[StanD]

Exactly. And no one would know better than jj where the weaknesses are. Extensive blind testing was carried out in the development of the codec.

se

"Who knows what evil lurks in the heart of the codec?"

 

[anetode]

Since this is the objectivist lounge, where's the data?

This data set is kept confidential by Head-Fi, though I have a fine collection of anecdotes. Unfortunately we can't discuss either one according to the forum rules.
 

I'd put in an offer to help moderate the sound science subforum... thought it was a pretty long stretch myself but you never know.

One of the reasons given for (what I presume is, for now) my rejection was that "and, being a sponsor, I'm sure the science guys would just LOVE you (not)".

So--what would you guys say to that? Would you have trusted Joe Bloggs not to stamp out all dissenting opinion threatening the portable HiFi status quo represented by FiiO et al?

Or would Joe Bloggs have bent over backwards in over-lenience of the quacks at Science, putting all our livelihoods in jeopardy?

Or would Joe's personal viewpoint and official job position have struck a nice balance cancelling out each other to produce the perfectly impartial Sound Science mod for head-fi?

 
You don't have the requisite ninja training and Curra is not currently accepting any new students

 

[Steve Eddy]

"Who knows what evil lurks in the heart of the codec?"

"The jj knows."

I remember an old Flip Wilson album we had back in the '60s. He did a take on that with the punchline being "The Shadow do."

Here come the judge... Here come the judge... Here come the judge...

se

 

[StanD]

"The jj knows."

I remember an old Flip Wilson album we had back in the '60s. He did a take on that with the punchline being "The Shadow do."

Here come the judge... Here come the judge... Here come the judge...

se

Sometimes I'd like to set Geraldine's boyfriend loose on some of the audio mythologists.

 

[dprimary]

I know the person who developed the codec. I'll ask him.

What are we considering as high bitrate? 320kbps?

se

256 to 320kbs would be the high bit rate. Multiple people developed the codec since they has been a dozen mp3 codec's over the last 20 years. But it would be interesting to know his prospective.

 

[lamode]

  the fact that people fail the abx isn't proof that it's impossible to pass it. that much is sure, anybody who bothered to try and understand abx will acknowledge that.
but we also know a lot about statistics, and we can reliably use stats for what they are. when 50000people fail a test, and 1 comes telling he passed, you know it's possible, but you also have legitimate right to doubt the procedure and ask to know more.

 
That's now how it works. For a given number of 'questions' in an ABX test, there is always a probability of a certain percentage of subjects getting a pass score by purely random chance. For most of the ABX tests we see here, comparing MP3 and lossless, the odds are often in the ballpark of 1%. So out of 50,000 subjects, we would expect around 500 to get a pass score even if they were all picking random answers. The result would not start to become meaningful until a much larger number of subjects passed.

 

[lamode]

I'd put in an offer to help moderate the sound science subforum...

 
Nothing personal, but a sponsor should never be a mod.

 

[cjl]

   
That's now how it works. For a given number of 'questions' in an ABX test, there is always a probability of a certain percentage of subjects getting a pass score by purely random chance. For most of the ABX tests we see here, comparing MP3 and lossless, the odds are often in the ballpark of 1%. So out of 50,000 subjects, we would expect around 500 to get a pass score even if they were all picking random answers. The result would not start to become meaningful until a much larger number of subjects passed.

Alternatively, the person who passed could repeat the test with a stricter criterion - a 20/20 result for example would only have about a 1 in a million chance of occurring by luck, and 30/30 would be about 1 in a billion. A single 30/30 result would be quite convincing, in my opinion...

 

[lamode]

  Alternatively, the person who passed could repeat the test with a stricter criterion - a 20/20 result for example would only have about a 1 in a million chance of occurring by luck, and 30/30 would be about 1 in a billion. A single 30/30 result would be quite convincing, in my opinion...

 
Yes, the 8/10 pass mark that some people here like to use is far too meaningless. Then when some people achieve a pass score (as some inevitably will by pure chance) the audiofools say "See! We told you some people can hear the difference!"
 
30/30 sounds good to me

 

[Steve Eddy]

Sometimes I'd like to set Geraldine's boyfriend loose on some of the audio mythologists.

Sic 'em, Killa!

se

 

[StanD]

Then there are people that can't hear a difference and are willing to admit it.

 

[Steve Eddy]

256 to 320kbs would be the high bit rate. Multiple people developed the codec since they has been a dozen mp3 codec's over the last 20 years. But it would be interesting to know his prospective.

Not sure what you're talking about. There is only the one MP3 standard that I am aware of and it hasn't changed since 1998.

se

 

[castleofargh]

 

  the fact that people fail the abx isn't proof that it's impossible to pass it. that much is sure, anybody who bothered to try and understand abx will acknowledge that.
but we also know a lot about statistics, and we can reliably use stats for what they are. when 50000people fail a test, and 1 comes telling he passed, you know it's possible, but you also have legitimate right to doubt the procedure and ask to know more.

 
That's now how it works. For a given number of 'questions' in an ABX test, there is always a probability of a certain percentage of subjects getting a pass score by purely random chance. For most of the ABX tests we see here, comparing MP3 and lossless, the odds are often in the ballpark of 1%. So out of 50,000 subjects, we would expect around 500 to get a pass score even if they were all picking random answers. The result would not start to become meaningful until a much larger number of subjects passed.

I said "when 50000 people fail a test", not "when 50000 try an abx with p=0.01". not to say you're wrong, just that it's not what I was saying ^_^.
and by saying "you know it's possible", I was pointing at the statistical chances that luck would lead to a positive result. 
I wasn't arguing that no one guy could ever succeed. just pointing at the idea that we're trying to prove a positive, not a negative. so when we get a positive, we need to at least get the method(number of trials, predetermined number of trials ...)to know how seriously we should take the result.

 

[MacacoDoSom]

  Then there are people that can't hear a difference and are willing to admit it.

I can't...

 

[dprimary]

Not sure what you're talking about. There is only the one MP3 standard that I am aware of and it hasn't changed since 1998.

se

There is the standard, but there is LAME and Fraunhofer versions to start. Then you have things like Xing is it a codec or just an encoder using someone else's codec?