DSD Capability

[Owen Smith]

I
But the Oppo 103 does allow for 4.0 playback, 5.1 with no center speaker or sub is quad! With no center speaker that output is supposed to go to both front speakers. Likewise when 5.1 is selected 7.1 back channels are mixed with the surrounds. Yes it says that it may discard the audio if redirection is not possible. Why would it not be possible? I suspect it can't mix DSD the same as it can PCM. IMHO the mixing should be done via analog circuits for the analog outputs!

Analog mixing would cost money to add, and there are quite a few different types of downmix that would need to be supported.

Try selecting playing DSD as PCM, you may find it now downmixes correctly.

I find Tubular Bells SACD plays from the wrong speaker positions as DSD, whereas if I select DSD as PCM it plays out of the correct speaker positions. Given I can detect no audible difference I have DSD as PCM permanently enabled.

 

[werfpsa]

So I gave my SACD rip of Scorpions & Berliner Philharmoniker - "Moment Of Glory" another listen and determined that the centre channel does not mix down to the front outputs even though the player is set up for four speakers only. The vocals are still present at the centre output jack. I haven't noticed a problem with any other rips or SACD's. I don't know if the other SACDs that I've auditioned just aren't making much use of the centre channel or if it's something to do with this particular rip. I don't own the disc but would purchase a copy if I could find one at an affordable price, so I don't know if it has anything to do with the rip itself.

Could it be that the mixing is done in the digital domain which works fine for PCM, but that such mixing is not possible with DSD? I suppose that the next step would be to switch SACD playback to PCM instead of DSD to see what happens via downmixing. If the player won't mix DSD then I'm forced back to an analog solution!

DSD (DSF and DFF) are one bit signals compared to PCM which is 16, 24 or 32 bit. Audio manipulation (Speaker Management and/or Digital Signal Processing) can only be done with PCM signals. So if you have Speaker Management active with 4 speakers, this is done with PCM. So your player should convert the DSD signal to PCM and process the signal accordingly. You'll miss the advantages of the DSD signal.....
You should always use (Pure) Direct on your player/receiver with DSD sources to take the advantage of the one bit signal....

 

[par4ken]

DSD (DSF and DFF) are one bit signals compared to PCM which is 16, 24 or 32 bit. Audio manipulation (Speaker Management and/or Digital Signal Processing) can only be done with PCM signals. So if you have Speaker Management active with 4 speakers, this is done with PCM. So your player should convert the DSD signal to PCM and process the signal accordingly. You'll miss the advantages of the DSD signal.....
You should always use (Pure) Direct on your player/receiver with DSD sources to take the advantage of the one bit signal....

Agreed, that's where the problem lies. The only solution would be to mix the analog outputs myself.

Analog mixing would cost money to add, and there are quite a few different types of downmix that would need to be supported.

Try selecting playing DSD as PCM, you may find it now downmixes correctly.

I find Tubular Bells SACD plays from the wrong speaker positions as DSD, whereas if I select DSD as PCM it plays out of the correct speaker positions. Given I can detect no audible difference I have DSD as PCM permanently enabled.

I can hear a big difference between DSD and PCM, so setting the player to PCM for SACD playback is not a solution for me. While analog mixing would of added cost (not really that much IMHO) not a lot of people would of required it, I suppose. I'm unsure why some .dsf files that show a strong center channel, still sound fine played via the Oppo, the vocals seem to be all there! I'm going to do more investigating (testing) on this. In the end an analog solution will likely be the only real way to go!

 

[marpow]

DSD (DSF and DFF) are one bit signals compared to PCM which is 16, 24 or 32 bit. Audio manipulation (Speaker Management and/or Digital Signal Processing) can only be done with PCM signals. So if you have Speaker Management active with 4 speakers, this is done with PCM. So your player should convert the DSD signal to PCM and process the signal accordingly. You'll miss the advantages of the DSD signal.....
You should always use (Pure) Direct on your player/receiver with DSD sources to take the advantage of the one bit signal....

