Comparing DSD Downloads, SACD Sonoma Copies & PS3 SACD Copies

[user 10264]

To a software engineer / computer scientist, this assertion doesn't make sense, i.e., it isn't well-formed. It's digital data. The only perfect copy is the same bits in the same order. Anything else is imperfect, and inferior (unless it's trivially possible to reconstruct the original bit stream from the other copy). So, for example, if a box must do an extra digitization because it's only capable of accepting analog input, then it cannot make a copy as good as a device that precisesly reproduces the bits on the original.

If the Sonoma workistation can't read an SACD, how does it get the input from the SACD? Again, I'm only asking out of curiosity (as I avoid SACD and DSD).

An SACD stores the audio data organised in files along with various metadata. The PS3 is able to create an exact copy of the full filesystem structure. In the Sonoma setup, the SACD is played back in a regular player with a digital output connected to a capture interface on the workstation. When the play button is pressed, the player reads the files on the disc and sends out the all the audio data as one continuous stream which is picked up by the Sonoma workstation's capture card. The link between the player and the workstation is only capable of transporting raw DSD data. No track markers or other metadata is sent. On the workstation, you end up with one long file which must be manually cut if you want a file per track. The digital data captured is exactly what was stored on the disc, but the specific arrangement of bits in tracks and files is not preserved, and all metadata is lost.

Perhaps a useful analogy is to imagine a disc storing a book with front/back cover and chapter/page numbers etc. Copying this book-disc with the Sonoma process would give you all the text, but you'd lose the original pagination and the covers.

 

[Eclectic]

An SACD stores the audio data organised in files along with various metadata. The PS3 is able to create an exact copy of the full filesystem structure. In the Sonoma setup, the SACD is played back in a regular player with a digital output connected to a capture interface on the workstation. When the play button is pressed, the player reads the files on the disc and sends out the all the audio data as one continuous stream which is picked up by the Sonoma workstation's capture card. The link between the player and the workstation is only capable of transporting raw DSD data. No track markers or other metadata is sent. On the workstation, you end up with one long file which must be manually cut if you want a file per track. The digital data captured is exactly what was stored on the disc, but the specific arrangement of bits in tracks and files is not preserved, and all metadata is lost.

Perhaps a useful analogy is to imagine a disc storing a book with front/back cover and chapter/page numbers etc. Copying this book-disc with the Sonoma process would give you all the text, but you'd lose the original pagination and the covers.

Thanks. This makes perfect sense. As a practical matter, it means that the PS3 and the Sonoma are both perfect, with a (very) slight nod to to the PS3, as it gets you metadata as well as data. (The NSA has a particular fondness for metadata .)

 

[grill]

There's an interesting discussion at the Computer Audiophile forum on the differences between DSD Downloads created from:

1. Album Edit Master Tapes (the final mix)
2. Cutting Masters (the final mix with DST compression added)
3. SACD Transport or modified SACD player plus a Sonoma DSD Workstation
4. Modified PS3 Sony Playstation

The resulting DSD music files are most accurate from the final mix/album edit master tapes (level 1) and not as accurate at level 4 - created with a modified PS3 Sony Playstation...

FYI

I did a quick test and I think I can prove that PS3 ripping method is audio data bit perfect for 1) Album Edit DSD Masters. I'll check this out for 2) DST masters, too. My test design is the following:

1) creating a test DSD master source
2) authoring an SACD-R from the source
3) PS3 iso rippng the SACD-R
4) extracting the Edit Master file from the iso using sacd_extract
5) checking and aligning the source and the extracted masters in a Hex editor
6) MD5 CRC comparison of the two masters

Please give me a feedback if you want me to perform and post the full test.

 

[ssully]

Really??

Really.

 

[grill]

FYI

I did a quick test and I think I can prove that PS3 ripping method is audio data bit perfect for 1) Album Edit DSD Masters. I'll check this out for 2) DST masters, too. My test design is the following:

1) creating a test DSD master source
2) authoring an SACD-R from the source
3) PS3 iso rippng the SACD-R
4) extracting the Edit Master file from the iso using sacd_extract
5) checking and aligning the source and the extracted masters in a Hex editor
6) MD5 CRC comparison of the two masters

Please give me a feedback if you want me to perform and post the full test.

I've just finished my test. Based on my findings now I can prove that PS3 iso ripping method is audio data bit perfect for both 1) Album Edit DSD Masters and 2) DST masters. This is also a proof that DST compression is audio data lossless. I'll share the details and all of my test files with you soon.

 

[Jim the Oldbie]

I've just finished my test. Based on my findings now I can prove that PS3 iso ripping method is audio data bit perfect for both 1) Album Edit DSD Masters and 2) DST masters. This is also a proof that DST compression is audio data lossless. I'll share the details and all of my test files with you soon.

Thanks for all your time and effort with this. It's very interesting. I'm really starting to think about trying to snag an old 3.55 PS3 somewhere, so I appreciate every bit of info I can find on this topic.

-- Jim

 

[grill]

Thanks for all your time and effort with this. It's very interesting. I'm really starting to think about trying to snag an old 3.55 PS3 somewhere, so I appreciate every bit of info I can find on this topic.

