5 Oct
2012
5 Oct
'12
6:30 p.m.
Hi
I play keyboards in a couple of bands. When setting up my sounds I often
find that on some songs my keyboard is much louder than on others. I'm
pretty sure it's because to the extremely mixed level of the mp3 files I
use for rehearsing and programming the synth sounds.
So my question is: how can I match the levels of the mp3 files so that
they will playback in levels "as close as possible to the level a my
band plays them"? I imagine a ballad would be played back at lower level
than a power rock tune, not sure if there's a way to compensate for
that, though. So at least something that sounds like the same level.
NB: The tunes vary a great deal in audio quality and
loudness-war-compression (some are new dance/pop tunes, some are old
soul tunes, some are something else).
Thanks in advance for any input!
--
Atte
http://atte.dk http://modlys.dk
5 Oct
5 Oct
7:12 p.m.
Hallo!
Do you just want to adjust the loudness between multiple files?
(this means there is just a constant gain for every file so the they are
similar on average)
Or do you also want to change the loudness inside one track?
For case 1, you have to measure the loudness somehow (best way would be
to use a loudness standard like used in EBU R128 - see
https://auphonic.com/blog/2012/08/02/loudness-measurement-and-normalization…),
and then you apply a constant gain, combined with a limiter (and maybe
compressor) if the gain is positive.
For case 2, it's more complex ;)
BTW:
we build an audio processing webservice (http://auphonic.com), which
does both cases, for case 1 just use the "global loudness normalization"
algorithm, for case 2 use the "adaptive leveler" too!
For case 1 you could also use libebur128 (or some other tools) to
measure the loudness, then apply a gain and limiter somehow ...
LG
Georg
Am 2012-10-05 18:30, schrieb Atte André Jensen:
Hi
I play keyboards in a couple of bands. When setting up my sounds I often
find that on some songs my keyboard is much louder than on others. I'm
pretty sure it's because to the extremely mixed level of the mp3 files I
use for rehearsing and programming the synth sounds.
So my question is: how can I match the levels of the mp3 files so that
they will playback in levels "as close as possible to the level a my
band plays them"? I imagine a ballad would be played back at lower level
than a power rock tune, not sure if there's a way to compensate for
that, though. So at least something that sounds like the same level.
NB: The tunes vary a great deal in audio quality and
loudness-war-compression (some are new dance/pop tunes, some are old
soul tunes, some are something else).
Thanks in advance for any input!
--
auphonic - automatic audio post production
http://auphonic.com
audio development, machine learning, open source and more
9:13 p.m.
On Fri, Oct 05, 2012 at 07:12:00PM +0200, Georg Holzmann wrote:
Do you just want to adjust the loudness between
multiple files?
(this means there is just a constant gain for every file so the they
are similar on average)
Or do you also want to change the loudness inside one track?
For case 1, you have to measure the loudness somehow (best way would
be to use a loudness standard like used in EBU R128 - see
https://auphonic.com/blog/2012/08/02/loudness-measurement-and-normalization…),
and then you apply a constant gain, combined with a limiter (and
maybe compressor) if the gain is positive.
Or use ebumeter (real-time, GUI) or ebur128 (batch processing).
The latter combined with some scripting and sox (or doing it
manually) will provide what you need.
Both available from:
<http://kokkinizita.linuxaudio.org/linuxaudio/downloads>.
LAC paper on the theory behind this:
<http://kokkinizita.linuxaudio.org/papers>
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
6 Oct
6 Oct
1:08 p.m.
On 10/05/2012 09:13 PM, Fons Adriaensen wrote:
Thanks for your reply, Fons!
> Do you just want to adjust the loudness between
multiple files?
> (this means there is just a constant gain for every file so the they
> are similar on average)
> Or do you also want to change the loudness inside one track?
Not sure what you mean. I don't to change the internal volume of each
track. Basically I just want to avoid that when I program the levels of
the patches on my hardware synth, I later (on stage or rehearsing with
the band) realize that on some songs my synth is too loud and on some
songs it's to soft.
I tried it on a couple of files:
atte@blokhus:~/music/bands/eldorado/mp3$ ls *.wav | xargs -iarg ebur128 arg
Integrated loudness: 13.7 LU
Loudness range: 3.8 LU
Integrated loudness: 13.0 LU
Loudness range: 2.2 LU
Integrated loudness: 6.0 LU
Loudness range: 3.6 LU
Integrated loudness: 10.8 LU
Loudness range: 4.5 LU
Integrated loudness: 12.9 LU
Loudness range: 3.2 LU
Ok, so it seems the third file is softer than the rest, which is very
true. But how do I interpret those values, and how do I call sox for a
gain adjustment?
Sorry for being stupid...
--
Atte
http://atte.dk http://modlys.dk
1:51 p.m.
Hallo!
