Head-Fi.org › Forums › Equipment Forums › Computer Audio › 24/96 Files?
New Posts  All Forums:Forum Nav:

24/96 Files? - Page 2

post #16 of 80
Quote:
Originally Posted by oscarHiFi View Post

I've done my personal testing and I did hear a difference between 16/44.1 and 24/96. It was the same track from the same artist, just different rates. I heard more detail in 24/96 than 16/44.1. The instruments sounded more alive and with even more detail. It all comes down to if you can hear it or not.

 

Did you try this method ?

- take a high resolution 24-bit/96 kHz audio file, let's call it A.wav

- convert A.wav to 16-bit/44.1 kHz with a good resampler (like the SoX sample rate converter plugin of foobar2000); make sure that the output is not clipped, and it should preferably be properly dithered

- convert the low resolution file back to 24-bit/96 kHz, let's call the result B.wav

- make sure that A.wav and B.wav are perfectly (sample accurately) synchronized, are of the exact same length, and that there is no artifact like a pop or click at the beginning or end of one of the files

- compare A.wav and B.wav with the foobar2000 ABX comparator plugin. Do not forget to enable ReplayGain, or make sure that the files are accurately level matched, this is important. Even a fraction of a dB volume difference can give a false positive result

Can you get less than 5% chance reported that you are just guessing, without any cheating ?

 

It is a common mistake to compare different masters of the same music, they will sound different for reasons other than the resolution and sample rate. Also, only with careful blind testing (with matched levels etc.) can you know for sure if you really hear a difference.


Edited by stv014 - 12/20/12 at 11:48am
post #17 of 80

If you don't mind buying physical media, a lot of albums and singles can be found in LP format. There are other formats such as SACD, but I don't think these are as popular, unless it's a remaster. I believe the LPs are usually mastered differently, because I've noticed on sites like http://dr.loudness-war.info/, the dynamic range is different with the vinyl pressings. So that's probably the main or only reason people hear a difference between 16/44.1 and 24/96 (even with digital downloads, they probably come from an LP).

post #18 of 80
CSGO you play CS:GO?

Are you on any leagues? Play ESEA?

My team is rank 6.
post #19 of 80
Thread Starter 
Quote:
Originally Posted by cssarrow View Post

CSGO you play CS:GO?
Are you on any leagues? Play ESEA?
My team is rank 6.


LOL.

I played GO for 1 day.

I don't play CS anymore, but I used to play in ESEA IM for 1.6.

post #20 of 80
Quote:
Originally Posted by Brooko View Post


How did you test? If it wasn't double blind, and volume matched, then it was flawed. Foobar2000 has an abx tool that will allow you to set up a proper test. If you take the original 24/96 master an resample same file to redbook, volume match and abx using 15-20 blind tests - I'd guarantee all those differences you supposedly hear will disappear smily_headphones1.gif

I've forgot about the Foobar2000's abx tool. It interests me a lot.

post #21 of 80
Quote:
Originally Posted by Puranti View Post

There is no difference, in terms of sound quality, between a 16/44.1 file and 24/96 just a way to make people pay more thinking they're getting something better.

The quality is the same and it takes less space on your HDD than a 24/96 file.

 

 

In my experience this simply could not be more false.

 

I do a lot of deedledropping, recording vinyl to digital.  I've have ample opportunity to directly compare the difference between redbook and 96/24 files, both recorded using the exact same signal train, the exact same mastering.  There is definitely a difference.

 

True, the differences can be very subtle.  they manifest themselves in things like sense of the room the recording was made in, the decay of notes, the sense of air and space around each instrument, the depth, breadth and focus of the soundstage, etc.

post #22 of 80
Quote:
Originally Posted by sharkz View Post

 

Unless I am mistaken, all Redbook CD's are 16bit 44.1khz. The only way to go higher with physical media would be SACD's, DVD-A's, or Blue Ray Audio.

 

There's also needledropping...

post #23 of 80
Quote:
Originally Posted by chewy4 View Post

You are correct that these files are usually mastered differently. That's the only audible difference.

 

 

If I make a needledrop of the same LP, sampled at 44.1/16 and also at 96/24, that eliminates that variable.

 

I have done this, many times.

 

There is a difference.

post #24 of 80
Quote:
Originally Posted by nelamvr6 View Post

 

 

If I make a needledrop of the same LP, sampled at 44.1/16 and also at 96/24, that eliminates that variable.

 

I have done this, many times.

 

There is a difference.

 

Interesting - some questions then.

 

Did you do two separate needle drops or was it one at 24/96 and then resampled to 44.1/16?

 

Which resampler?

 

What did you use to abx, and how did you volume match?

post #25 of 80
Quote:
Originally Posted by Brooko View Post

 

Interesting - some questions then.

 

Did you do two separate needle drops or was it one at 24/96 and then resampled to 44.1/16?

 

Which resampler?

 

What did you use to abx, and how did you volume match?