-- Jim

Thanks and great. If you're not in hurry to jump into the SACD archiving world then I'd suggest you to wait for the completion of Mr Wicked's project posted here: http://www.computeraudiophile.com/f...tween-computers-and-electronic-devices-22124/

 

[grill]

I've just finished my test. Based on my findings now I can prove that PS3 iso ripping method is audio data bit perfect for both 1) Album Edit DSD Masters and 2) DST masters. This is also a proof that DST compression is audio data lossless. I'll share the details and all of my test files with you soon.

I uploaded all my test and result files to https://drive.google.com/open?id=0B1MNXd0OOQsNOWtzdHphTTJiU00 I'm hereby describing my test process and the content of my upload:

1) I created test DSD master sources from a HD Dolby TrueHD Channel Check file (https://www.youtube.com/watch?v=IfaiHl_WJqM). I used Weiss Saracon and Philips DST Encoder for PCM2DSD and DST encoding, respectively.

01a Source\DSD_mch.dff
01a Source\DSD_stereo.dff
01a Source\DST_mch.dff

2) I authored a stereo/multichannel SACD-R from the DSD_stereo.dff and DST_mch.dff files using Philips SuperAuthor. For the sake of simplicity I created only one track for each SACD area. I burned the original SuperAuthor SACD ISO (DAT) onto a DVD-RW using ImgBurn. Before burning ImgBurn warned me that it would change the image track format MODE1/2064 of the SuperAuthor ISO to DVD track format MODE1/2048. Great, this is what I wanted! I also got a warning message from ImgBurn alerting the disc may contain encrypted data. Don't worry, it won't

02 SACD-R\1 Original SuperAuthor ISO\DVDIMAGE.iso
02 SACD-R\2 ImgBurn ISO\DISC1.ISO

3) I ripped the burned SACD-R using a PS3 and Mr Wicked's sacd-ripper/sacd_extract through network. Interestinly my first ripping attempt failed because sacd_extract freezed. The following trick helped: I started ripping a factory pressed SACD for a couple of seconds and my SACD-R rip worked like a charm afterwards.

03 PS3 ISO Rip\PS3 sacd-ripper test SACD.iso
03 PS3 ISO Rip\sacd_log.txt

4) I extracted the stereo and multichannel Edit Master files from the ripped iso using sacd_extract/sacd_extract-gui. I applied DST->DSD decompression on the multichannel track during extraction.

04a Stereo EM DSD extract\PS3 sacd-ripper test SACD.cue
04a Stereo EM DSD extract\PS3 sacd-ripper test SACD.dff
04b Mch EM DSD extract\PS3 sacd-ripper test SACD.cue
04b Mch EM DSD extract\PS3 sacd-ripper test SACD.dff

5) I checked and aligned by trimming the source and the extracted masters in XVI32 Hex Editor (http://www.chmaas.handshake.de/delphi/freeware/xvi32/xvi32.htm). I had to trim off some bytes from the start and the end of the files because they contained different descriptive (non-audio) data. I made screenshots of these parts of the files. In the pictures the blue cursor rectangles show where the trimmed parts started or ended.

01a Source\Hex_mch_end.jpg
01a Source\Hex_mch_start.jpg
01a Source\Hex_stereo_end.jpg
01a Source\Hex_stereo_start.jpg
05a Trimmed Stereo EM DSD extract with XVI32\Hex_stereo_trimmed_end.jpg
05a Trimmed Stereo EM DSD extract with XVI32\Hex_stereo_trimmed_start.jpg
05b Trimmed Mch EM DSD extract with XVI32\Hex_mch_trimmed_end.jpg
05b Trimmed Mch EM DSD extract with XVI32\Hex_mch_trimmed_start.jpg

6) I did MD5 CRC comparisons of the

a) two ISO's (ImgBurn original and the ripped ones)

02 SACD-R\2 ImgBurn ISO\DISC1.md5
03 PS3 ISO Rip\PS3 sacd-ripper test SACD.md5

b) two master pairs (trimmed stereo source/trimmed extracted, trimmed mch DSD source/trimmed extracted and decompressed)

01b Trimmed source with XVI32\DSD_mch.md5
01b Trimmed source with XVI32\DSD_stereo.md5
05a Trimmed Stereo EM DSD extract with XVI32\PS3 sacd-ripper test SACD.md5
05b Trimmed Mch EM DSD extract with XVI32\PS3 sacd-ripper test SACD.md5

All the respective MD5 CRC hashes matched - proof done.

 

[Jim the Oldbie]

Thanks and great. If you're not in hurry to jump into the SACD archiving world then I'd suggest you to wait for the completion of Mr Wicked's project posted here: http://www.computeraudiophile.com/f...tween-computers-and-electronic-devices-22124/

I have been watching that project with interest, although it looks like life has intervened in Mr. Wicked's plans of late...

-- Jim

 

[Jim the Oldbie]

...All the respective MD5 CRC hashes matched - proof done.

Dang, dude! This is some real meat-and-potatoes stuff here. Thanks again for all your time and effort.

-- Jim