Ripping SACDs the right way

[tailspn]

Except that's not the case. Look at the screenshot again. The fade-in starts at the same point as the music on the PS3 track.

Believe me, that's exactly what's happening on post 191. The two pair of stereo tracks are not time aligned on that Pyramix screen shot. Look closely and compare the even number and odd number tracks, they're not identical after the Sonoma fade point. I have no clue what Bruce did in this example to time align the stereo pair of tracks, but the Sonoma rip either is affected by a fade-in from the transport (my Meitner doesn't do that), or more likely, the start marker is grossly misplaced to be at the second music repeat. But the fact remains that the PS3/SACD_extract combo tracks start and stop at the position of the track PQ markers.

 

[miguelito]

It seems the folks who use the PS3 have a problem with it. I'm totally satisfied, as are my clients, with the results, including gapless playback with no artifacts.

I don't know... I do know that some players don't do gapless properly regardless of the source of the dsf files. In particular downloaded dsf. But unfortunately I don't have an example in mind of a downloaded dsf which does not properly do gapless in a particular software player. Perhaps someone could chip in.

And if the PS3 rips do have issues (like the non-zero crossing you mentioned) then that is likely possible to fix in the ISO->DSF process.

 

[miguelito]

I figured I'd test this, so I ripped a random CD using cdparanoia and no special options to a wav file. Then I played the same CD in a PS3 (I don't have any other CD transport) while capturing the optical S/PDIF output. After aligning the start, every last bit matches. That's what I expected.

Fabulous! That is ALL I want to see!!!

Please someone with a PS3 rip and a Sonoma workstation do the same from the same exact disc?

(I realize that religious objections might make this set empty... )

 

[miguelito]

You explained the start/stop markers. That does not explain why the Sonoma capture has a fade-in at the very start of the first track where the PS3 rip doesn't.

I VERY much doubt this comes from the transport. It would require fairly ellaborate massaging of the data stream. Why would a transport do this (mute/not mute is not fade in/out).

 

[miguelito]

Track start and stop timings for SACD rips performed by PS3 and SACD_extract (or similar software), are determined by the PQ markers embedded in the cutting master for SACD production. Those markers are intended to run the SACD player display metadata, and have no actual conection to the music content, other than where the mastering engineer placed them. Normally that's about one second prior to music start. It's simply a situation of where the extract software used to facilitate automatic track slicing of the continuous album file uses data (PQ markers) not intended for that purpose.

Sonoma track starts and stops are determined manually by when the Record button is hit. The SACD transport is free running, and has no control of Sonoma's operation. So therefore some time buffer amount is usually built-in. Like I start the Sonoma's Record, then hit Play on the transport.

So in Bruce's examples, some PQ Marker(s) are within the music content, and therefore the track data prior was not included in the track by the SACD_extract software. Also, PQ markers are both Start and Stop types. Most DSD edited masters I work with have an album beginning Start marker, then more Start markers at the beginning of each new track, and a Stop marker at the end of the last track. But sometimes the mastering guy put a Stop marker at a track end, followed by an inter track time, then a Start marker at the new track beginning. I have no idea how SACD_extract treats these occasional Stop markers, and whether it cuts out the time between the Stop and succeeding Start marker when slicing.

Makes perfect sense. Doesn't sound like rocket science to device an sacd_extract routine that behaves exactly like the Sonoma. From your explanation I also understand why the mastering engineer might conceivably be less than careful about where the markers are placed, especially in a gapless album.

Question: When I skip to the next track how does the player know where to start? Using these markers?

 

[tailspn]

Question: When I skip to the next track how does the player know where to start? Using these markers?

I don't know, but I believe so. Bruce would certainly know. I do know from my Pyramix and Sonoma experience that the PQ markers are common to both the CD and SACD layers. I'm unfamiliar though how and where an SACD player control system obtains its information.

 

[mansr]

Believe me, that's exactly what's happening on post 191. The two pair of stereo tracks are not time aligned on that Pyramix screen shot. Look closely and compare the even number and odd number tracks, they're not identical after the Sonoma fade point. I have no clue what Bruce did in this example to time align the stereo pair of tracks, but the Sonoma rip either is affected by a fade-in from the transport (my Meitner doesn't do that), or more likely, the start marker is grossly misplaced to be at the second music repeat. But the fact remains that the PS3/SACD_extract combo tracks start and stop at the position of the track PQ markers.

