[bigshot]

Musicians are concerned about the same thing. They might call it dynamic shading or shaping. It's commonly discussed. Musicians that do it well are praised or lauded. Etc.

I've mentioned microdynamic resolution, PRaT, and beauty. No one else has mentioned what they enjoy about music since I joined this discussion.

We're talking about sound reproduction, not sound creation, and none of those things have anything to do with circuit boards and wires. That's all a product of flesh and blood musicians.

I regularly discuss music in the music forum as well as on my blog and several music forums.

 

[kiteki]

So if you believe in peer reviewed papers and critical test conditions, shouldn't you be saying MP2 at 256 kbps is transparent and the rest is mythology? That's what Wikipedia and the ABX tests of expert listeners indicate, yes?

 
Some people overinterpret the data or for simplicity's sake (of wording, not throwing in the usual caveats), maybe make looser claims than appropriate. Sure. Anyway, as many have already pointed out, null results don't prove anything.
 
I'm saying that peer-reviewed papers are preferable to non-peer-reviewed papers, and experiments with proper controls are better than experiments without proper controls. It includes blinding, to control for potential biases. That said, a lot of experiments that do blind testing or even double-blind testing, may be junk in other ways.
 
A lot of audiophile claims being made don't really make much sense in the context of the engineering analysis, so current evidence consists of much more than just listening tests, for the record.
 
I'd just point out that statements like these:
 

If two medicines are 100% equal down to the chemical level then they are equal. If there are differences in composition or chemicals it becomes likely they could have different chemical reactions. That's why I look at technology and design in audio and for differences in raw data, instead of completely useless academic papers.

 
seem to reflect a popular sentiment—that different designs / whatevers should sound different. It's not just you. However, the reality is that very different topologies and circuits can produce very similar results, so your intuition from other fields is hurting you. That's one danger of running too many analogies without understanding how they all apply.

 
 
I'm honestly not following you here.  If I believed in the peer reviewed DBT / ABX AES expert listener wikipedia links / papers whatever I'd believe that MP2 256kbps is transparent.
 
If I look at the differences in data / technology / design and trust my hearing, and several others reporting the exact same differences, I'd believe it's not transparent.  So which is closer to the truth here, the first or second.
 
?

 

[kiteki]

By the way Mikeaj I did a vague blind-test on op-amps with 1/21 significance that I could have arrived at the results via chance, which do you think is more convincing to me, random dScope lifeblood theories on the internet, or that my listening arrived at very similar results to others via 1/21 chance?

 

[kiteki]

Anyway, as many have already pointed out, null results don't prove anything.

 
I think you (or Maverick) really need to explain this part too.  If the "null results" don't prove anything, why are they defended so violently, calling the other theories pixie dust, subjective hallucination, sea monsters, audio salesmen, elitism, or preaching that no one should believe, take seriously, or invest resources into?
 
If anything, shouldn't the extremely negative track record of the "null results" make not believing in their current 2012 status a more scientific viewpoint.
 
?

 

[kiteki]

I believe that my current digital setup (modified PS Audio transport, custom DAC by K Works) is far superior to a basic CD player (by "basic" I mean, say, a commonly available CD player which measures with vanishingly low distortion) and I'd be willing to blind-test that if anyone in the LA area wants to help run the experiment.

 
Not familiar with your DAC or anything, but you can buy the ODAC from here if you want - http://www.jdslabs.com/item.php?fetchitem=39, and find someone in the LA area with a new thread, to help your experiment.
 
With two laptops and two different DAC's, connected to the same amplifier or stereo receiver with a "Y cable" or "piggyback cable", someone presses play X or Y computer at random by rolling a dice, you directly identify which DAC is playing.
 
Or, the same DAC is played 4 times in a row (YYYY) versus mixed (XYXY), you identify if it was the same or mixed.
 
Have fun.

 

[maverickronin]

Quote:

I think you (or Maverick) really need to explain this part too.  If the "null results" don't prove anything, why are they defended so violently, calling the other theories pixie dust, subjective hallucination, sea monsters, audio salesmen, elitism, or preaching that no one should believe, take seriously, or invest resources into?
 
If anything, shouldn't the extremely negative track record of the "null results" make not believing in their current 2012 status a more scientific viewpoint.

 
Didn't I explain the idea of a null hypothesis to you before?