 

 

One needledrop, resampled using Adobe Audition 3, the same software I used to record.  All my flacs have replaygain tags, so volume matching was not a problem.   I did not use any tools to do a formal ABX.  I just listened.

post #26 of 80
Quote:
Originally Posted by nelamvr6 View Post

 

 

One needledrop, resampled using Adobe Audition 3, the same software I used to record.  All my flacs have replaygain tags, so volume matching was not a problem.   I did not use any tools to do a formal ABX.  I just listened.

 

In other words your test was not blind - therefore not valid.  ABX the two tracks using foobar2000's ABX tool.  It's a true blind test.  Run about 15-20 iterations.  You'll find the perceived differences suddenly disappear.  The mind is a wonderful thing wink.gif

 

And just to add - none of us would be able to tell the difference either.  Two tracks from same master at 16/44 and 24/96 - all that's really changed is the noise floor.  Humans can't perceive at the range you're talking about.  So like I said - to our ears, as long as you've introduced no artifacts in the resampling ...... there is no difference.


Edited by Brooko - 12/21/12 at 3:54pm
post #27 of 80
Quote:
Originally Posted by Brooko View Post

 

In other words your test was not blind - therefore not valid.  ABX the two tracks using foobar2000's ABX tool.  It's a true blind test.  Run about 15-20 iterations.  You'll find the perceived differences suddenly disappear.  The mind is a wonderful thing wink.gif

 

And just to add - none of us would be able to tell the difference either.  Two tracks from same master at 16/44 and 24/96 - all that's really changed is the noise floor.  Humans can't perceive at the range you're talking about.  So like I said - to our ears, as long as you've introduced no artifacts in the resampling ...... there is no difference.

 

 

  I'll do an ABX and report back.


Edited by nelamvr6 - 12/21/12 at 4:27pm
post #28 of 80
Quote:
Originally Posted by Brooko View Post

 

In other words your test was not blind - therefore not valid.  ABX the two tracks using foobar2000's ABX tool.  It's a true blind test.  Run about 15-20 iterations.  You'll find the perceived differences suddenly disappear.  The mind is a wonderful thing wink.gif

 

And just to add - none of us would be able to tell the difference either.  Two tracks from same master at 16/44 and 24/96 - all that's really changed is the noise floor.  Humans can't perceive at the range you're talking about.  So like I said - to our ears, as long as you've introduced no artifacts in the resampling ...... there is no difference.

 

I think that  after, say, 30 iterations our hearing system  (ears/brain) is exhausted and therefore we can't always tell the difference.

post #29 of 80
Quote:
Originally Posted by JakeJack_2008 View Post

 

I think that  after, say, 30 iterations our hearing system  (ears/brain) is exhausted and therefore we can't always tell the difference.

 

Ah yes - but all I'm asking for is 15-20 :

  • same volume
  • same master
  • using a resample that did not introduce obvious artifcats
  • done in a controlled manner - true blind
  • performed with complete honesty (there are ways to cheat the system - but it helps nobody)

 

Then post the results.

 

I was one of those a few years ago who would swear I could spot the difference between mp3 320 and lossless, and also redbook and hires.  Once I learned to perform an actual controlled abx, all of the percieved differences disappeared.  I have semi-decent gear.  Ok - I do have older ears, and my hearing is not perfect (low level tinnitus).  But my wife (who can hear a cat walk on carpet at 20-30ft away) has done the same tests and she can't perceive a difference either.  People on this forum - who have far better gear than I - and whose hearing is a lot more acute than mine - cannot tell the difference.

 

So what conclusion can we draw from this - comparing known results with those who claim to be able to tell the difference, but have never performed a controlled blind abx .......?

 

I'm not trying to be a smart ***** with any of these.  I just encourage enlightenment.

post #30 of 80

Well, I admit it was harder that I thought it would be, but I could tell the difference:

 

 

foo_abx 1.3.4 report
foobar2000 v1.1.18
2012/12/21 21:38:02
 
File A: C:\Users\Nat\Music\flacrip\sehr.flac
File B: C:\Users\Nat\Music\flacrip\selr.flac
 
21:38:02 : Test started.
21:38:37 : 01/01  50.0%
21:39:07 : 01/02  75.0%
21:39:40 : 01/03  87.5%
21:40:20 : 02/04  68.8%
21:41:14 : 03/05  50.0%
21:42:06 : 04/06  34.4%
21:43:14 : 05/07  22.7%
21:44:09 : 05/08  36.3%
21:45:11 : 06/09  25.4%
21:46:22 : 07/10  17.2%
21:46:51 : Test finished.
 
 ---------- 
Total: 7/10 (17.2%)
 
 

1000

 

 

Of course, this was music I was very familiar with, on my own system.  I'm not certain I could duplicate the results under other circumstances.

 

But, to me at least, the differences were there.  As I mentioned before, they are quite subtle, but they are there.

 

One thing that helped me no doubt was that I chose the music.  This was a needle drop of Talk Talk's Spirit of Eden.  A LOT of ambient stuff going on!  :D

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Computer Audio
Head-Fi.org › Forums › Equipment Forums › Computer Audio › 24/96 Files?