More > JRiver Media Center 21 for Linux
I can play DSD ISOs, but only in redbook
Hendrik:
--- Quote from: blgentry on October 14, 2015, 04:00:14 pm ---I disagree. Any time you "create" interpolated samples you're introducing something that isn't in the original. How can that *not* be worse?
--- End quote ---
No matter what you resample to, you have to introduce new data. The point only is that integer multiple or not makes no difference, as the algorithms don't care about that difference.
Obviously not resampling at all is best, but thats not the point I was making.
bamyasi:
OK, progress!!
I set 'Audio Device' to "front:CARD=x20 [ALSA]".
If output encoding is set to "none" then pcm plays natively, but DSD plays at 352.8khz. If I set output encoding to "DSD in DoP format" EVERYTHING is resampled to DSD 128.
edit: I might have had it set to 2xDSD, but I still haven't found a setting for universal native playback.
Awesome Donkey:
Is bitstreaming enabled (assuming it'd work on Linux)?
blgentry:
--- Quote from: Hendrik on October 14, 2015, 04:22:33 pm ---No matter what you resample to, you have to introduce new data.
--- End quote ---
Not when you're down sampling. 352.8 -> 44.1 just discards samples. No samples have to be interpolated at all. That was my point. ...and honestly I didn't post this to talk about quality. Just that integer resampling is easier when going down in sample rate in particular. ...and of course less calculations (interpolation) when going up.
Brian.
Hendrik:
--- Quote from: blgentry on October 14, 2015, 05:09:01 pm ---Just that integer resampling is easier when going down in sample rate in particular. ...and of course less calculations (interpolation) when going up.
--- End quote ---
But thats the entire point, its not "easier" with any sophisticated resampler. Its not like you just drop every second sample when going 88.2->44.1, that would result in terrible audio (strong aliasing). Instead, you fully interpolate the final sample rate from the original.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version