 

[mikeaj]

I'm honestly not following you here.  If I believed in the peer reviewed DBT / ABX AES expert listener wikipedia links / papers whatever I'd believe that MP2 256kbps is transparent.
 
If I look at the differences in data / technology / design and trust my hearing, and several others reporting the exact same differences, I'd believe it's not transparent.  So which is closer to the truth here, the first or second.
 
?

 
 
I didn't say to fully take "peer reviewed DBT / ABX AES expert listener wikipedia links / papers" as truth, but as potential evidence, data points to consider or throw out.  Unless you've been trolling us hard, it doesn't particularly seem like you understand the data or technology very well, but if you've done proper listening tests of your own, that counts as potential evidence too.  I would say that a study that is rigorous enough to pass peer review is less likely to have fatal flaws, but peer review doesn't screen everything questionable out, and some good studies don't get through the process.  It's not perfect by any means.  The publication and research quality in some fields are better than in others, too.
 
 

By the way Mikeaj I did a vague blind-test on op-amps with 1/21 significance that I could have arrived at the results via chance, which do you think is more convincing to me, random dScope lifeblood theories on the internet, or that my listening arrived at very similar results to others via 1/21 chance?

 
 
I don't think what's most important is what's most convincing to you (as in you specifically), or to anybody in particular.  [To answer the question, it seems to be the latter.]  But in general (not talking about any particular test) I would say that people rely on personal experience far, far too much.  People that don't have the educational background are less able to figure out which theories have more merit than others, based on the theoretical merits.  It seems like you're treating all theories equally, or without the proper weighting, and are just going by empirical data—dismissing many as just "random dScope lifeblood theories on the internet".  Maybe this assessment is off the mark though.
 
Quote:

 
I think you (or Maverick) really need to explain this part too.  If the "null results" don't prove anything, why are they defended so violently, calling the other theories pixie dust, subjective hallucination, sea monsters, audio salesmen, elitism, or preaching that no one should believe, take seriously, or invest resources into?
 
If anything, shouldn't the extremely negative track record of the "null results" make not believing in their current 2012 status a more scientific viewpoint.
 
?

 
Null results don't prove anything, but they're evidence.  If a null result comes from some well-run experiment, it means a lot.  If a null result comes from a mediocre experiment, it means less, but still something.  People are saying that you shouldn't be throwing out results as if they don't mean anything.  That said, many audio-related studies are not run nearly as well as they could be, and some could use some more rigorous statistical analysis as well.
 
Null results mean that we probably need some more evidence of positive results to start believing the hypothesis.
 
Anyhow, some people make a lot stronger (too strong, too loose, and maybe less accurate) statements than I do, in my estimation.  I try not to overinterpret data, with some moderate success.

 

[kiteki]

 
Didn't I explain the idea of a null hypothesis to you before?

 
Yes and it's clear you don't believe in it.

 

[maverickronin]

It's clear you didn't understand it...

 

[kiteki]

I'm honestly not following you here. If I believed in the peer reviewed DBT / ABX AES expert listener wikipedia links / papers whatever I'd believe that MP2 256kbps is transparent.
 
If I look at the differences in data / technology / design and trust my hearing, and several others reporting the exact same differences, I'd believe it's not transparent. So which is closer to the truth here, the first or second.?

 
I didn't say to fully take "peer reviewed DBT / ABX AES expert listener wikipedia links / papers" as truth, but as potential evidence, data points to consider or throw out. Unless you've been trolling us hard, it doesn't particularly seem like you understand the data or technology very well, but if you've done proper listening tests of your own, that counts as potential evidence too.

 
Sorry but this doesn't answer the question.  I didn't say listening tests I just said my hearing, and we can assume I completely don't understand the technology of codecs or any of the data.
 
Now which was more reliable in finding the transparency in 256kbps in this case - the peer reviewed papers or just my hearing alone?
 
 

Originally Posted by mikeaj /img/forum/go_quote.gif
 
It seems like you're treating all theories equally, or without the proper weighting, and are just going by empirical data—dismissing many as just "random dScope lifeblood theories on the internet". Maybe this assessment is off the mark though.

 
I try to weight everything with it's evidence and likelihood of correctness as much as possible.  If someone says the Clip+ sounds the same as the iBasso DX100 due to RMAA, what do you think?
 
