself, it's quite different and if we just do measurements then sinewaves testing is sufficient to establish the transfer function of a device.
iow: sinewaves aren't good enough for establishing audibility metrics. But they're good enough to know what kind of distortions a particular DAC will produce.
This is what you posted "Once you measure variations in frequency response, phase, amplitude, noise and harmonic spectrum at all power and frequency points, measure IMD (again at various frequency/power points), check what's happening out of band and keep the system operating in specs (no digital clipping and the like), you pretty much have all you need to know about the ability of a particular DAC to be faithful to the input signal. There are a few more implementations concerns like crosstalk and emi interferences to account for. You can also check for jitter independently to isolate the problem."
I read this as a claim but now you are saying is that you can list off a series of measurements that SHOULD be made but you can't actually produce your list for any DAC, that's fine
Yea, next it will be the whole spiel about Nobel prize - please desist from these well worn techniques!
Ah, so you don't have any questioning about DAC acting differently with complex dynamic signals? Why? Because you only ever measure with Sinewaves!!
And the whole point is that you don't know this - you just state it as if it's true.
I simply don't have the time to do it. All those measurements taken one by one are routinely made on this forum but taking the time to do them all on one particular device is serious work and requires specialized equipment. But again, taken separately, they're all well established. If you want a starting point, see how Tomchr did characterize the modulus-86.
Sure, sue me.
Agreed, taking measurements is a lot of work. Here is a DAC-3 manual: https://cdn.shopify.com/s/files/1/0321/7609/files/DAC3_Series_Manual_Rev_B.pdf?9982830537634228604
Performance graphs and specifications take up about 20 pages.
However, most manufacturers don't measure all that stuff, and most people that build DACs here in the forum don't either. They probably would all measure different and sound different too. Question is to what extent can the sound be predicted from the measurements? Not completely, I would say.
And some things can be hard to pin down with measurements unless maybe one knows what one is hunting for. Thermal distortion?
What does limit-cycling sound like, and what kind of distortion does it show up as?
What does transient phase distortion sound like and how do you measure it in a meaningful way that gives an idea as to how much perceptual sound quality is affected?
Performance graphs and specifications take up about 20 pages.
However, most manufacturers don't measure all that stuff, and most people that build DACs here in the forum don't either. They probably would all measure different and sound different too. Question is to what extent can the sound be predicted from the measurements? Not completely, I would say.
And some things can be hard to pin down with measurements unless maybe one knows what one is hunting for. Thermal distortion?
What does limit-cycling sound like, and what kind of distortion does it show up as?
What does transient phase distortion sound like and how do you measure it in a meaningful way that gives an idea as to how much perceptual sound quality is affected?
This argument about measurements is the same flawed one as used for ABX testing:
- Sinewave testing is an industry standard; ABX is an industry standard
- We measure using sinewaves & extrapolate to how the device will perform with music; we run ABX listening tests & extrapolate this to how we listen to music casually
- we know measurements are not complete but we believe they are 'good enough'; we know our ABX testing is casual/flawed but we believe it's 'good enough'
However, there's at least one major difference between measurements & ABX testing - the tools used for measurements are usually reliable/sensitive - ABX tests run on forums have no such sensitivity/calibration - the test may not be capable of revealing the level of differences being claimed.
So, based on both these flawed approaches, it's claimed that all DACs (amplifiers, etc) sound the same!!
- Sinewave testing is an industry standard; ABX is an industry standard
- We measure using sinewaves & extrapolate to how the device will perform with music; we run ABX listening tests & extrapolate this to how we listen to music casually
- we know measurements are not complete but we believe they are 'good enough'; we know our ABX testing is casual/flawed but we believe it's 'good enough'
However, there's at least one major difference between measurements & ABX testing - the tools used for measurements are usually reliable/sensitive - ABX tests run on forums have no such sensitivity/calibration - the test may not be capable of revealing the level of differences being claimed.
So, based on both these flawed approaches, it's claimed that all DACs (amplifiers, etc) sound the same!!
Last edited:
Of course, prediction is not a given since we definitely lack studies establishing correlation between various measurements and their impact on our auditory system. So in the imperfect world we have to live in, I've got to agree.
Thermal distortion does already show up in "stationary" thd measurements at low frequencies (not at the common 1khz or 20khz though). Thermal distortion in resistors used for feedback created some puzzle on a preamp thread recently because the harmonic spectrum measurements were significantly different depending on the resistors used. Bob Cordell also suggested to run smpte im tests, using a 20hz signal, to check for thermal distortion.
Wrt sinewaves, you have to keep in mind that a sinewave 20khz signal is by definition much harder on any gear than any musical signal can ever be (if proper recording and mastering techniques were used). Careless testing can easily destroy your amp or speakers.
And not all testing is done with sinewave. Clipping and stability testing uses square waves. You also have multitones and IMD test as well as impulse tests.
If you really agree with your first statement above then I don't see what your problem is with my position - maybe you can clarify? Neither measurements nor ABX tests as we see here are sufficient evidence for the claims made about the auditory perception of DAC differences
Last edited:
You think that being able to destroy your amp/speakers means that sinewave test signals exercises the device in the same way as music signal does?
Did you not just claim the following "Sinewave testing is a perfectly valid substitute for audio measurements btw (at least for DACs, under the constraint that we are dealing only with legit signals). Please make a search on that. The DAC knows nothing of the "dynamic complexity of music".
Any idea why then would anybody use a different test signal if sinewaves are "perfectly valid"?
Agreed on that point, but how many people building DACs bother with it? Of all the DAC projects here in the forum, I haven't seen one with a set of comprehensive measurements. Also, I don't think many commercial manufactures bother. Certainly, publishing comprehensive measurements like Benchmark does is far more the exception than the rule.
Of forum members, one exception I did see is that PMA posted detailed measurements on his website of his DACMagic Plus, but whatever it is that made the one I heard compare poorly to an old Benchmark DAC-1 wasn't apparent to me from his measurements.
It seems to me that we are still reliant on listening tests to differentiate the perceptual performance quality of most or all DACs on the market (which it sounds like we may be in agreement on). Therefore choosing a DAC to practice listening skills, or for use in non-DAC ABX audibility discrimination listening test systems remains problematic for most people who don't have access to a skilled listener to help validate system performance. The only reliable way I could otherwise recommend a DAC to use would be to specify a Benchmark Media product. Unfortunately, their cost will likely cause many people to rationalize that less expensive products are equally good for the above purposes.
Last edited:
In trying to attach meaning to this I can only assume that "the level of differences being claimed" is a claim made by the DAC industry that their DACs all sound different yet this difference is so small that it cannot be detected in these "amateur" ABX tests. This can only have meaning if the DAC industry know how different their DACs are (yet they cannot or will not tell us in technical terms, in most cases?) and the DAC industry knows how sensitive the "amateur" tests are. So we have two people deducing opposite conclusions from the same test:mmerrill99 said:
1. "Your DACs are all so similar that I cannot tell the difference."
2. "Our DACs are so different that if you cannot hear this then your test is not sufficiently sensitive."
An obvious solution is for the DAC industry to tell us how different their products are, so a suitable test can be defined. If they cannot do this then they cannot bleat on about tests being insensitive.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.