I meant, if you have longer files with a high loudness range, maybe you
should also adjust the loudness of different regions within your file -
but I guess it's not the case ... ;)
> Do you
just want to adjust the loudness between multiple files?
> (this means there is just a constant gain for every file so the they
> are similar on average)
> Or do you also want to change the loudness inside one track?
Not sure what you mean. I don't to change the internal volume of each
track. Basically I just want to avoid that when I program the levels of
the patches on my hardware synth, I later (on stage or rehearsing with
the band) realize that on some songs my synth is too loud and on some
songs it's to soft. I tried it on a couple of files:
atte@blokhus:~/music/bands/eldorado/mp3$ ls *.wav | xargs -iarg ebur128 arg
Integrated loudness: 13.7 LU
Loudness range: 3.8 LU
Integrated loudness: 13.0 LU
Loudness range: 2.2 LU
Integrated loudness: 6.0 LU
Loudness range: 3.6 LU
Integrated loudness: 10.8 LU
Loudness range: 4.5 LU
Integrated loudness: 12.9 LU
Loudness range: 3.2 LU
Ok, so it seems the third file is softer than the rest, which is very
true. But how do I interpret those values, and how do I call sox for a
gain adjustment?
That's great - now you only have to bring all your files to the same
loudness!
For example, let's say you want all files to be at 13 LU.
You can apply a simple gain on all files:
-0.7 dB for the first file (13.7LU), 0dB for the second, +7dB for the
third (6LU) and so on ...
I think you can do that with sox - and maybe it's also good to use an
additional limiter (at -1dBFS or lower), to avoid clipping!
AFAIK sox has the compad effect, which can do that ...
Have a nice weekend,
LG
Georg
--
auphonic - automatic audio post production
http://auphonic.com
audio development, machine learning, open source and more
7 Oct
7 Oct
8:56 p.m.
On 10/06/2012 01:51 PM, Georg Holzmann wrote:
That's great - now you only have to bring all your
files to the same
loudness!
For example, let's say you want all files to be at 13 LU.
You can apply a simple gain on all files:
-0.7 dB for the first file (13.7LU), 0dB for the second, +7dB for the
third (6LU) and so on ...
I think you can do that with sox - and maybe it's also good to use an
additional limiter (at -1dBFS or lower), to avoid clipping!
AFAIK sox has the compad effect, which can do that ...
Thanks for the help! That's great.
Now I started looking for an sox alternative that could do gain
adjustment without having to decode/reencode the mp3, since I seemed to
remember such a tool existed.
This led me to mp3gain. Reading it's man page:
"mp3gain can analyze and adjust mp3 files so that they have the same
volume."
Hmmmm. Now I'm thinking, isn't this exactly what I wanted in the first
place?
After running "mp3gain -r *.mp3" on the same files as I tested with in
the original mail, and checking them with ebur128, it seems that the two
programs do not agree how to measure the perceived loudness. mp3gain
uses the replay gain algorithm, what ever that means. Any ideas whether
mp3gains measurements should fare worse than ebur128?
The reason I ask is, that I'm tempted to just use "mp3gain -r *.mp3" :-)
--
Atte
http://atte.dk http://modlys.dk
10:40 p.m.
On Sun, October 7, 2012 11:56 am, Atte André Jensen wrote:
After running "mp3gain -r *.mp3" on the same
files as I tested with in
the original mail, and checking them with ebur128, it seems that the two
programs do not agree how to measure the perceived loudness. mp3gain
uses the replay gain algorithm, what ever that means. Any ideas whether
mp3gains measurements should fare worse than ebur128?
The reason I ask is, that I'm tempted to just use "mp3gain -r *.mp3" :-)
Use your ears. Close your eyes and listen. If you like what you hear use
what makes it that way.
--
Len Ovens
www.OvenWerks.net
11:34 p.m.
On Sun, Oct 07, 2012 at 01:40:43PM -0700, Len Ovens wrote:
On Sun, October 7, 2012 11:56 am, Atte André Jensen wrote:
The two algorithms can be expected to produce different results,
in the sense that when you apply the replay gain adjustment then
the LU values are likely not to be zero, probably all positive.
But at least with the examples you posted (which have little
dynamic range), the LU values after replay gain should all be
close. Is that not the case ?
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
After running "mp3gain -r *.mp3" on the
same files as I tested with in
the original mail, and checking them with ebur128, it seems that the two
programs do not agree how to measure the perceived loudness. mp3gain
uses the replay gain algorithm, what ever that means. Any ideas whether
mp3gains measurements should fare worse than ebur128?
The reason I ask is, that I'm tempted to just use "mp3gain -r *.mp3" :-)
Use your ears. Close your eyes and listen. If you like what you hear use
what makes it that way.
8 Oct
8 Oct
8:18 a.m.
