Serious Question: How can DAC's have a SOUND SIGNATURE if they measure as transparent? Are that many confused?

[PabloAlbita]

Must be the 'or someone.'



Glitch might have been on your end, but you can certainly continue to participate in a good faith discussion if you'd like to.

I tried to log in a bunch over the last few days and it said something like "Oops we ran into some problems. You have been banned."

Anyway, apologies.

 

[PabloAlbita]

Since we've got a cowboy betting $50K on the thread now, I can be demanding too: assuming all the other proper controls are in place, I insist on a p<0.01 result or difference ain't proved.

Anyone who believes, as you do, that “DACs well below 'best' in SINAD rankings will be audibly indistinguishable in level-matched blind comparison.” by definition believes that any distinction made during such a comparison will be purely luck.

It's equivalent to saying, “any guess of the flip of a proper coin will be purely luck”

So therefore, if someone says, “I will bet $X that I can guess correctly enough times that it would only be expected 5% of the time”, a person confident in their position would eagerly take up that bet for as much money as possible.

They would not say, as you did, “no, that's too scary for me. To take a bet, I will only accept an outcome that could only be expected to happen 1% of the time.” On a coin flip.

One reasonable interpretation of this irrational and unprofitable response would be a lack of confidence in your original statement.

 

[Blumlein 88]

Anyone who believes, as you do, that “DACs well below 'best' in SINAD rankings will be audibly indistinguishable in level-matched blind comparison.” by definition believes that any distinction made during such a comparison will be purely luck.

It's equivalent to saying, “any guess of the flip of a proper coin will be purely luck”

So therefore, if someone says, “I will bet $X that I can guess correctly enough times that it would only be expected 5% of the time”, a person confident in their position would eagerly take up that bet for as much money as possible.

They would not say, as you did, “no, that's too scary for me. To take a bet, I will only accept an outcome that could only be expected to happen 1% of the time.” On a coin flip.

One reasonable interpretation of this irrational and unprofitable response would be a lack of confidence in your original statement.

No that's an incorrect analysis. It's means someone does not wish to gamble with a 5% chance of losing. Every one has a level of risk threshold for risking their money. The amount matters differently to each person as well.

 

[PabloAlbita]

No that's an incorrect analysis. It's means someone does not wish to gamble with a 5% chance of losing. Every one has a level of risk threshold for risking their money. The amount matters differently to each person as well.

Very true. We all have our own risk thresholds.

As you also point out, the sum may matter. If $50k is all one has, perhaps a 95% chance at $100k is less valuable than just keeping the $50k.

But it's a rare person indeed who would turn down a 95% chance of doubling their money. So it's also a very reasonable conclusion that that person does not privately believe what they're publicly professing about DACs being indistinguishable, since they refuse to stand behind that statement in any way beyond empty words.

 

[Blumlein 88]

Very true. We all have our own risk thresholds.

As you also point out, the sum may matter. If $50k is all one has, perhaps a 95% chance at $100k is less valuable than just keeping the $50k.

But it's a rare person indeed who would turn down a 95% chance of doubling their money. So it's also a very reasonable conclusion that that person does not privately believe what they're publicly professing about DACs being indistinguishable, since they refuse to stand behind that statement in any way beyond empty words.

Many have your idea. I don't think it holds water, certainly not universally or even close to it. Mr. Clark and his amplifier challenge offered people the chance to win $10,000. Didn't cost them anything other than time and effort to try. Later on with so many takers I think he required someone to put up $500 for the trouble. Still mostly a chance to win $10k with little downside. Since the risk was all his he did want better than 95% confidence. Some offers made in this forum were of that type. Some weren't. I see it as more of a dick waving bet.

Counter side to that is why do you need a bet to get involved? If you or someone is so sure, then try the test and see what happens. I'd want more like the 99% confidence version, but 95% would be interesting under those circumstances. Mr. Clark's offer was to induce people to try since so few seem willing to do it for nothing. Plenty have the opportunity to test themselves for nothing and few wish to do so.

 

[PabloAlbita]

Many have your idea. I don't think it holds water, certainly not universally or even close to it. Mr. Clark and his amplifier challenge offered people the chance to win $10,000. Didn't cost them anything other than time and effort to try. Later on with so many takers I think he required someone to put up $500 for the trouble. Still mostly a chance to win $10k with little downside. Since the risk was all his he did want better than 95% confidence. Some offers made in this forum were of that type. Some weren't. I see it as more of a dick waving bet.