 

Originally Posted by mikeaj /img/forum/go_quote.gif
 
If a null result comes from a mediocre experiment, it means less, but still something. People are saying that you shouldn't be throwing out results as if they don't mean anything. That said, many audio-related studies are not run nearly as well as they could be, and some could use some more rigorous statistical analysis as well.

 
In that case, the last time I checked the NwAv articles they are only referencing "less, but still something".  (such as Matrix Hifi and M&M etc.)
 
Did you miss my comments about the A/D/A loop instead of A/D/A/D/A/D/A^12 when they had so much time on their hands?  The transparency of transducers and the playback system aside, if an A/DC and D/AC are perfectly clean, they'll still be clean after 200 conversions, or 200,000 conversions, right?

 

[mikeaj]

Sorry but this doesn't answer the question.  I didn't say listening tests I just said my hearing, and we can assume I completely don't understand the technology of codecs or any of the data.
 
Now which was more reliable in finding the transparency in 256kbps in this case - the peer reviewed papers or just my hearing alone?

 
Between those two options, in this case it's your hearing.  I would also look at the test conditions, equipment used, and so on.
 

I try to weight everything with it's evidence and likelihood of correctness as much as possible.  If someone says the Clip+ sounds the same as the iBasso DX100 due to RMAA, what do you think?

 
That it's not something RMAA can determine, particularly not with a single run, and looking at the numbers only (not the graphs).  Given enough data with a legitimate audio analyzer and other test gear, if it's demonstrated that two devices are similar enough in enough ways in a given usage scenario—no matter what the devices are or their topology—I could be reasonably convinced that they would sound pretty much the same and maybe be indistinguishable.  If two devices are not close to transparent for practical purposes, it's not too likely that they should be indistinguishable, since there are many ways to deviate from ideal:  it's not particularly likely that two devices would deviate in the exact same ways.
 

Did you miss my comments about the A/D/A loop instead of A/D/A/D/A/D/A^12 when they had so much time on their hands?  The transparency of transducers and the playback system aside, if an A/DC and D/AC are perfectly clean, they'll still be clean after 200 conversions, or 200,000 conversions, right?

 
I did miss that.
 
But what do you mean by "perfectly clean"?  If they're literally ideal with no change at all (impossible for any real-world system), then you can make as many loops as you want.  If one loop is audibly transparent for practical purposes on a given setup, multiple loops will eventually not be.  Errors start compounding until they're audible after a sufficient number.  The number of loops it takes will depend on the listener, listening setup, and so on.  It might be two, or ten.  It would be one for a lower-quality A/D or D/A.
 
I hope you weren't implying that one loop being audibly transparent implies that any number of loops will be?

 

[mike1127]

Quote:

We're talking about sound reproduction, not sound creation, and none of those things [microdynamics, PRaT, and beauty] have anything to do with circuit boards and wires. That's all a product of flesh and blood musicians.
I regularly discuss music in the music forum as well as on my blog and several music forums.

Yes they do -- speakers, microphones, and miking techniques (among other things) either get those things right or mess them up.
 
Still no response to my point about abstracted perception. As far as I can tell, that's the biggest gulf between my experience of sound/music and yours-- that you seem to think it doesn't exist. Not only does it exist, musicians couldn't function without it. If there were no abstracted perception, a musician wouldn't be able to recognize his own technique the first time he plays in a new room. A pianist would fall to pieces the first time they use an unfamiliar piano.
 
I'll check out the music forum.

 

[bigshot]

We're talking about playback, not creating recordings, right? I'm talking about how to make recordings sound good on your home stereo, not how to record and perform music. Those are two different things.

 

[mike1127]

Quote:

We're talking about playback, not creating recordings, right? I'm talking about how to make recordings sound good on your home stereo, not how to record and perform music. Those are two different things.

I always try to talk about the whole chain, the whole system. But even if the discussion is playback only, and even if I grant you modern electronics are 100% transparent, you still have to choose speakers/headphones somehow. I choose headphones that do a good job of getting right the aspects of music that are most important to me, all of which happen to be
 
....
 
here it comes
 
..
 
abstracted perceptions

 

[bigshot]

Speakers and headphones are really the only thing worth discussing. It's tricky too because the best sounding speakers in the store might sound lousy in your living room. I wish I knew more about room acoustics. I do know that equalization helps a LOT.