[Krutsch]

  To 16 or to 24 that is the Question!  Or what was the question again?
 
You need to scroll back the last 10-20 pages or so for my other posts.  My summary:
 
The OP made a lengthy statement advocating that 16 bit is all that is needed and then invited rebuttals.  I attempted to ABX test 24 bit vs 16 bit tracks to prove it one way or the other.  Unfortunately my tests have not been as conclusive as I thought they would be.  I'm still searching for that one track where I can get 9/10 and really definitely show there is an audible difference in 24.  I took a few week hiatus, RAM is my newest attempt.  I've gone through the HD Tracks sampler, the Linn sampler, and a few other suggestions.  RAM album was suggested by someone way back so I thought I'd have a crack at it.  In short, so far I've got tantalizingly close on a couple of tracks, but on many others it's clearly guessing.

 
So, I'm not a sound scientist, but like to think I am a good active listener. RAM is a great album (in a Grammy winning sort of way), but it's "busy" in that there is a lot going on, sound-wise, and it's kind of loud (IMO, although well recorded).
 
May I make a suggestion? Try something that was originally recorded with/for DSD to start with that will have a mix of loud and quiet passages; it might be easier to find recognizable artifacts to get yourself to the goal of passing the test, so to speak. It's fascinating following your journey here.
 
Anyway, maybe find a recording from the Pentatone catalog (assuming you like Classical) and look for something with enough variation to stretch the dynamic range (like, say, Mahler). I am a huge fan of Pentatone's recordings and they have a great reputation for sound engineering.

 

[bigshot]

I did a listening test with a sound engineer friend of mine using the Pentatone Parvo Jarvi Stravinsky disk. Couldn't tell any difference between the CD layer and the SACD layer. So I ripped it to AAC 256 VBR. Still no difference.
 
Why waste time doing tests that have already been established (except to audiophools)

 

[RRod]

  I did a listening test with a sound engineer friend of mine using the Pentatone Parvo Jarvi Stravinsky disk. Couldn't tell any difference between the CD layer and the SACD layer. So I ripped it to AAC 256 VBR. Still no difference.
 
Why waste time doing tests that have already been established (except to audiophools)

 
I just got that disc. Audiophilia sure is a small world of a small world ^_^

 

[RRod]

   
So, I'm not a sound scientist, but like to think I am a good active listener. RAM is a great album (in a Grammy winning sort of way), but it's "busy" in that there is a lot going on, sound-wise, and it's kind of loud (IMO, although well recorded).
 
May I make a suggestion? Try something that was originally recorded with/for DSD to start with that will have a mix of loud and quiet passages; it might be easier to find recognizable artifacts to get yourself to the goal of passing the test, so to speak. It's fascinating following your journey here.
 
Anyway, maybe find a recording from the Pentatone catalog (assuming you like Classical) and look for something with enough variation to stretch the dynamic range (like, say, Mahler). I am a huge fan of Pentatone's recordings and they have a great reputation for sound engineering.

 
The thing is that even something like Mahler done up by Telarc isn't even using all the dynamic range even of the CD, and already it's about all I can stand before needing to adjust the pot mid-symphony.

 

[Stillhart]

I'm really confused.  It looks like he's sampled more than enough to prove that there's no difference.  Is he being disingenuous?  In other words, am I missing out on the obvious sarcasm?

 

[Krutsch]

  I did a listening test with a sound engineer friend of mine using the Pentatone Parvo Jarvi Stravinsky disk. Couldn't tell any difference between the CD layer and the SACD layer. So I ripped it to AAC 256 VBR. Still no difference.
 
Why waste time doing tests that have already been established (except to audiophools)

 
Simply put: @Greenears wants to prove the assertion on its own merit, and I think some credit is deserved for the effort. I've always wanted to do a similar exercise, but I just don't have the time now. So, I thought I'd add my $0.02 to the discussion and recommend music that may or may not help.

 

[RRod]

  I'm really confused.  It looks like he's sampled more than enough to prove that there's no difference.  Is he being disingenuous?  In other words, am I missing out on the obvious sarcasm?

 
He's first operating under the hypothesis that there is a difference, and trying to hear what a positive test would sound like. There are in fact audible differences between 16 and 24-bits IF you jack up the volume to insane enough levels to hear near the noise floors, but no one listens to music this loudly. Hence the impossibility in finding an actual music track where one can detect a difference using a properly setup ABX test.

