RME ADI-2 Pro FS R vs. MOTU M2

[manisandher]

And oh, I finished my physics 'career' in 1993.

 

[watchnerd]

I also have an RME ADI-2 Pro.

I've also played around with vinyl capture a lot, as well as 15 IPS reel-to-reel.

On playback, reel-to-reel vs the digital capture thereof are closer than the vinyl equivalent.

Vinyl playback, in real life (as opposed to ripping with speakers off), involves acoustic feedback from the speakers impacting the cartridge and turntable.

@Frank Dernie has measured this.

Reel to reel doesn't have this phenomenon to contend with.

Except for experiments, I don't bother with vinyl rips anymore.

Amongst modern, good DACs, I don't think swapping DACs does much to solve the "problem" of good vinyl ripping.

 

[outerspace]

What was the difference between A and B?

Because you don't want to ABX them can you share your impressions about my fragments first. With your "relaxing listening". Or you don't really hear the difference? After that I'll share my impressions backed up by abx logs and reveal the real names.

 

[daftcombo]

I've done blind ABX tests, and really dislike doing them. I find them exhausting, and not fun in anyway whatsoever. My brain is forced into a state that it simply never goes into when ordinarily listening to music.

I did a blind ABX test a couple of years ago where I scored 9/10. The guy I did it with didn't believe it was possible that I could really be hearing a difference (he couldn't hear it himself) and finally put it down to pure luck. For my part, had I scored 5-6/10, I would have contuinued to believe that I could indeed hear a difference (because it was so obvious to me during casual listening) and that the ABX method is flawed.

So what did the ABX achieve? Nothing.

Mani.

The ABX is not flawed but it can definitely be exhausting.

 

[daftcombo]

@Frank Dernie has measured this.

Interesting, Where is it?

 

[watchnerd]

Interesting, Where is it?

On paper in some drawer, I would assume. You'll have to ask him.

It was research done while he was working for a TT manufacturer in the 1980s, I believe.

 

[manisandher]

Vinyl playback, in real life (as opposed to ripping with speakers off), involves acoustic feedback from the speakers impacting the cartridge and turntable.

Yes. All my vinyl rips are done with headphones. The TTs are mounted on Townshend seismic stands on solid concrete floors. No acoustic feedback here!

@Frank Dernie has measured this.

Ha, Frank has the same speakers I have!

Mani.

 

[Tks]

I think any 9-10/10 is very unlikely to be a false positive. However, I think a 5-6/10 can easily be a false negative, due to stress involved in doing the test. Are ABX tests still useful? Maybe.

Okay, so we've gotten from achieving nothing - to maybe useful. That's good.

Also stress? That's a classic, but at some point you have to bite bullets and say nothing is testable (like scholastic testing in educational systems) due to "stress". While there is some of that in every notion of a human performance test, at some point it doesn't really hold water if you repeat the test enough.

Also a 9/10 here isn't described, simply because we don't know what the test itself was. There's a massive difference between a 9/10 test between binary decisions, and one where you're asked to provide elaboration on such choice (like if there is a difference, which file is which for example).

The details of your test are undisclosed to make proper comment. But I don't really need the details. The point of contention was how as someone involved in physics, could you deduce the notion that ABX testing achieves nothing if properly conducted with enough factors under control.

A 9/10 has a p=0.01.

Sure, I guess? I mean I don't know what that needed stating for.

 

[watchnerd]

Yes. All my vinyl rips are done with headphones. The TTs are mounted on Townshend seismic stands on solid concrete floors. No acoustic feedback here!

Do you always listen with headphones?

The acoustic feedback is also through the air.

I've done vinyl rips where my wife yelling at the cats showed up in the rip due to being picked up by the cart.

 

[manisandher]

Because you don't want to ABX them can you share your impressions about my fragments first. With your "relaxing listening". Or you don't really hear the difference? After that I'll share my impressions backed up by abx logs and reveal the real names.

If I was certain I could hear a difference during "relaxed listening", I wouldn't be averse to doing an ABX. But I really can't hear a difference in your files, so no point in doing an ABX.

I'd love to hear your impressions and what I'm missing.

Mani.

 

[manisandher]

Do you always listen with headphones?

I only record using headphones. I listen exclusively through speakers.

The acoustic feedback is also through the air.

I've done vinyl rips where my wife yelling at the cats showed up in the rip due to being picked up by the cart.

Haha...

I had a little 'studio' built in my basement especially for my vinyl rips. It's acoustically insulated... and has a Faraday cage around it! No idea if the latter makes any difference though, but seemed like a good idea at the time.

Mani.

 

[outerspace]

I'd love to hear your impressions and what I'm missing.

OK. Conclusions from this aligned fragments of your last recording session: "2 ABX" (originally A) and "3 ABX" (originally B) have some subtle harshness at 28.0 sec that "1 ABX" (originally C) don't. Seems like this is due to variations in two recording takes.