On 10/07/2012 10:40 PM, Len Ovens wrote:
Use your ears. Close your eyes and listen. If you like
what you hear use
what makes it that way.
I totally agree.
However for now, I think I'll start with mp3gain and that's not good
enough, I'll dig into ebur128.
Thanks for the pointers!
--
Atte
http://atte.dk http://modlys.dk
2:48 p.m.
On 10/08/2012 05:18 PM, Atte André Jensen wrote:
On 10/07/2012 10:40 PM, Len Ovens wrote:
replaygain only works for players that have it enabled. It is written as
metadata or tag to the audio file without actually changing the audio part.
Cheers
Roger
Use your ears. Close your eyes and listen. If you
like what you hear use
what makes it that way.
I totally agree.
However for now, I think I'll start with mp3gain and that's not good
enough, I'll dig into ebur128.
Thanks for the pointers!
8:01 p.m.
On 10/08/2012 02:47 PM, Roger wrote:
replaygain only works for players that have it
enabled. It is written as
metadata or tag to the audio file without actually changing the audio part.
From the man page (of mp3gain):
"mp3gain optionally writes gain adjustments directly into the encoded
data. In this case, the adjustment works with all mp3 players, i.e. no
support for a special tag is required. This mode is activated by any
of the options -r, -a, -g, or -l."
--
Atte
http://atte.dk http://modlys.dk
10:21 p.m.
On Mon, Oct 08, 2012 at 11:47:55PM +1100, Roger wrote:
replaygain only works for players that have it
enabled. It is
written as metadata or tag to the audio file without actually
changing the audio part.
Which would explain things if ebur128 still shows large
differences (which I wouldn't expect) after applying replaygain...
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
12:30 a.m.
On Sun, Oct 07, 2012 at 08:56:50PM +0200, Atte André Jensen wrote:
After running "mp3gain -r *.mp3" on the same
files as I tested with
in the original mail, and checking them with ebur128, it seems that
the two programs do not agree how to measure the perceived loudness.
mp3gain uses the replay gain algorithm, what ever that means. Any
ideas whether mp3gains measurements should fare worse than ebur128?
The reason I ask is, that I'm tempted to just use "mp3gain -r *.mp3" :-)
I'd expect R-128 to do a better job on content that has a somewhat
wider dynamic range than typical pop music: classical, film sound,
etc. That said, if 'mp3gain -r *.mp3' produces the results you need,
then that's what you should use.
No loudness adjustment algorithm can do 'the right thing' fully
automatically in all cases and replace the sound engineer's
judgement, for the simple reason that not everything should have
the same loudness. What R-128 and similar systems can do is to
provide reliable inputs even in conditions when you can't trust
your ears.
Yesterday I recorded a live performance of 'Vortex Temporum', a
marvellous piece for piano, string trio, clarinet and flute (and
the 'bass' versions of the latter two), by the French composer
Gerard Grisey. I don't have the file here ATM, but I'll run it
through ebur128 when I'll do the editing one of these days, as
I'm curious to know what that will produce.
I was watching jkmeter during the recording. The loudest parts
were around -20 dB RMS (having attended the rehearsal as well I
allowed for some margin, peak level was around -5 dB or so). The
softest parts, lasting for a minute or so, were some 45 dB lower
which is extreme even for this kind of music.
No loudness measurement algo will be able to make sense of that.
As the recording is intended for radio broadcast, I'll have to
adjust levels by something like 15 dB, which won't be easy.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
6 Oct
6 Oct
2:55 p.m.
On Sat, Oct 06, 2012 at 01:08:43PM +0200, Atte André Jensen wrote:
I tried it on a couple of files:
atte@blokhus:~/music/bands/eldorado/mp3$ ls *.wav | xargs -iarg ebur128 arg
Integrated loudness: 13.7 LU
Loudness range: 3.8 LU
Integrated loudness: 13.0 LU
Loudness range: 2.2 LU
Integrated loudness: 6.0 LU
Loudness range: 3.6 LU
Integrated loudness: 10.8 LU
Loudness range: 4.5 LU
Integrated loudness: 12.9 LU
Loudness range: 3.2 LU
Ok, so it seems the third file is softer than the rest, which is
very true. But how do I interpret those values, and how do I call
sox for a gain adjustment?
A 'Loudness Unit' (LU) is the same as a dB. So if you want a loudness L
and a file measures M, then you need to adjust by L - M dB. Beware of
clipping if you apply positive gain. Future releases of ebumeter and
ebur128 will measure peak level as well to make this easier.
The basic sox command is IIRC:
sox infile outfile gain value-in-db
See the sox manpage for details.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
4472
days inactive
4475
days old
linux-audio-user@lists.linuxaudio.org
13 comments
5 participants
participants (5)
-
Atte André Jensen -
Fons Adriaensen -
Georg Holzmann -
Len Ovens -
Roger