Counter side to that is why do you need a bet to get involved? If you or someone is so sure, then try the test and see what happens. I'd want more like the 99% confidence version, but 95% would be interesting under those circumstances. Mr. Clark's offer was to induce people to try since so few seem willing to do it for nothing. Plenty have the opportunity to test themselves for nothing and few wish to do so.

Well I started the bet idea because I happen to know someone here who has funds to handle it like I do, and thought it would be fun to see how much faith in their convictions those who advocate the "they all sound the same" side have. And because I think the results (with the video) would make for very interesting discussion. And, of course, because I known I'd win because I've done the same test a bunch of times with my wife just to make sure I could win; one would be an abject fool to go into something like this not knowing for sure when it would be easy to do so.

The bottom line is that if one truly believes 100% that all of properly-functioning DACs sound the same, it should be the easiest thing in the world to accept a p<0.05 bet and that we'd have it set up in a week. But clearly that hasn't been the result. On the contrary, it devolved into all sorts of doubts and fears and caveats micro-requirements and in your case a requirement of almost perfect certainty, which is a completely irrational outcome (excepting, as you well note, the high sum involved and one's personal risk profile), and so forth. I didn't expect that, and find that psychologically fascinating and hugely revealing.

 

[ahofer]

because I've done the same test a bunch of times with my wife just to make sure I could win

The trouble is that your test is a test of the headphone amplification and level mismatch, not the DAC. Nobody here has suggested that there isn’t a difference when you confound with those types of variables. The DAC you chose has a massive output voltage (by design, I think). I don’t have the equipment to knock it down and level match. If you want to call level matching and impedance/ power matching “special requirements” go ahead, but nobody is going to take you seriously in a technical forum (see my signature quote). And if you keep strawmaning as “all DACs sound the same” when we have always specified audibility thresholds doesn’t help your case.

Not to mention the statistical problems of one listener/one test. I don’t think this is as “revealing“ as you think it is. It is just normal people trying to make sure they don’t get had by small variations in test conditions (just as you tried with your wife tests).

 

[Mart68]

Well I started the bet idea because I happen to know someone here who has funds to handle it like I do, and thought it would be fun to see how much faith in their convictions those who advocate the "they all sound the same" side have. And because I think the results (with the video) would make for very interesting discussion. And, of course, because I known I'd win because I've done the same test a bunch of times with my wife just to make sure I could win; one would be an abject fool to go into something like this not knowing for sure when it would be easy to do so.

The bottom line is that if one truly believes 100% that all of properly-functioning DACs sound the same, it should be the easiest thing in the world to accept a p<0.05 bet and that we'd have it set up in a week. But clearly that hasn't been the result. On the contrary, it devolved into all sorts of doubts and fears and caveats micro-requirements and in your case a requirement of almost perfect certainty, which is a completely irrational outcome (excepting, as you well note, the high sum involved and one's personal risk profile), and so forth. I didn't expect that, and find that psychologically fascinating and hugely revealing.

For a big bet it's best to be sure that all the parameters are covered.

Poker players are fond of proposition bets. At a tournament one declared that he could drive a golf ball more than a mile. The bet was immediately taken up by two others, at ten thousand dollars each. Which was a lot of money back in the day.

The Sunday break came around so all three got in a car to drive to a golf course. Except they didn't. It was the depths of the winter and they drove to a frozen lake. The ball travelled well over a mile on the low friction surface, and money changed hands.

 

[PabloAlbita]

The trouble is that your test is a test of the headphone amplification and level mismatch, not the DAC. Nobody here has suggested that there isn’t a difference when you confound with those types of variables. The DAC you chose has a massive output voltage. I don’t have the equipment to knock it down and level match. If you want to call level matching and impedance/ power matching “special requirements” go ahead, but nobody is going to take you seriously in a technical forum (see my signature quote).

Not to mention the statistical problems of one listener/one test. I don’t think this is as “revealing“ as you think it is.

But I offered to do it with very sensitive IEMs, not headphones. At least I believe I did; maybe I forgot? In any case, I do now.

 

[PabloAlbita]

For a big bet it's best to be sure that all the parameters are covered.

Poker players are fond of proposition bets. At a tournament one declared that he could drive a golf ball more than a mile. The bet was immediately taken up by two others, at ten thousand dollars each. Which was a lot of money back in the day.

The Sunday break came around so all three got in a car to drive to a golf course. Except they didn't. It was the depths of the winter and they drove to a frozen lake. The ball travelled well over a mile on the low friction surface, and money changed hands.

Wow, I knew the drive the ball a mile thing, but I was never aware that was a true story!