Spoiler: ABX log

foo_abx 2.0.6d report
foobar2000 v1.4
2020-09-07 17:13:18

File A: 1 ABX.flac
SHA1: 3da013661ecd1747af759334a392729884e0f5bb
File B: 2 ABX.flac
SHA1: 468790fbe8770a939be6e19581fa2231ea44c8be

Output:
DS : DAC (3- CA DacMagic XS)
Crossfading: NO

17:13:18 : Test started.
17:13:37 : 01/01
17:13:43 : 02/02
17:13:59 : 03/03
17:14:08 : 04/04
17:14:19 : 05/05
17:14:27 : 06/06
17:15:33 : 07/07
17:17:14 : 08/08
17:19:25 : 09/09
17:20:59 : 10/10
17:20:59 : Test finished.

----------
Total: 10/10
p-value: 0.001 (0.1%)

-- signature --
760c10f586e0433276496887531a62ed35ee3431

 

[watchnerd]

I only record using headphones. I listen exclusively through speakers.

That's the point.

Listening to a vinyl rip, though speakers, made with speakers off, doesn't replicate the real-time acoustic feedback you get when listening to LP through speakers.

So while it might sound good, it will never sound quite the same because that extra bit of distortion has been eliminated.

Swapping DACs doesn't change that.

 

[manisandher]

Also stress? That's a classic, but at some point you have to bite bullets and say nothing is testable (like scholastic testing in educational systems) due to "stress".

This is interesting. It's been a long time since I sat an exam, but I never found them stressful. I studied. I revised. I prepared. Sitting down at the exam table, I suspect my mind was in a very similar state to that in which it was doing the revision and past papers.

Listening to music relaxes me. The better the music and SQ, the more easily I seem to get put into some sort of 'deep meditative' (almost dream-like) state. For me, doing an ABX is about conscious focus and is hard work... and nothing like listening to music.

Mani.

 

[outerspace]

For me, doing an ABX is about conscious focus and is hard work... and nothing like listening to music.

But why can't you do the ABX in the same relaxed state if you think you can spot the differences without hard concentration? Do it without "focus and hard work". Just listen and relax if you think it's help you.

The reality is our ability to spot audio differences degrade almost to zero without hard concentration and instantaneous A/B'ing. Our audio memory is very short. If you can't spot the difference in instantaneous A/B'ing you'll never can do it with long relaxed session. The ABX test was proven to be more sensitive than long therm listening.

 

[manisandher]

OK. Conclusions from this aligned fragments of your last recording session: "2 ABX" (originally A) and "3 ABX" (originally B) have some subtle harshness at 28.0 sec that "1 ABX" (originally C) don't. Seems like this is due to variations in two recording takes.

Apart from this (which is no doubt due to the different takes) can you hear any difference in sound quality between the files? That's the only thing I'm interested in.

In the raw files I linked, there are subtle differences in levels, clicks, ticks and distortion. You could ABX these easily, which is not a very useful thing to do, and I understand (and appreciate) therefore why you've gone to lengths to normalise them ready for ABX testing.

However... In one of the raw files, the transients are better delineated, and there is more ambient information coming through. These things are subtle, but I believe they're there. As a result, it's simply more enjoyable to listen to than the other two raw files. (To my ears at least.)

I'm interested in knowing if other people hear what I believe I'm hearing. If the answer is "no" and "therefore you're probably just imaging things Mani", that's absolutely fine with me. Really.

Mani.

 

[manisandher]

But why can't you do the ABX in the same relaxed state if you think you can spot the differences without hard concentration? Do it without "focus and hard work". Just listen and relax if you think it's help you.

Yeah, I agree. But it's more easily said than done.

With your files, I listened to them in a 'relaxed state', but really couldn't hear a difference. I'd be going into the ABX knowing that I'd just be guessing - not a particularly useful state of mind to be in.

 

[outerspace]

However... In one of the raw files, the transients are better delineated, and there is more ambient information coming through. These things are subtle, but I believe they're there.

But when you listen aligned files with hidden names you don't hear this. How do you explain this for yourself?

 

[manisandher]

So while it might sound good, it will never sound quite the same because that extra bit of distortion has been eliminated.

Sure. It's actually one of the (many) reasons why I've gone down the vinyl rip route.

 

[watchnerd]

Sure. It's actually one of the (many) reasons why I've gone down the vinyl rip route.

Interesting.

And that's why I stopped ripping.

I didn't want the distortion removed.

If I want clean and pristine, I listen to digital.

 

[manisandher]

But when you listen aligned files with hidden names you don't hear this. How you explain this for yourself?

I'll open a can of worms if I answer this right now. But I promise to address this in a future thread when I can substantiate things a bit more. OK?