More > JRiver Media Center 25 for Windows
Multichannel over DoP
Hendrik:
Bitstreaming has no particular deep meaning for DSD, especially if DoP is being used. Don't attribute too much into that word. All it tells MC to do is to avoid decoding the selected formats.
All "Bitstreaming" is basically just packing of non-PCM data into a PCM-style stream. This applies equally to DoP (where DoP is this sort of packing), or movie audio formats like TrueHD. You could equally wrap such a movie audio format into a PCM stream, put it in a file, and play it as if it were PCM, and it would probably work.
"Native" DSD is actually the only exception, since it uses ASIO interfaces to directly communicate the non-PCM DSD data to the driver, without additional wrapping.
For the issue at hand, this seems like a shortcoming in the DAC, doesn't it? If it detects a 5 channel input, shouldn't it be smart about which channels are in there? I would be surprised if every single player is expected to mess with the DSD signal instead of the DAC just being smart for you.
With PCM, this gets solved by inserting empty channel. One of the key problems in inserting empty DSD channels is the way DSD works, there is no absolute value for a silent stream. In PCM, you just set everything to zero, and its perfectly silent. Such a value does not exist for DSD, what I have seen to generate "DSD silence" is generating a very high frequency noise signal and hoping that the DAC will filter that out completely.
dtc:
This goes back to an old discussion of how to handle 3,4,5,6 channel signals. They can either have a .1 channel or not. 4 channel can be 4.0 or 3.1. 6 channel can be 5.1 or 6.0. The DAC has no way to sort out those issues. I think the software is the place to handle it, but it needs additional tags to indicate if a LFE channel is present. The problem is the standards dictate a standard order, which the funky file formats don't always follow. A channel map in the header would help, but that is very unlikely to happen. There is no easy answer. Plus, trying to determine if a channel is totally silent requires looking at the channel for the entire length of the track, something you probably don't want to the DAC to do on the fly.
Software is the most flexible place to handle this, but it is by no means easy and these funky formats are just not very common.
kr4:
--- Quote from: Hendrik on June 30, 2019, 05:42:48 pm ---For the issue at hand, this seems like a shortcoming in the DAC, doesn't it? If it detects a 5 channel input, shouldn't it be smart about which channels are in there? I would be surprised if every single player is expected to mess with the DSD signal instead of the DAC just being smart for you.
--- End quote ---
For the issue at hand, the interface is not a DAC but a USB-to-AES3 (via I2S) bridge. It has 10 input and 10 output channels but is relatively dumb. It does, however, allow the use of multiple synch-ed stereo DACs for multichannel.
kr4:
--- Quote from: dtc on June 30, 2019, 06:13:57 pm ---Software is the most flexible place to handle this, but it is by no means easy and these funky formats are just not very common.
--- End quote ---
Just not very common? I have thousands. 5.0 is the standard for Classical multichannel recordings as very few are 5.1. Currently, MC fixes this if one simply specifies "5.1 channels" in DSP Studio and uses PCM. Not so for DSD.
dtc:
--- Quote from: kr4 on June 30, 2019, 06:39:49 pm ---Just not very common? I have thousands. 5.0 is the standard for Classical multichannel recordings as very few are 5.1. Currently, MC fixes this if one simply specifies "5.1 channels" in DSP Studio and uses PCM. Not so for DSD.
--- End quote ---
Sorry, did not mean to mininize it. The point was simply that there are far more 5.1 disks/files sold than 5.0 disks/files.
I do agree that if MC could basically do the same thing with DSD that it does with PCM that would be a solution to your problem. And, given that bitstreaming is not necessary, there could be a solution, like Roon has found a solution.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version