All my SACD rips I do as a DFF file that plays the 1 bit DSD. JRiver is my player through the Exasound38 DAC. I use Pure Direct on the AVR. Thanks for explaining why I do this. I like the way it sounds but I never knew why, so your explanation is good for me.

 

[Owen Smith]

PCM of sufficient sample rate and bit depth can more than contain DSD resolution. 24 bit 96 KHz is more than enough for SACD bit rate DSD. If you are hearing any difference it is either not really there (since this isn't a double blind test) or your DAC is processing them differently in some manner that it should not. When done properly the conversion in the digital domain ie. to PCM, should produce identical analogue output to direct DSD out of the DAC. Any other result is an error somewhere, either in the DSD to PCM conversion or in the way the DAC renders the DSD or PCM or both. The mathematics do not lie, this is sampling theorem.

 

[ted_b]

PCM of sufficient sample rate and bit depth can more than contain DSD resolution. 24 bit 96 KHz is more than enough for SACD bit rate DSD. If you are hearing any difference it is either not really there (since this isn't a double blind test) or your DAC is processing them differently in some manner that it should not. When done properly the conversion in the digital domain ie. to PCM, should produce identical analogue output to direct DSD out of the DAC. Any other result is an error somewhere, either in the DSD to PCM conversion or in the way the DAC renders the DSD or PCM or both. The mathematics do not lie, this is sampling theorem.

Sorry but this is just wrong. Converting DSD to PCM is lossy process, period. Your dac might present something like 24/176 (often chosen as close to DSD64) as sonically similar or maybe indistinguishable sonically but it’s a lossy process.

 

[Owen Smith]

Sorry but this is just wrong. Converting DSD to PCM is lossy process, period. Your dac might present something like 24/176 (often chosen as close to DSD64) as sonically similar or maybe indistinguishable sonically but it’s a lossy process.

Have you read up on sampling theorem? What matters here is the analogue signal that comes out of the DAC. And within the range of human hearing that is the same. What differs is how DSD's dirty little secret of ultrasonic noise is dealt with, but we can't hear any of that.

 

[par4ken]

Have you read up on sampling theorem? What matters here is the analogue signal that comes out of the DAC. And within the range of human hearing that is the same. What differs is how DSD's dirty little secret of ultrasonic noise is dealt with, but we can't hear any of that.

I do know that they sound different.

 

[Owen Smith]

I do know that they sound different.

Which could be that the DSD to PCM conversion is being done incorrectly, or that the DAC is treating the DSD and PCM differently in ways it should not be. They can be made to sound different.

Now if you prefer the DSD rendition by all means listen to it.

 

[LuvMyQuad]

I do know that they sound different.

This is hard to substantiate without a setup that allows AB testing. And that kind of setup usually isn't available with most home based systems. You really cant do A/B testing with disks. And if your statement is based on listening tests other than A/B, no one should take it seriously anyway.

Back when I started ripping disks I had to make a decision on what to do with the SACD format. Keep it as DSD or convert it to PCM. So I did some A/B testing across a half dozen or so titles. I was only able to do that by ripping the DSD and converting it to PCM (@ 24/88.2) and A/Bing the original disk against the ripped file. The processor allowed me to define presets for volume and input source and I could switch between the disc and the file instantaneously. I would offset the track start time so there was about a 10 second overlap in playback which would allow me to hear a snippet of music from one source followed by the same snippet repeated from the other source. Even this was not a true A/B test by any means. The disc was being decoded with the DAC in my Oppo while the PCM data stream was being processed by my pre/pro. And it certainly wasn't blind testing either. It wasn't ideal but it was the best I could do with my home equipment and i felt it was good enough for my needs. Not I, nor either of the other two listeners I used for the testing could detect a significant difference. I started converting everything to PCM from that point forward. Take it for what it's worth.

Sorry but this is just wrong. Converting DSD to PCM is lossy process, period.

