Pink Floyd - Dark Side of the Moon 2011 blu-ray 96kHz and 2003 SACD spectral analysis

Here are some spectrogram views from blu-ray and SACD with some comments.

Here is spectrogram of stereo channels from 2011 96kHz/24bit stereo mix from the blu-ray:
http://i.imgur.com/ZyNPZ.jpg
I will comment this with the Quad because they are probably sourced the same way.

Here is spectrogram of LR channels from the Quad mix from 2011 blu-ray:
http://i.imgur.com/XcdN3.jpg
Many peaks with not much high frequency noise - this is more than just some old 44.1kHz saved as 96kHz/24bit. This is clearly tape sourced and noise gets quieter with higher frequencies - this is how PCM works. Verdict this was most probably tape > PCM ADC. Both Quad and stereo are PCM sourced.

Here is spectrogram of LR channels from the 5.1 mix from 2003 SACD - converted to PCM @ 88.2kHz with foo_input_sacd:
http://i.imgur.com/YMJbC.jpg
High frequency noise that is inherit in DSD format can be clearly seen. Many players/software use noise shaping to get rid of it. Foo_input_sacd does it too to some extent. This noise is just how DSD works. Especially visible in the beginning of the song where there should be quiet but clearly it is high frequency noise filled.

Here is spectrogram of LR channels from the 5.1 mix from 2011 blu-ray:
http://i.imgur.com/vKnCp.jpg
As you can see this clearly has the SAME NOISE as DSD format!!! Look at the "silence" in the beginning of the song - pure high frequency noise! It is HIGHY probable that this was DSD sourced and converted to PCM @ 96kHz! No noise shaping!

DSD is 44.1kHz based format, it is sampled @ 64 x 44.1kHz = 2.8224MHz or later 128 x 44.1kHz. When converting it to 48kHz base (96kHz = 2 x 48kHz) you are on the hands of the algorithm writers - this is never recommended and is much more difficult operation than resampling from any X times 44.1kHz sample rate. DSD should be convert to 44.1, 88.2 etc PCM for optimal quality.

In conclusion the stereo and quad mixes are PCM sourced and probably "holy grail" versions but 2003 SACD has probably the best 5.1 version when using porper tools to convert to 88.2kHz PCM!

 

Thanks for all that data. How come all the frequency graphs max out at 44.1 kHz? Even the supposed 88 and 96 kHz?

 

What graphs are you looking? The 96kHz graphs max out @ 48kHz because PCM cointains frequency information up to half the sampling rate - http://en.wikipedia.org/wiki/Nyquist_rate
SACD is 88.2kHz PCM so information up to 44.1kHz. 44.1kHz CD contains information up to 22050Hz. This is basic digital sound reproduction knowledge.

 

Does that mean we're better off listening to the 5.1 mix directly from the SACD?

 

Well, I'm not surprised. I doubt there even is an analog tape of the 5.1 mix, so the DSD source is all they have available to them.

 

I gotta admit that I'm a little bit confused...

 

Based on these spectrogram views it is HIGHLY probable that the 5.1 channel 2003 mix on the blu-ray is sourced from the same DSD stream that was used to create the 2003 SACD (with SACD having the original DSD stream and blu-ray having it converted to PCM that is not a lossless and 1:1 conversion and needs care to be done correctly). When you have a SACD player or DSD compatible DAC there is only one conversion DSD>analog audio. Most of todays DAC chips are sigma-delta modulation DAC's so (to keep it simple) basically they make PCM>DSD inside the chip and then to analog audio. With that DAC chip the information path with the blu-ray is (most probably) original 2003 DSD source>96kHz/24bit PCM>DSD again in DAC chip>analog audio! See the multiple conversions?

If you have and older resistor ladder DAC then it plays PCM>analog audio directly. In that case you need an optimal conversion from DSD to PCM and that is always 44.1kHz based (88.2kHz being optimal) and use a good noise shaper. This blu-ray version has it in 96kHz (48kHz base) and no noise shaping.

 

oops, sorry. I see the differences now.

 

where does the quad mix (4.1?) fit in SQ wise? To me, there is tape hiss but sonically first rate.

 

Thanks, nice work.

 

Does it seem to you, as if you're just being used?

 

He could always pack up and fly down south !!

 

I seem to remember that the 2003 5.1 mix was done in pcm, 24/96 then converted to dsd. The stereo however was tape to dsd.

 

The Quad mix is 4.0, not 4.1.

 

No worries. You are doing just fine.

Standby for Wish You Were Here

 

...or hide his head in the sand

 

I'm just another sad old man...

 

All alone... hold on, the threads going off the rails! Topic will be locked soon?

 

But what about the 2003 stereo version??? Isn't DSD better than 9k6kHz PCM? WHy no analysis of this version, it should be a grail even holier.

And not to mention my MFSL vinyl which is the holiest of grails ...

 

Olev, how did you get the DSD off the SACD? I thought that a computer would only read the red-book layer (which is OK in this case as this release is hybrid disc - but if you've got an SACD with no red book layer, I am not aware that a computer can play it at all).

 

This topic explains it in detail:
Backing up SACDs now possible..
http://www.stevehoffman.tv/forums/showthread.php?t=254772

 

You are not alone

I don't know what the hell those spectral analysis mean, frankly.
I know they're quite useful to detect uncompressed v/s MP3-sourced files but other than that ...

 

It recently became possible to extract DSD using a hacked first generation PS3.


It's becoming a more and more popular method, though there are some serious technical hurdles to overcome. There's also a serious shortage of hardware since any PS3 updated past a certain firmware version is essentially broken for this purpose.

Edit: Doh! Missed Averara's post!

 

Frequency response.

 

Thanks for the investigative work.