DSD Playback

[simoncn]

(08-03-2015 10:38)beckphotonik Wrote:  I agree with Anton, there seems to be more freedom to the music with DSD. Personally I suspect it is the hard filtering needed to maintain the audio bandwidth but reject the Nyquist frequency and beyond in pcm formats that make listening to pcm less pleasurable than DSD. DSD filtering is more noise shaping so in band phase and group delay distortions are less obtrusive.

Another possibility is that the on-the-fly DSD to PCM conversion algorithms used by ffmpeg aren't "state of the art" compared with other converters such as the Weiss Saracon offline conversion software.

Also, I have read that a perfect lossless conversion of DSD64 to PCM would produce a 64-bit 352.8 KHz PCM file. Downsampling this to 24/176 will invevitably degrade sound quality.

[AntonD]

(08-03-2015 19:07)simoncn Wrote:  

(08-03-2015 10:38)beckphotonik Wrote:  I agree with Anton, there seems to be more freedom to the music with DSD. Personally I suspect it is the hard filtering needed to maintain the audio bandwidth but reject the Nyquist frequency and beyond in pcm formats that make listening to pcm less pleasurable than DSD. DSD filtering is more noise shaping so in band phase and group delay distortions are less obtrusive.

Another possibility is that the on-the-fly DSD to PCM conversion algorithms used by ffmpeg aren't "state of the art" compared with other converters such as the Weiss Saracon offline conversion software.

Also, I have read that a perfect lossless conversion of DSD64 to PCM would produce a 64-bit 352.8 KHz PCM file. Downsampling this to 24/176 will invevitably degrade sound quality.

Hi Simon
Thanks for this information
If you know, when setting dopwav as the output, what makes the output setting to wav 24/176? (Seen on my control point)
Assuming my streamer could accept 32/384, would it be possible to set output to wav32;352 to get as close as possible to the original DSD source?
Im interested to get as close as possible for best sound quality.
Many thanks in advance
Anton

[beckphotonik]

(09-03-2015 22:58)AntonD Wrote:  

(08-03-2015 19:07)simoncn Wrote:  

(08-03-2015 10:38)beckphotonik Wrote:  I agree with Anton, there seems to be more freedom to the music with DSD. Personally I suspect it is the hard filtering needed to maintain the audio bandwidth but reject the Nyquist frequency and beyond in pcm formats that make listening to pcm less pleasurable than DSD. DSD filtering is more noise shaping so in band phase and group delay distortions are less obtrusive.

Another possibility is that the on-the-fly DSD to PCM conversion algorithms used by ffmpeg aren't "state of the art" compared with other converters such as the Weiss Saracon offline conversion software.

Also, I have read that a perfect lossless conversion of DSD64 to PCM would produce a 64-bit 352.8 KHz PCM file. Downsampling this to 24/176 will invevitably degrade sound quality.

Hi Simon
Thanks for this information
If you know, when setting dopwav as the output, what makes the output setting to wav 24/176? (Seen on my control point)
Assuming my streamer could accept 32/384, would it be possible to set output to wav32;352 to get as close as possible to the original DSD source?
Im interested to get as close as possible for best sound quality.
Many thanks in advance
Anton

Anton, my understanding is that dopwav is not like transcoding from one format to another but a packaging format to enable a bitstream to be sent as pcm over ethernet. The streamer reinstates the bitstream so you should have full resolution of the DSD source.
Donna.

[simoncn]

(09-03-2015 23:11)beckphotonik Wrote:  

(09-03-2015 22:58)AntonD Wrote:  Hi Simon
Thanks for this information
If you know, when setting dopwav as the output, what makes the output setting to wav 24/176? (Seen on my control point)
Assuming my streamer could accept 32/384, would it be possible to set output to wav32;352 to get as close as possible to the original DSD source?
Im interested to get as close as possible for best sound quality.
Many thanks in advance
Anton

Anton, my understanding is that dopwav is not like transcoding from one format to another but a packaging format to enable a bitstream to be sent as pcm over ethernet. The streamer reinstates the bitstream so you should have full resolution of the DSD source.
Donna.