True. By definition it isn't bit perfect by any stretch, so it is lossy. But probably no more so than the conversion that was done to change the original PCM stream to DSD when the SACD was mastered.

 

[marpow]

This is hard to substantiate without a setup that allows AB testing. And that kind of setup usually isn't available with most home based systems. You really cant do A/B testing with disks. And if your statement is based on listening tests other than A/B, no one should take it seriously anyway.

Back when I started ripping disks I had to make a decision on what to do with the SACD format. Keep it as DSD or convert it to PCM. So I did some A/B testing across a half dozen or so titles. I was only able to that by ripping the DSD and converting it to PCM (@ 24/88.2) and A/Bing the original disk against the ripped file. The processor allowed me to define presets for volume and input source and I could switch between the disc and the file instantaneously. I would offset the track start time so there was about a 10 second overlap in playback which would allow me to hear a snippet of music from one source followed by the same snippet repeated from the other source. Even this was not a true A/B test by any means. The disc was being decoded with the DAC in my Oppo while the PCM data stream was being processed by my pre/pro. And it certainly wasn't blind testing either. It wasn't ideal but it was the best I could do with my home equipment and i felt it was good enough for my needs. Not I, nor either of the other two other listeners I used for the testing could detect a significant difference. I started converting everything to PCM from that point forward. Take it for what it's worth.


True. By definition it isn't bit perfect by any stretch, so it is lossy. But probably no more so than the conversion that was done to change the original PCM stream to DSD when the SACD was mastered.

Hi Luv, I understand your statement and it makes sense. But I didn't get the why?

If there is not a significant listening difference to you, why do you choose PCM over DSD?

I rip my SACD's from the OPPO 105 with the Sonore key. Always to a DFF file which plays back as DSD64 1 bit. I like it.
I seem to hear all my DSD rips and my DSD downloads as a little fuller in sonics, less thin as I do with even a 24/192 FLAC PCM file.
The other reason I keep them at DSD is when I rip the SACD it is just the one step process, I don't have to re convert to PCM.
So, I like the ease of keeping it DSD and I think the DSD sounds a hair better.

 

[LuvMyQuad]

Hi Luv, I understand your statement and it makes sense. But I didn't get the why?

If there is not a significant listening difference to you, why do you choose PCM over DSD?

I rip my SACD's from the OPPO 105 with the Sonore key. Always to a DFF file which plays back as DSD64 1 bit. I like it.
I seem to hear all my DSD rips and my DSD downloads as a little fuller in sonics, less thin as I do with even a 24/192 FLAC PCM file.
The other reason I keep them at DSD is when I rip the SACD it is just the one step process, I don't have to re convert to PCM.
So, I like the ease of keeping it DSD and I think the DSD sounds a hair better.

Definitely good question.

Going back to the time this all started I explored JRiver, Kodi, and Foobar for use as music players. For me, Kodi had the best interface hands down. It still does. But it could not process DSD (i think its coming though). JRiver could process DSD, but it had the poorest interface of them all. It still does, for my purposes anyway. I'm not sure if there was a DSD plugin for Foobar at that time.

Although you can keep metadata with Sony's DSD file format, it is not as in depth with how you easily and fully can tag FLAC files. Kodi needs the tags.

Also, I ran into issues with a loud pop occurring at the start of every DSD track that played through JRiver. I managed to reduce it somewhat, but never actually eliminated it. I believe it had something to do with an early version of the sonare software or the early version of JRiver I was using, or both.

At any rate, once I decided I couldn't hear a difference, it became a moot point. I did still save all the SACD ISOs however, in case I ever decide to revert back. Maybe Ill try it when Kodi can handle DSD.

 

[marpow]

Definitely good question.

Going back to the time this all started I explored JRiver, Kodi, and Foobar for use as music players. For me, Kodi had the best interface hands down. It still does. But it could not process DSD (i think its coming though). JRiver could process DSD, but it had the poorest interface of them all. It still does, for my purposes anyway. I'm not sure if there was a DSD plugin for Foobar at that time.

