After noticiting that playing a track for the second time sounds better then the first time, I did some research.
The second time a track plays, it is read from the linux OS disk cache instead from disk. This means the track is played from RAM memory instead of from the hard drive that contains the file (track).
So I was looking for a way to copy a track to RAM memory before it is being played. I came up with a very simple but effective solution that works fine.
By modifying my custom-convert.conf from:
to:
Search for the differences ;D
It's this:
cp $FILE$ /dev/null;
before the sox upsampling command.
What this does is: copy the file (track) to the null device. This is very fast, because the null device is virtual and does not require disk writes. As a result the file (track) has been read once from disk and is now in the linux disk cache.
On the OS level I could test and prove this really works! So all music tracks now play from RAM memory, resulting in a more stable stereo image. And yes that is very subjective ;-)
Maybe someone else likes to try this as well :-)
The second time a track plays, it is read from the linux OS disk cache instead from disk. This means the track is played from RAM memory instead of from the hard drive that contains the file (track).
So I was looking for a way to copy a track to RAM memory before it is being played. I came up with a very simple but effective solution that works fine.
By modifying my custom-convert.conf from:
Code:
flc flc * *
# FT:{START=--skip=%t}U:{END=--until=%v}
[flac] -dcs $START$ $END$ -- $FILE$ | [sox14.4.2] -v0.98 --buffer 32768 -q -t wav - -t flac -e signed -C0 -b24 - rate -v -M -a -b 90.7 96000 dither -SCode:
flc flc * *
# FT:{START=--skip=%t}U:{END=--until=%v}
cp $FILE$ /dev/null; [flac] -dcs $START$ $END$ -- $FILE$ | [sox14.4.2] -v0.98 --buffer 32768 -q -t wav - -t flac -e signed -C0 -b24 - rate -v -M -a -b 90.7 96000 dither -SIt's this:
cp $FILE$ /dev/null;
before the sox upsampling command.
What this does is: copy the file (track) to the null device. This is very fast, because the null device is virtual and does not require disk writes. As a result the file (track) has been read once from disk and is now in the linux disk cache.
On the OS level I could test and prove this really works! So all music tracks now play from RAM memory, resulting in a more stable stereo image. And yes that is very subjective ;-)
Maybe someone else likes to try this as well :-)