The information from Donna is correct. The control point is seeing the DoP/WAV stream as 24/176 because 24/176 is the packaging format required by DoP for a DSD64 file. This can't be changed to any other value and it doesn't mean that the DSD data played by the streamer is 24/176.

[AntonD]

(09-03-2015 23:21)simoncn Wrote:  

(09-03-2015 23:11)beckphotonik Wrote:  

(09-03-2015 22:58)AntonD Wrote:  Hi Simon
Thanks for this information
If you know, when setting dopwav as the output, what makes the output setting to wav 24/176? (Seen on my control point)
Assuming my streamer could accept 32/384, would it be possible to set output to wav32;352 to get as close as possible to the original DSD source?
Im interested to get as close as possible for best sound quality.
Many thanks in advance
Anton

Anton, my understanding is that dopwav is not like transcoding from one format to another but a packaging format to enable a bitstream to be sent as pcm over ethernet. The streamer reinstates the bitstream so you should have full resolution of the DSD source.
Donna.

The information from Donna is correct. The control point is seeing the DoP/WAV stream as 24/176 because 24/176 is the packaging format required by DoP for a DSD64 file. This can't be changed to any other value and it doesn't mean that the DSD data played by the streamer is 24/176.

Thank you both. That makes sense now.
That explains why I could hear a difference when using DoP as the output.
Anton

[ppienter]

This is a great feature ! Thanks !!

I tried following setting : 'dsf:wav24;88' on my QNAP NAS and get the message "Unsupported transcoding combination 'dsf:wav24;88' ".
Using alac:wav is working fine.

I have no idea which ffmpeg version is installed nor how to upgrade it.

Any help is more than welcome !

Thanks,

Piet.

[simoncn]

(19-03-2015 19:14)ppienter Wrote:  This is a great feature ! Thanks !!

I tried following setting : 'dsf:wav24;88' on my QNAP NAS and get the message "Unsupported transcoding combination 'dsf:wav24;88' ".
Using alac:wav is working fine.

I have no idea which ffmpeg version is installed nor how to upgrade it.

Any help is more than welcome !

Thanks,

Piet.

This message means your version of ffmpeg doesn't support DSF to PCM conversion. Which QNAP NAS are you using?

[ppienter]

(19-03-2015 19:45)simoncn Wrote:  This message means your version of ffmpeg doesn't support DSF to PCM conversion. Which QNAP NAS are you using?

QNAP TS419P-II with QTS 4.1.0

[simoncn]

(19-03-2015 20:55)ppienter Wrote:  QNAP TS419P-II with QTS 4.1.0

To get a newer version of FFmpeg for this NAS, you would need to download an ARMv5 static build of FFmpeg from the link in this post and copy the ffmpeg executable to your NAS, then put the full Linux path to the ffmpeg executable into the MinimServer stream.converter property.

Even after doing this, you might have problems because the processor in your NAS doesn't have floating-point hardware and this might cause dropouts when trying to do DSD to PCM conversion in real time. I have a similar NAS with a slower processor (1.2 GHz instead of 2.0 GHz) and DSD to PCM transcoding is unusable because of dropouts every few seconds.

[ppienter]

(19-03-2015 22:44)simoncn Wrote:  

(19-03-2015 20:55)ppienter Wrote:  QNAP TS419P-II with QTS 4.1.0

To get a newer version of FFmpeg for this NAS, you would need to download an ARMv5 static build of FFmpeg from the link in this post and copy the ffmpeg executable to your NAS, then put the full Linux path to the ffmpeg executable into the MinimServer stream.converter property.

Even after doing this, you might have problems because the processor in your NAS doesn't have floating-point hardware and this might cause dropouts when trying to do DSD to PCM conversion in real time. I have a similar NAS with a slower processor (1.2 GHz instead of 2.0 GHz) and DSD to PCM transcoding is unusable because of dropouts every few seconds.

Thanks. I'll try it later on.