The tracks are not exactly identical, but that could simply be follow-on effects from the fade. We know nothing about the filters used to calculate the waveform for display. I don't see how the tracks in the screenshot could be aligned any better than they are.

 

[mansr]

Fabulous! That is ALL I want to see!!!

Please someone with a PS3 rip and a Sonoma workstation do the same from the same exact disc?

(I realize that religious objections might make this set empty... )

Note that this was a plain old Redbook CD, not an SACD.

 

[bonzo75]

Hi, so the start points aside, are the sonic qualities of the main track the same on both PS3 and Sonoma?

 

[mansr]

I VERY much doubt this comes from the transport. It would require fairly ellaborate massaging of the data stream. Why would a transport do this (mute/not mute is not fade in/out).

I have no idea why a transport would do that. I have no idea why anything would do that, but something did.

 

[miguelito]

I don't know, but I believe so. Bruce would certainly know. I do know from my Pyramix and Sonoma experience that the PQ markers are common to both the CD and SACD layers. I'm unfamiliar though how and where an SACD player control system obtains its information.

Actually, it can't be this way I don't think since if I skip to the next track it would have to read the whole thing until it finds the marker! When you insert a disc in a CD/SACD player the machine reads some sort of directory structure from the most inner tracks. It's probably here that it gets information on where each track begins. This info might not match the marker positions I suppose.

 

[miguelito]

I have no idea why a transport would do that. I have no idea why anything would do that, but something did.

Maybe it was in the track to start with, the PS3 dsf just started later due to a belated marker? Or maybe it's Sonoma post processing?

 

[bmoura]

Hi, so the start points aside, are the sonic qualities of the main track the same on both PS3 and Sonoma?

Not on the example shown earlier by Bruce Brown. The waveforms are very different.

 

[miguelito]

Not on the example shown earlier by Bruce Brown. The waveforms are very different.

How can this be??? Really... How can this be???

 

[mansr]

Actually, it can't be this way I don't think since if I skip to the next track it would have to read the whole thing until it finds the marker! When you insert a disc in a CD/SACD player the machine reads some sort of directory structure from the most inner tracks. It's probably here that it gets information on where each track begins. This info might not match the marker positions I suppose.

It is obviously possible to create a disc where the table of contents disagrees with the embedded markers. Although that would probably be a spec violation, there's almost certainly some such disc out there. I wouldn't worry about it though.

 

[mansr]

Not on the example shown earlier by Bruce Brown. The waveforms are very different.

They look very similar to me apart from the missing inter-track gaps. Differences seen in the waveforms in the screenshots are small enough to be rounding errors in the rendering.

 

[tailspn]

Actually, it can't be this way I don't think since if I skip to the next track it would have to read the whole thing until it finds the marker! When you insert a disc in a CD/SACD player the machine reads some sort of directory structure from the most inner tracks. It's probably here that it gets information on where each track begins. This info might not match the marker positions I suppose.

The track data (PQ markers and metadata) are independent of the .dff/.dsf music data, and are treated and stored in the player as TOC information for the entire album, not as read along side the music. So the player knows from loading the SACD where the various track locations exist.

 

[miguelito]

It is obviously possible to create a disc where the table of contents disagrees with the embedded markers. Although that would probably be a spec violation, there's almost certainly some such disc out there. I wouldn't worry about it though.

I don't worry about it except this means the motivation to put proper markers is not strong, which means tools that rely on markers rather than TOC will get the wrong track start/ends.

There's tons of spec violations out there. Hidden tracks are surely not in the spec.

 

[Yuri Korzunov]

Because the fade-in occurred prior to the track start marker on the SACD, so therefore wasn't included in the PS3/SACD_extract track. That was the point of my explanation.

Fade-in/out always DSD editing operation. I.e., for this case read from SACD will not bit-perfect.

I suppose SACD audio player device do any fading after conversion to analog.

I don't think that PS3 do it.

 

[mansr]

Fade-in/out always DSD editing operation. I.e., for this case read from SACD will not bit-perfect.

I suppose SACD audio player device do any fading after conversion to analog.

I don't think that PS3 do it.

This is supposed to be a digital capture.