hey Olaf, list,
is there any trick to get a mono stream working with the shoutcast~ external?
i seem to be getting some wierd effects with all mono streams.
stereo works like a charm. downsampling is also not a problem.
mono, however, generates some distorted output.
thanks - august.
Hi August,
in line 313 of shoutcast~.c (version 0.3c) starts code that is only under Linux. I understand the code used for Win but never realised why it should be different under Linux... You could try to use the code used for Win. BTW, which version of Lame are you using? I use 3.90 under Win / OS X and had problems with other versions. But I haven't tried mono on OS X yet.
Olaf
august schrieb:
hey Olaf, list,
is there any trick to get a mono stream working with the shoutcast~ external?
i seem to be getting some wierd effects with all mono streams.
stereo works like a charm. downsampling is also not a problem.
mono, however, generates some distorted output.
thanks - august.
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
hi,
i can't really see what motivates this answer since :
- it's not been said this happened on Linux
- on the contrary, i did a fast test, it works on Linux
- read the code, Linux solution is more logic to my eyes.
next,
sevy/yves
Olaf Matthes wrote:
Hi August,
in line 313 of shoutcast~.c (version 0.3c) starts code that is only under Linux. I understand the code used for Win but never realised why it should be different under Linux... You could try to use the code used for Win. BTW, which version of Lame are you using? I use 3.90 under Win / OS X and had problems with other versions. But I haven't tried mono on OS X yet.
Olaf
august schrieb:
hey Olaf, list,
is there any trick to get a mono stream working with the shoutcast~ external?
i seem to be getting some wierd effects with all mono streams.
stereo works like a charm. downsampling is also not a problem.
mono, however, generates some distorted output.
thanks - august.
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
Hi August and Yves,
first of all, there's a small bug: shoutcast~ takes the number of channels as first creation argument. This should allways be 2, even if encoding to mono! I'll correct this for the next release where the first creation argument will determine the number of inlets only, not channels.
Solution for the moment: use "shoutcast~ 2" and connect the audio signal to both inlets.
Yves Degoyon schrieb:
- read the code, Linux solution is more logic to my eyes.
in lame.h it says:
- NOTE:
- if gfp->num_channels=2, but gfp->mode = 3 (mono), the L & R channels
- will be averaged into the L channel before encoding only the L channel
- This will overwrite the data in buffer_l[] and buffer_r[].
and
- NOTE:
- num_samples = number of samples in the L (or R)
- channel, not the total number of samples in pcm[]
Since we allways submit two channels and have to tell the encoder the number of samples per channel I find the solution used for Win more logic.
Maybe the lame code has changed / differs between Linux and Win?
the confused Olaf
hi Olaf,
why are you arguing soooo long while i still don't know it this happened on Linux or Win + with which version ???
besides, it definitely works on Linux. period.
next,
sevy/yves
Olaf Matthes wrote:
Hi August and Yves,
first of all, there's a small bug: shoutcast~ takes the number of channels as first creation argument. This should allways be 2, even if encoding to mono! I'll correct this for the next release where the first creation argument will determine the number of inlets only, not channels.
Solution for the moment: use "shoutcast~ 2" and connect the audio signal to both inlets.
Yves Degoyon schrieb:
- read the code, Linux solution is more logic to my eyes.
in lame.h it says:
- NOTE:
- if gfp->num_channels=2, but gfp->mode = 3 (mono), the L & R channels
- will be averaged into the L channel before encoding only the L channel
- This will overwrite the data in buffer_l[] and buffer_r[].
and
- NOTE:
- num_samples = number of samples in the L (or R)
- channel, not the total number of samples in pcm[]
Since we allways submit two channels and have to tell the encoder the number of samples per channel I find the solution used for Win more logic.
Maybe the lame code has changed / differs between Linux and Win?
the confused Olaf
On Fri, 6 Sep 2002, Olaf Matthes wrote:
Hi August,
BTW, which version of Lame are you using? I use 3.90 under Win / OS X and had problems with other versions. But I haven't tried mono on OS X yet.
When speaking of lame-versions, I had to fix the lame.h file to get the shout* objects to work with lame 3.92. One of the structures in the file has got two new int values since 3.89.
good point... the error is to distribute lame's include files with objects ... we should use system-wide installed includes of course... ( i had never noticed this really )
regards,
sevy/yves
Kjetil S. Matheussen wrote:
On Fri, 6 Sep 2002, Olaf Matthes wrote:
Hi August,
BTW, which version of Lame are you using? I use 3.90 under Win / OS X and had problems with other versions. But I haven't tried mono on OS X yet.
