Re: [Savonet-users] source mixing
> Regarding the LADSPA plugins, I tried using them, but "liquidsoap
--list-plugins" doesn't show any.
Nevermind, I think it's a local problem. I managed to run the plugins on
two other machines.
Am 25.11.18 um 23:02 schrieb Bernhard Schmidt:
Hi,
thanks for your responses. I think I'm going with this line for now (
plus another normalize()), though I have to test if it works in all
cases and what's the limit of number of sources...
mixed = add(normalize=true,weights=[],[])
Regarding the LADSPA plugins, I tried using them, but "liquidsoap
--list-plugins" doesn't show any. Neither does "liquidsoap -h
plugin."... -> "Plugin not found!"
I installed the ladspa-swh-plugins (CentOS7), which then seem to be
located in /usr/lib64/ladspa/.
I tried "export LIQ_LADSPA_PATH=/usr/lib64/ladspa/", but that didn't
make a difference.
So I tried installing the ladspa tap-plugins
(https://github.com/kmatheussen/tap-plugins) both in
/usr/local/lib/ladspa/ and /usr/lib/ladspa/, but nothing worked. (make
and make install)
Any idea why?
Regards
Am 17.11.2018 um 17:06 schrieb Romain Beauxis:
Hey there!
Le sam. 17 nov. 2018 à 09:35, Bernhard Schmidt <mailto:b...@pluto24.com>> a écrit :
> I'm trying to use liquidsoap for merging multiple input files into a
> single stream, which then goes out to icecast. This raises a couple of
> questions: (hope you don't mind I list them here all at once)
Of course not!
> 1) So far mixing/merging seems to work with add(), although I'm not
sure
> if this is the right way. (e.g. what is the difference to mix()?) Are
> there any limitations?
add() is the basic operator for sources. mix is more evolved, with
telnet interface and etc. Unless you plan on using that you should
stick with add.
> 2) Apparently the more sources I add, the lower the volume gets.
> normalize() didn't seem to make a difference, so I added an amplify()
> with the volume equals the number of input-sources. Is this the right
> way to go?
>
>
> pl0 = single("file1.wav")
> pl1 = single("file2.wav")
> pl1 = single("file3.wav")
>
> mixed = amplify(3.,add([pl0,pl1,pl2]))
>
By default, if you have 3 sources, the volume of each source is
divided by 3, and etc. This is to avoid clipping when adding sources.
However, if 2 of the added source, for instance, are silent, this
will just end up dividing the volume of the last one by 3.
I don't think that there's an easy default here that can work for
everybody. You can disable the normalization or use custom weights. I
believe the amplify above is equivalent to just doing:
mixed = add(normalize=false,[pl0,pl1,pl2])
> 3) When I want the sources to have different volumes, is it
> better/faster to use the weights in add(), or run each through an
> amplify() function?
I think it makes more sense to use custom weights just because that
processing is already going to happen there. Adding separate
amplify() shouldn't be very CPU intensive, though.
> 4) Is it possible to add an reverb effect?
We support LADSPA plugins. You should be able to find a ladspa reverb
plugin and use it with liquidsoap.
> 5) What is the best format for the sources to work with? Lossless I
> guess, WAV? FLAC?
> I'm intending to start multiple liquidsoap instances, so performance is
> important..
Internally, liquidsoap decodes everything and manipulates raw PCM
audio. Decoding isn't very CPU intensive but if you really want to
minimize that part you can use WAV input files.
Hope this helps, all the best with your project!
Romain
___
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users
___
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users
___
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users
Re: [Savonet-users] source mixing
Hi,
thanks for your responses. I think I'm going with this line for now (
plus another normalize()), though I have to test if it works in all
cases and what's the limit of number of sources...
mixed = add(normalize=true,weights=[],[])
Regarding the LADSPA plugins, I tried using them, but "liquidsoap
--list-plugins" doesn't show any. Neither does "liquidsoap -h
plugin."... -> "Plugin not found!"
