[zoltankr]

Hello,
I purchased a lossless Van Halen album, and I would like to know if it's real.
I checked the FLAC files with auCDtect. The result:
https://dl.dropbox.com/u/1924024/aucdtect.jpg
 
It looks like a CDDA, but I don't understand why the spectogram is low at 22khz? (at all songs)
https://dl.dropbox.com/u/1924024/spectro.jpg
 
I would be grateful for any suggestion. Thank you!

 

[Jupiterknight]

Quote:

Hello,
I purchased a lossless Van Halen album, and I would like to know if it's real.
I checked the FLAC files with auCDtect. The result:
https://dl.dropbox.com/u/1924024/aucdtect.jpg
 
It looks like a CDDA, but I don't understand why the spectogram is low at 22khz? (at all songs)
https://dl.dropbox.com/u/1924024/spectro.jpg
 
I would be grateful for any suggestion. Thank you!

 
I have CD of this album and have ripped it to FLAC. I checked the same song "Running With The Devil" and my spectogram looks exact the same as the one you have posted. Frequency cutoff at 21.7 KHz which is quite normal for lossy files. If it had been a lossless file around 320 kbps, you would typically have seen a frequency cutoff at around 18 KHz. A 128 kbps file even lower.
 
So I believe that your album is a true lossless

 

[mich41]

[ATTACHMENT=888]chinfonia.zip (4,193k. zip file)[/ATTACHMENT]

chinfonia.wav - excerpt from the original recording
chinfonia.aac - same thing compressed to 120kb/s AAC
chinfonia.aac.wav - above AAC file decompressed to wav

Enjoy

auCDtect: CD records authenticity detector, version 0.8.2
Copyright (c) 2004 Oleg Berngardt. All rights reserved.
Copyright (c) 2004 Alexander Djourik. All rights reserved.
------------------------------------------------------------
Processing file: [chinfonia.wav]
Data analysis: [100%]
------------------------------------------------------------
This track looks like CDDA with probability 100%

------------------------------------------------------------
Processing file: [chinfonia.aac.wav]
Data analysis: [100%]
------------------------------------------------------------
This track looks like CDDA with probability 100%

 

[zoltankr]

Quote:

 
I have CD of this album and have ripped it to FLAC. I checked the same song "Running With The Devil" and my spectogram looks exact the same as the one you have posted. Frequency cutoff at 21.7 KHz which is quite normal for lossy files. If it had been a lossless file around 320 kbps, you would typically have seen a frequency cutoff at around 18 KHz. A 128 kbps file even lower.
 
So I believe that your album is a true lossless

 
Thank you! Your time is greatly appreciated.
That album was released in 1978. I have the remastered version (Warner Bros). Is this frequency cutoff normal for older album? (I mean music before 90's)

 

[zoltankr]

Quote:

 

chinfonia.wav - excerpt from the original recording
chinfonia.aac - same thing compressed to 120kb/s AAC
chinfonia.aac.wav - above AAC file decompressed to wav

Enjoy

Code:

 auCDtect: CD records authenticity detector, version 0.8.2 Copyright (c) 2004 Oleg Berngardt. All rights reserved. Copyright (c) 2004 Alexander Djourik. All rights reserved. ------------------------------------------------------------ Processing file: [chinfonia.wav] Data analysis: [100%] ------------------------------------------------------------ This track looks like CDDA with probability 100% ------------------------------------------------------------ Processing file: [chinfonia.aac.wav] Data analysis: [100%] ------------------------------------------------------------ This track looks like CDDA with probability 100%

What does that mean, that auCDtect isn't reliable?

 

[Jupiterknight]

Quote:

 
Thank you! Your time is greatly appreciated.
That album was released in 1978. I have the remastered version (Warner Bros). Is this frequency cutoff normal for older album? (I mean music before 90's)

Remastered HDCD from 2008 is the one I have, so it must be the same as yours. Yeah, the frequency cutoff around 21-22K is pretty much normal for older albums as well. Although I have seen some albums with lower cutoff. The worst I have seen is probably as low as 16K (although not the whole album) and we are talking about a lossless cd rip.

 

[mich41]

What does that mean, that auCDtect isn't reliable?

Kinda. It seems to rely on high frequency cutoff detection and is easily fooled by configuring the encoder to preserve high frequencies even when it doesn't make sense. But it also appears to correctly identify files compressed with default settings and such files are probably more common "in the wild".

I just wanted to show that comparing spectrograms isn't a very reliable way of telling lossy files apart. There are lossy files without treble cutoff and, as Jupiterknight already pointed out, there are genuine recordings with little treble content.

 

[Jupiterknight]

I converted the aac 128 kbps to a wav file in dBpoweramp and in this case it showed, what I assume would be the correct cutoff for low lossy file, somewhere around around 14K.
 
 I also converted the same AAC file using Foobar2000, which Mitch41 directed me to use or try out, and I got the same result as he did. Converting the aac 128 kbps to a wav file, thereby creating a "fake" wav file that the spectrogram didn't recognize or expressed more correctly, it showed a similar cutoff as the original wav file. 
 
Tried the same procedure with an Lame mp3 file in Foobar, but this time the output disguised as a wav file measured as it should.  So it does seem possible to create a fake lossless file and when using spectrogram etc. it does looks like a real lossless file, but apparently depending on the type converter and the file format involved. 
 
I'm aware of that this topic may not quite belong in the music forum, but it shows that our eyes can't be trusted completely or the most common measuring tools at least.  The "fake" wav file does not sound like the original, so in this case our ears are more reliable, that's music...