 

[Mart68]

Wow, I knew the drive the ball a mile thing, but I was never aware that was a true story!

From memory it was Doyle Brunson who won the money, but it could well be apocryphal. I mean it's poker. No-one tells the truth.

 

[PabloAlbita]

From memory it was Doyle Brunson who won the money, but it could well be apocryphal. I mean it's poker. No-one tells the truth.

Hahahaha point taken!!

 

[ahofer]

But I offered to do it with very sensitive IEMs, not headphones. At least I believe I did; maybe I forgot? In any case, I do now.

You even offered to use a headphone amp, which is the only way to do it properly. But you still have to make sure the input voltage to the headphone amp is matched.

And for your sake, and the durability of the test outside of just two people you still need to be able to switch rapidly to A, B, and X.

Thread 'Audio Blind Testing - You Are Doing It Wrong! (Video)'
https://audiosciencereview.com/foru...d-testing-you-are-doing-it-wrong-video.26809/

 

[PabloAlbita]

You even offered to use a headphone amp, which is the only way to do it properly. But you still have to make sure the input voltage to the headphone amp is matched.

Well that could easily be done using a little Mackie mixer, though I'd hate to have to add another signal path and distort the sound even more. I'd have to test that first to make sure I could still pass the test. Mackies are notably reasonably lousy.

 

[PabloAlbita]

Well that could easily be done using a little Mackie mixer, though I'd hate to have to add another signal path and distort the sound even more. I'd have to test that first to make sure I could still pass the test. Mackies are notably reasonably lousy.

Actually, even better, a passive attenuator for the hotter signal. The would take gain-staging electronics out of the signal path.

 

[DSJR]

I remember when I 'heard' the M-Scaler in a dem conducted by the Chord rep and after the introduction/lecture by Rob Watts. The M-Scaler DEFINITELY made an audible difference (track was by 'Uncle Lucious') but I had a nagging doubt the levels were slightly different, the result being a deeper perceived soundstage with M-Scaler in situ. tests subsequently showed how inserting it reduced the volume levels which could quite easily explain it!

More recently, I tried comparing my SMSL SU1 (very happy with it and it's a great cheap tool for decoding the USB feed from the PC, the PC supply seemingly okay now I've taken out the wallwart it previously used) with a midi sized Sony La Scala CD player. The Sony via its RCA outputs 'sounds' ever so slightly 'soft' up top, or appears to when compared loosely with the optical out into the SU1. The SU1 does have a higher output and being more diligent in matching levels, these 'differences' seem to magically all but disappear - I suspect that a religiously applied level match would totally remove any audible difference and anyway, the optical digital output is fine if I want to be anal about it.

So, even if I was in the locale, I certainly cannot simply 'trust my ears' as subjectivists all claim to do so well. Maybe I feel better relieved of that burden these days!

 

[Yuhasz01]

Well I started the bet idea because I happen to know someone here who has funds to handle it like I do, and thought it would be fun to see how much faith in their convictions those who advocate the "they all sound the same" side have. And because I think the results (with the video) would make for very interesting discussion. And, of course, because I known I'd win because I've done the same test a bunch of times with my wife just to make sure I could win; one would be an abject fool to go into something like this not knowing for sure when it would be easy to do so.

The bottom line is that if one truly believes 100% that all of properly-functioning DACs sound the same, it should be the easiest thing in the world to accept a p<0.05 bet and that we'd have it set up in a week. But clearly that hasn't been the result. On the contrary, it devolved into all sorts of doubts and fears and caveats micro-requirements and in your case a requirement of almost perfect certainty, which is a completely irrational outcome (excepting, as you well note, the high sum involved and one's personal risk profile), and so forth. I didn't expect that, and find that psychologically fascinating and hugely revealing.

Read Darrell Huffs' classic paperbacks, How to Lie with Statistics or How to Take A Chance: Probability. Most of the 300 or so posts here are nonsense statistically.

 

[ahofer]

I remember when I 'heard' the M-Scaler in a dem conducted by the Chord rep and after the introduction/lecture by Rob Watts. The M-Scaler DEFINITELY made an audible difference (track was by 'Uncle Lucious') but I had a nagging doubt the levels were slightly different, the result being a deeper perceived soundstage with M-Scaler in situ. tests subsequently showed how inserting it reduced the volume levels which could quite easily explain it!