I installed the ladspa-swh-plugins (CentOS7), which then seem to be
located in /usr/lib64/ladspa/.
I tried "export LIQ_LADSPA_PATH=/usr/lib64/ladspa/", but that didn't
make a difference.
So I tried installing the ladspa tap-plugins
(https://github.com/kmatheussen/tap-plugins) both in
/usr/local/lib/ladspa/ and /usr/lib/ladspa/, but nothing worked. (make
and make install)
Any idea why?
Regards
Am 17.11.2018 um 17:06 schrieb Romain Beauxis:
Hey there!
Le sam. 17 nov. 2018 à 09:35, Bernhard Schmidt <mailto:b...@pluto24.com>> a écrit :
> I'm trying to use liquidsoap for merging multiple input files into a
> single stream, which then goes out to icecast. This raises a couple of
> questions: (hope you don't mind I list them here all at once)
Of course not!
> 1) So far mixing/merging seems to work with add(), although I'm not sure
> if this is the right way. (e.g. what is the difference to mix()?) Are
> there any limitations?
add() is the basic operator for sources. mix is more evolved, with
telnet interface and etc. Unless you plan on using that you should
stick with add.
> 2) Apparently the more sources I add, the lower the volume gets.
> normalize() didn't seem to make a difference, so I added an amplify()
> with the volume equals the number of input-sources. Is this the right
> way to go?
>
>
> pl0 = single("file1.wav")
> pl1 = single("file2.wav")
> pl1 = single("file3.wav")
>
> mixed = amplify(3.,add([pl0,pl1,pl2]))
>
By default, if you have 3 sources, the volume of each source is
divided by 3, and etc. This is to avoid clipping when adding sources.
However, if 2 of the added source, for instance, are silent, this will
just end up dividing the volume of the last one by 3.
I don't think that there's an easy default here that can work for
everybody. You can disable the normalization or use custom weights. I
believe the amplify above is equivalent to just doing:
mixed = add(normalize=false,[pl0,pl1,pl2])
> 3) When I want the sources to have different volumes, is it
> better/faster to use the weights in add(), or run each through an
> amplify() function?
I think it makes more sense to use custom weights just because that
processing is already going to happen there. Adding separate amplify()
shouldn't be very CPU intensive, though.
> 4) Is it possible to add an reverb effect?
We support LADSPA plugins. You should be able to find a ladspa reverb
plugin and use it with liquidsoap.
> 5) What is the best format for the sources to work with? Lossless I
> guess, WAV? FLAC?
> I'm intending to start multiple liquidsoap instances, so performance is
> important..
Internally, liquidsoap decodes everything and manipulates raw PCM
audio. Decoding isn't very CPU intensive but if you really want to
minimize that part you can use WAV input files.
Hope this helps, all the best with your project!
Romain
___
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users
___
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users
[Savonet-users] source mixing
Hi,
I'm trying to use liquidsoap for merging multiple input files into a
single stream, which then goes out to icecast. This raises a couple of
questions: (hope you don't mind I list them here all at once)
1) So far mixing/merging seems to work with add(), although I'm not sure
if this is the right way. (e.g. what is the difference to mix()?) Are
there any limitations?
2) Apparently the more sources I add, the lower the volume gets.
normalize() didn't seem to make a difference, so I added an amplify()
with the volume equals the number of input-sources. Is this the right
way to go?
pl0 = single("file1.wav")
pl1 = single("file2.wav")
pl1 = single("file3.wav")
mixed = amplify(3.,add([pl0,pl1,pl2]))
3) When I want the sources to have different volumes, is it
better/faster to use the weights in add(), or run each through an
amplify() function?
4) Is it possible to add an reverb effect?
5) What is the best format for the sources to work with? Lossless I
guess, WAV? FLAC?
I'm intending to start multiple liquidsoap instances, so performance is
important..
Thanks in advance :)
___
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users