Although you can keep metadata with Sony's DSD file format, it is not as in depth with how you easily and fully can tag FLAC files. Kodi needs the tags.

Also, I ran into issues with a loud pop occurring at the start of every DSD track that played through JRiver. I managed to reduce it somewhat, but never actually eliminated it. I believe it had something to do with an early version of the sonare software or the early version of JRiver I was using, or both.

At any rate, once I decided I couldn't hear a difference, it became a moot point. I did still save all the SACD ISOs however, in case I ever decide to revert back. Maybe Ill try it when Kodi can handle DSD.

OK, that makes perfect sense and I knew it was something like that.
We tend to forget, and sometimes if we are new to this world, that there are reasons for what we do.
For those who are reading, Luv's answer is perfect, and we all have to remember is to try different things.
I use JRiver and never have experienced those issues. I do see others using Foobar, etc and think I really like those cool Black Knight interfaces, and wish I had that, plus JRiver is practically a job in itself as it does so much, but I use it as a push and play and works always fine.

 

[Kal Rubinson]

If there is not a significant listening difference to you, why do you choose PCM over DSD?

Because I can apply DSP to the PCM (for various purposes) but not to DSD.

Always to a DFF file which plays back as DSD64 1 bit.

So, why DFF and not DSF?

 

[par4ken]

I would have to suggest that if you can't hear a difference between native DSD and DSD converted to PCM perhaps you are using an amplifier that uses DSP. Digital signal processing requires conversion to PCM. Any benefits of DSD would be lost.

Theoretically you would think that you could convert to PCM and not be able to tell the difference, but you can tell (and it's not subtle) at least the way it's implemented by the Oppo.

Again I'm not suggesting that DSD is better, the best sounding discs seem to be some of the Blu-ray audio's., they have a silky smooth character. SACD's have a rougher more dynamic sound to them, much of that sonic character is lost when converted to PCM.

 

[marpow]

Because I can apply DSP to the PCM (for various purposes) but not to DSD.

So, why DFF and not DSF?

Great question, I don't know why. I went back in my JRiver and I have 308 DFF files and 7779 files of DSF.
I am not sure how some where ripped to DFF and others to DSF. But obviously I need to correct myself and facts are I prefer DSF according to my library.

 

[ted_b]

DFF and DSF are both proper DSD files (thank you Sony and Philips), but only DSF supports tagging correctly. Converting DFF to DSF is a minimal (header in file) process and worthwhile if you want to catalog properly. Moreover, some of the pops and clicks that some dacs have with starting or stopping a DSF track (cuz of the way they read and process non-zero, dc offsets that sacd-extract can often produce...oversimplified) are often not there with DFF tracks. Test for your own dac.

 

[marpow]

Exploring my personal library and I looked at my DFF vs my DSF SACD rips.
All my DFF rips just happen to be Mobile Fidelity.
All my DSF rips are all the rest of the SACD's with many MOFI DSF rips.

Why, when I use the exact same ripping process these select 31 MOFI albums rip to DFF?
When all other SACD's including other MOFI's rip to DSF?

 

[ted_b]

Hmm, that's strange. The argument in sacd_extract is different (-p vs -s) but you'd have to manually choose that (or via a gui like ISO2DSD). Are you using more than one tool?

 

[marpow]

Hmm, that's strange. The argument in sacd_extract is different (-p vs -s) but you'd have to manually choose that (or via a gui like ISO2DSD). Are you using more than one tool?

Yeah, I just don't know. Maybe I changed the settings accidentally on my Sonore software ripper? Or the import settings on my JRiver player was changed?
I only have used the OPPO105 and the Sonore extract, nothing ever used, oh well, I have a little stack of SACD's that I will rip in a week or so and see what happens.
I am tempted to re rip one of those discs in DFF just to see what I get for experimental purposes.
But not today, they all play good.