 

[ascension278]

Used Hugo + Roxannes to sample the tracks.

Blind testing results:
 
16/44.1
Songs play at a higher volume (need to be cranked up to have equal levels)
Songs have a less darker background. Somewhat grainier and more congested

24/96
Volume has to be cranked up
Songs have darker backgrounds
Song is more effortless
 
17 times out of 20 I could identify the 24bit  songs
 
* Using a variety of songs. But the 16/44 and 24/96 songs are different.

 

[RRod]

  Used Hugo + Roxannes to sample the tracks.

Blind testing results:
 
16/44.1
Songs play at a higher volume (need to be cranked up to have equal levels)
Songs have a less darker background. Somewhat grainier and more congested

24/96
Volume has to be cranked up
Songs have darker backgrounds
Song is more effortless
 
17 times out of 20 I could identify the 24bit  songs
 
* Using a variety of songs. But the 16/44 and 24/96 songs are different.

 
A proper testing framework would compare the same track at the two specifications (with the lower spec upsampled to prevent DAC hints), volume matched, ideally made from the same source as to avoid difference in mastering (say from a CD vs. an HD download).

 

[Head Injury]

  Used Hugo + Roxannes to sample the tracks.

Blind testing results:
 
16/44.1
Songs play at a higher volume (need to be cranked up to have equal levels)
Songs have a less darker background. Somewhat grainier and more congested

24/96
Volume has to be cranked up
Songs have darker backgrounds
Song is more effortless
 
17 times out of 20 I could identify the 24bit  songs
 
* Using a variety of songs. But the 16/44 and 24/96 songs are different.

If you're getting large volume differences you're comparing different masterings or your 16/44.1 encode is borked. If you're changing the volume manually you're very likely screwing up the level matching, and they'll sound different as a result.
 
If you're not comparing with the same songs, from the same mastering, and level matched, the results aren't going to be conclusive.

 

[bigshot]

Actually it's most likely that your equipment is borking the high sample rate files.

 

[headwhacker]

I'm the king of getting in trouble for that.

I doubt that coz you are still gere

 

[limpidglitch]

Bork, bork, bork.
 

 

[Greenears]

  Used Hugo + Roxannes to sample the tracks.

Blind testing results:
 
16/44.1
Songs play at a higher volume (need to be cranked up to have equal levels)
Songs have a less darker background. Somewhat grainier and more congested

24/96
Volume has to be cranked up
Songs have darker backgrounds
Song is more effortless
 
17 times out of 20 I could identify the 24bit  songs
 
* Using a variety of songs. But the 16/44 and 24/96 songs are different.

As has been the practice on this thread, please post your ABX logs and file information so that others can reproduce your results.  If you refuse to post logs, your results will be discounted.
 
To clarify the methodology a number of us agreed to, you need to downsample the 24 bit file to 16 bit and ABX that.  It doesn't mean you can't also ABX a published 16 bit version and provide two logs.  The reason for the downsample is to control for any possibility that the published 16 bit file has different mastering than the 24 bit.  By simply chopping the last 8 bits you eliminate mastering as a possible difference (and also in all of our experience the levels match perfectly).

 

[Greenears]

   
So, I'm not a sound scientist, but like to think I am a good active listener. RAM is a great album (in a Grammy winning sort of way), but it's "busy" in that there is a lot going on, sound-wise, and it's kind of loud (IMO, although well recorded).
 
May I make a suggestion? Try something that was originally recorded with/for DSD to start with that will have a mix of loud and quiet passages; it might be easier to find recognizable artifacts to get yourself to the goal of passing the test, so to speak. It's fascinating following your journey here.
 
Anyway, maybe find a recording from the Pentatone catalog (assuming you like Classical) and look for something with enough variation to stretch the dynamic range (like, say, Mahler). I am a huge fan of Pentatone's recordings and they have a great reputation for sound engineering.

I'm willing to try, even though not a big Mahler fan.  Where can I get this Pentatone recording legitimately in the US? (We can't access Qobuz yet).
 
Note I've already gone through 10 tracks of the HD Tracks sampler, 2 of the Linn sampler, and I have the full RAM album in 24 bit but only really tried one track so far.  I haven't had any success with louder passages so far, it tends to be on more moderate relative volme or single instrument sections.  Depsite the fact that initially I thought I could hear differences on some loud sections.