01-14-2024, 01:51 AM
(This post was last modified: 01-14-2024, 02:33 AM by TheOldPresbyope.
Edit Reason: fix the math and wording
)
@Doc77
First, this thread has stretched to 40+ posts and long since is not about the subject line. It's long since time to close it and open another if necessary with a relevant subject line.
Second...ah, well, I see Tim beat me to the punch while I was composing. In essence, you must be careful talking about those sample sizes (1-bit vs 24-bit vs 32-bit) and apparent frequencies (176.4kHz vs 352.8kHz). DSD and PCM audio are different beasts. Google is your friend if you want to do some homework on what DSD encoding (DSD64, DSD128, DSD256, and DSD512) is and how DSD-encoded material can be contained in a .dsf or .dff file. Study as well how DSD-encoding compares to PCM-encoding. Finally, study how a DSD-encoded stream can be encapsulated in PCM frames in order to slip it through PCM-only interfaces to DAC chips which know how to extract the DSD-encoded stream.
Third, I just tried sending a DSD64 track to my Tone1 with MPD set to use Native DSD format then tried sending that same track with MPD set to embed DSD as DoP. I don't want to add screen captures here so I'll give you the screen scrapes instead.
First, as native DSD
Note that "1-bit 2.822 MHz" is per channel. That MHz really ought to be MSs for mega-samples per second but that's the quirks of the audio world for you. Stereo is two channels so the total transmission rate is 5.645 Mbps. (mega-bits per sec).
Now again, with MPD set to embed DSD in PCM frames
Ah, now some interesting things happen. The 24 bit you see in "Decoded to" has been padded to "32 bit" in "Output format" because the DAC has reported it accepts only 32-bit formats. The total "Decoded to" bit rate and the total "Output format" bit rate have stretched by the same ratio simply because there are now more bits per second, albeit 1/3 1/4 of them now are zero.
Either way, the track plays fine on my system and I hear it fine. Mind you, my DAC does not support Hardware Volume control of DSD material so I have to set volume = 100 percent and use a downstream preamp to control the volume.
In your most recent reply, the first screen capture suggests that, for whatever settings you have in this case, MPD is not sending unaltered DSD material to the DAC. Rather, it is converting it on-the-fly to PCM just as described in the information provided on the MPD settings screen for DSD format, which Tim included in his reply. THIS IS NOT THE SAME AS DoP.
By the way, I have DSD-encoded material on hand only for testing purposes. Even if I thought there were an inherent advantage to this distribution mechanism, which I don't, my ears are long past their prime and I can't hear any advantage.
ETA - Oh, jeez. I see you slipped in a final reply while I was wordsmithing. Looks like we're done here.
Regards,
Kent
First, this thread has stretched to 40+ posts and long since is not about the subject line. It's long since time to close it and open another if necessary with a relevant subject line.
Second...ah, well, I see Tim beat me to the punch while I was composing. In essence, you must be careful talking about those sample sizes (1-bit vs 24-bit vs 32-bit) and apparent frequencies (176.4kHz vs 352.8kHz). DSD and PCM audio are different beasts. Google is your friend if you want to do some homework on what DSD encoding (DSD64, DSD128, DSD256, and DSD512) is and how DSD-encoded material can be contained in a .dsf or .dff file. Study as well how DSD-encoding compares to PCM-encoding. Finally, study how a DSD-encoded stream can be encapsulated in PCM frames in order to slip it through PCM-only interfaces to DAC chips which know how to extract the DSD-encoded stream.
Third, I just tried sending a DSD64 track to my Tone1 with MPD set to use Native DSD format then tried sending that same track with MPD set to embed DSD as DoP. I don't want to add screen captures here so I'll give you the screen scrapes instead.
First, as native DSD
Code:
INPUT / OUTPUT
Source NAS/OMV-Music/DSD/NDSD012/10_Welcome-to-NativeDSD-Sampler_NativeDSD-Presents_NDSD012_2ch64.dsf
Source format DSD 1 bit 2.822 MHz, Stereo
Decoded to DSD 1 bit 2.822 MHz, Stereo, 5.645 Mbps
Output format DSD 1 bit 2.822 MHz, Stereo, 5.645 Mbps
ALSA mode Direct (hw)
Audio chain MPD -> hw -> DeviceDSP OPERATIONS
Resample rate Off
Quality Off
Selective Off
CamillaDSP Off
Volume type Hardware (On-chip)AUDIO DEVICE
Device Tone1
Mixer name Tone1
Interface USB
Formats S32_LE, SPECIAL, DSD_U32_BE
Platform Pi-4B 1.1 2GNote that "1-bit 2.822 MHz" is per channel. That MHz really ought to be MSs for mega-samples per second but that's the quirks of the audio world for you. Stereo is two channels so the total transmission rate is 5.645 Mbps. (mega-bits per sec).
Now again, with MPD set to embed DSD in PCM frames
Code:
INPUT / OUTPUT
Source NAS/OMV-Music/DSD/NDSD012/10_Welcome-to-NativeDSD-Sampler_NativeDSD-Presents_NDSD012_2ch64.dsf
Source format DSD 1 bit 2.822 MHz, Stereo
Decoded to DoP 24 bit 176.4 kHz, Stereo, 8.467 Mbps
Output format PCM 32 bit 176.4 kHz, Stereo, 11.290 Mbps
ALSA mode Direct (hw)
Audio chain MPD -> hw -> DeviceDSP OPERATIONS
Resample rate Off
Quality Off
Selective Off
CamillaDSP Off
Volume type Hardware (On-chip)AUDIO DEVICE
Device Tone1
Mixer name Tone1
Interface USB
Formats S32_LE, SPECIAL, DSD_U32_BE
Platform Pi-4B 1.1 2GBAh, now some interesting things happen. The 24 bit you see in "Decoded to" has been padded to "32 bit" in "Output format" because the DAC has reported it accepts only 32-bit formats. The total "Decoded to" bit rate and the total "Output format" bit rate have stretched by the same ratio simply because there are now more bits per second, albeit 1/3 1/4 of them now are zero.
Either way, the track plays fine on my system and I hear it fine. Mind you, my DAC does not support Hardware Volume control of DSD material so I have to set volume = 100 percent and use a downstream preamp to control the volume.
In your most recent reply, the first screen capture suggests that, for whatever settings you have in this case, MPD is not sending unaltered DSD material to the DAC. Rather, it is converting it on-the-fly to PCM just as described in the information provided on the MPD settings screen for DSD format, which Tim included in his reply. THIS IS NOT THE SAME AS DoP.
By the way, I have DSD-encoded material on hand only for testing purposes. Even if I thought there were an inherent advantage to this distribution mechanism, which I don't, my ears are long past their prime and I can't hear any advantage.
ETA - Oh, jeez. I see you slipped in a final reply while I was wordsmithing. Looks like we're done here.
Regards,
Kent
