16 Mar
2008
16 Mar
'08
11:08 a.m.
Hello.
Part of my group project involves a music playing system. What we have
is a bunch of stages connected together with buffers and pipes. The
first stage decodes an mp3 file with mplayer and sends the raw audio
down the pipe in 44.1KHz 16bit. No headers are ever sent (as far as I
know). What I'm trying to do is add an EQ stage. The problem is that the
samples which come in seem to be equalised to the maximum level possible
with 16bit. Therefore if I increase the volume of any band it will cause
clipping. The only solution I can think of at the moment is to just EQ
down, but that sucks a bit.
Can anyone shed some light on what it is I'm doing wrong?
Thanks.
Simon
16 Mar
16 Mar
12:20 p.m.
On Sun, Mar 16, 2008 at 7:08 AM, Simon Williams
<simon(a)systemparadox.co.uk> wrote:
Hello.
Part of my group project involves a music playing system. What we have
is a bunch of stages connected together with buffers and pipes. The
first stage decodes an mp3 file with mplayer and sends the raw audio
down the pipe in 44.1KHz 16bit. No headers are ever sent (as far as I
know). What I'm trying to do is add an EQ stage. The problem is that the
samples which come in seem to be equalised to the maximum level possible
with 16bit. Therefore if I increase the volume of any band it will cause
clipping. The only solution I can think of at the moment is to just EQ
down, but that sucks a bit.
Can anyone shed some light on what it is I'm doing wrong?
Thanks.
Simon
Simon,
Actually, EQing down, as you call it, is really the right thing to
do and the way that all good recording teachers/books will teach you
to work with the audio. that said it isn't, for most people, the most
natural thing to do for some reason and most of us had to learn to get
through it.
Keeping the audio at maximum values as far as you can through the
audio chain is best for signal-to-noise, etc., and the audio at the
start of your pipe is doing that. EQing is now a subjective step you
are adding, presumably for good reason. To do this you cut the other
EQ bands so as to allow the band you want to be more promenent to come
through. You might just do this with a volume control at the front end
of your EQ unit, or you might adjust each of the unwanted bands. The
two options may have different sonic characteristics. Play with it and
hear for yourself.
Keep in mind that if you wanted 10% more of something and wanted to
boost the band to get it you would have had to have had 10% less of
everything in the first place to give you headroom to do it. Since
there's no way to know whether you will want 10%, 20% or even 50% more
of something down the road it's far better to keep everything as loud
as it can be through the whole process, without clipping at the top,
to maximize signal-to-noise and give you the most flexibility in all
stages.
Hope this helps,
Mark
17 Mar
17 Mar
8:47 a.m.
Mark Knecht wrote:
Keeping the audio at maximum values as far as you
can through the
audio chain is best for signal-to-noise, etc., and the audio at the
start of your pipe is doing that. EQing is now a subjective step you
are adding, presumably for good reason. To do this you cut the other
EQ bands so as to allow the band you want to be more promenent to come
through. You might just do this with a volume control at the front end
of your EQ unit, or you might adjust each of the unwanted bands. The
two options may have different sonic characteristics. Play with it and
hear for yourself.
Right. The question now is how do I control the overall volume? Is there
anything in the header which says how loud the track is? Not that that
will help here- I think my only option is to tell the mplayer at the end
of the pipe to turn it up a bit.
10:38 a.m.
On Mon, Mar 17, 2008 at 4:46 AM, Simon Williams
<simon(a)systemparadox.co.uk> wrote:
Mark Knecht wrote:
If the data is 'headerless' (I'm not a programmer so I cannot help you
with that part) then there is no header to tell you anything, is
there?
I think the general path you are looking for is that if you logically
want to raise a single band of EQ by 10db then what you really want to
do physically is cut the level of the other 9 bands by 10db leaving
the one you want at 0db cut. If you want to adjust the level of your
player at the end of the chain to get a final volume that's probably a
good feature to build in but in general I *think* that the final level
in mplayer should be viewed as a way of leveling different audio
sources as different audio files have a small/large degree of
variation in overall volume.
Hope this helps,
Mark
Keeping the audio at maximum values as far as
you can through the
audio chain is best for signal-to-noise, etc., and the audio at the
start of your pipe is doing that. EQing is now a subjective step you
are adding, presumably for good reason. To do this you cut the other
EQ bands so as to allow the band you want to be more promenent to come
through. You might just do this with a volume control at the front end
of your EQ unit, or you might adjust each of the unwanted bands. The
two options may have different sonic characteristics. Play with it and
hear for yourself.
Right. The question now is how do I control the overall volume? Is there
anything in the header which says how loud the track is? Not that that
will help here- I think my only option is to tell the mplayer at the end
of the pipe to turn it up a bit.
6042
days inactive
6043
days old
linux-audio-user@lists.linuxaudio.org
3 comments
2 participants
participants (2)
-
Mark Knecht -
Simon Williams