More recently, I tried comparing my SMSL SU1 (very happy with it and it's a great cheap tool for decoding the USB feed from the PC, the PC supply seemingly okay now I've taken out the wallwart it previously used) with a midi sized Sony La Scala CD player. The Sony via its RCA outputs 'sounds' ever so slightly 'soft' up top, or appears to when compared loosely with the optical out into the SU1. The SU1 does have a higher output and being more diligent in matching levels, these 'differences' seem to magically all but disappear - I suspect that a religiously applied level match would totally remove any audible difference and anyway, the optical digital output is fine if I want to be anal about it.

So, even if I was in the locale, I certainly cannot simply 'trust my ears' as subjectivists all claim to do so well. Maybe I feel better relieved of that burden these days!

It gets mundane when you reduce it to what you are doing with audibility thresholds and level-matching: If the distortion is really low and the FR and amplitude match, the transducer will produce the same signal. It’s really a ‘there’s nothing you can’t measure’ argument. But controlling for levels and FR is non-trivial.

 

[PabloAlbita]

I remember when I 'heard' the M-Scaler in a dem conducted by the Chord rep and after the introduction/lecture by Rob Watts. The M-Scaler DEFINITELY made an audible difference (track was by 'Uncle Lucious') but I had a nagging doubt the levels were slightly different, the result being a deeper perceived soundstage with M-Scaler in situ. tests subsequently showed how inserting it reduced the volume levels which could quite easily explain it!

More recently, I tried comparing my SMSL SU1 (very happy with it and it's a great cheap tool for decoding the USB feed from the PC, the PC supply seemingly okay now I've taken out the wallwart it previously used) with a midi sized Sony La Scala CD player. The Sony via its RCA outputs 'sounds' ever so slightly 'soft' up top, or appears to when compared loosely with the optical out into the SU1. The SU1 does have a higher output and being more diligent in matching levels, these 'differences' seem to magically all but disappear - I suspect that a religiously applied level match would totally remove any audible difference and anyway, the optical digital output is fine if I want to be anal about it.

So, even if I was in the locale, I certainly cannot simply 'trust my ears' as subjectivists all claim to do so well. Maybe I feel better relieved of that burden these days!

Yeah, when I first got the M-Scaler in to test, I was wholly unimpressed. It has a bypass on it so you can instantly hear its effect (which maintains the 3dB down when either in line or bypassed so there's no differential level issue) and it felt like more or less zero effect was going on, or at least not $5k worth. But after a few hours of listening, I began to hear what it's doing and now find it uncomfortable to listen without it. Sometimes I'll turn on the system and start listening and feel like, "meh, it's fine, but I'm not all that engaged" only to later find that I still had the upscaling turned off because I'd been watching a video (upscaling adds a delay so you can't watch video with it and be in sync). When I come out of video mode, it feels great again to listen to.

Of course I am fully aware this sort of reaction means less than zero to anyone here, so no need to start trashing me left and right; I only mention it in light of your comment.

However, as important as its effect is to my own listening enjoyment, I am completely sympathetic to the view that for almost everybody else in the world, there are much, much better ways to spend $5k.

 

[Purité Audio]

I remember when I 'heard' the M-Scaler in a dem conducted by the Chord rep and after the introduction/lecture by Rob Watts. The M-Scaler DEFINITELY made an audible difference (track was by 'Uncle Lucious') but I had a nagging doubt the levels were slightly different, the result being a deeper perceived soundstage with M-Scaler in situ. tests subsequently showed how inserting it reduced the volume levels which could quite easily explain it!

More recently, I tried comparing my SMSL SU1 (very happy with it and it's a great cheap tool for decoding the USB feed from the PC, the PC supply seemingly okay now I've taken out the wallwart it previously used) with a midi sized Sony La Scala CD player. The Sony via its RCA outputs 'sounds' ever so slightly 'soft' up top, or appears to when compared loosely with the optical out into the SU1. The SU1 does have a higher output and being more diligent in matching levels, these 'differences' seem to magically all but disappear - I suspect that a religiously applied level match would totally remove any audible difference and anyway, the optical digital output is fine if I want to be anal about it.

So, even if I was in the locale, I certainly cannot simply 'trust my ears' as subjectivists all claim to do so well. Maybe I feel better relieved of that burden these days!

The M-scaler does appear to be nothing more than a ‘mugs-eyeful’.

CHORD M-Scaler Review (Upsampler)

This is a review, listening tests and measurements of the CHORD Hugo M-scaler upsampling digital transport. It is on kind loan from a member and costs US $5,295 (some show it as $5,650). We have the usual lightshow of buttons as other CHORD digital products but here, due to less overloading...

www.audiosciencereview.com

Keith