When speaking of lame-versions, I had to fix the lame.h file to get the shout* objects to work with lame 3.92. One of the structures in the file has got two new int values since 3.89.
hi again,
i've looked into that further and the difference of includes explains a lot !!
in fact, mp3 objects were not working with latest releases of lame because of that !!
The includes ( mpglib.h, mpg123.h, interface.h ) have been bundled with the objects because they are part of mpglib ( the decoding library included in lame ) and ARE NOT installed system-wide when you install lame.
besides, they had to be modified to compile in the context of an extern.
so, i've updated these includes ( mainly mpglib.h ) in the linux version of mp3 objects ( mp3write~, shoucast~, shoutamp~, mp3live~ ) and, thus, they work now with latest releases of lame ( > 3.90 ).
i think those versions are more stable so, as far as i could run some tests, it's a kind of improvement.
if you update those objects, update lame also ( i used 3.92 ).
cheers,
sevy/yves
[root@dregs mp3live~]# grep 3.89 * mpg123.h:#define M_PI 3.14159265358979323846
Yves Degoyon wrote:
good point... the error is to distribute lame's include files with objects ... we should use system-wide installed includes of course... ( i had never noticed this really )
regards,
sevy/yves
Kjetil S. Matheussen wrote:
On Fri, 6 Sep 2002, Olaf Matthes wrote:
Hi August,
BTW, which version of Lame are you using? I use 3.90 under Win / OS X and had problems with other versions. But I haven't tried mono on OS X yet.
When speaking of lame-versions, I had to fix the lame.h file to get the shout* objects to work with lame 3.92. One of the structures in the file has got two new int values since 3.89.
Yves, Olaf,
I was using linux with lame 3.92 and starting shoutcast~ with stereo input like this : shoutcast~ 2 512
with the new lame headers, it seems to work like a charm though.
thanks , best - august.
hi,
these informations must have been given off-list. it's although funny because the shoutcast unauthorized version doesn't take any creation argument. yes, Olaf's and mine ARE definitely different.
the real error was to use lame 3.92 when the INSTALL file recommended 3.89.
regards,
sevy/yves
august wrote:
Yves, Olaf,
I was using linux with lame 3.92 and starting shoutcast~ with stereo input like this : shoutcast~ 2 512
with the new lame headers, it seems to work like a charm though.
thanks , best - august.
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
Yves Degoyon schrieb:
these informations must have been given off-list.
...see August's post to the pd-list from August, 28th....
it's although funny because the shoutcast unauthorized version doesn't take any creation argument. yes, Olaf's and mine ARE definitely different.
Yes, they are and I'm currently thinking about changing the names of my externals to avoid any confusion. - I just don't like the idea of talking about namclashes and on the other hand having 'branches' of objects. Yes, I know, it was me having introduced new versions with new / different features...
Olaf
Olaf Matthes wrote:
Yves Degoyon schrieb:
these informations must have been given off-list.
...see August's post to the pd-list from August, 28th....
it's although funny because the shoutcast unauthorized version doesn't take any creation argument. yes, Olaf's and mine ARE definitely different.
Yes, they are and I'm currently thinking about changing the names of my externals to avoid any confusion. - I just don't like the idea of talking about namclashes and on the other hand having 'branches' of objects. Yes, I know, it was me having introduced new versions with new / different features...
i can do it myself ... and rename them to "lincast~" & "linamp~", what do u think ??
Olaf
Yves Degoyon schrieb:
i can do it myself ... and rename them to "lincast~" & "linamp~", what do u think ??
my idea was to use the name icecast~... but this would suggest they only work with Icecast server.... lincast~ would be okay to me, but what's about mp3cast~ / mp3amp~ - this would fit 'more nicely' with your mp3stream~ externals.... ;-)
thanks anyway,
Olaf
Olaf Matthes wrote:
Yves Degoyon schrieb:
i can do it myself ... and rename them to "lincast~" & "linamp~", what do u think ??
my idea was to use the name icecast~... but this would suggest they only work with Icecast server....
<grin> what about the name shoutcast ? </grin>
lincast~ would be okay to me, but what's about mp3cast~ / mp3amp~ - this would fit 'more nicely' with your mp3stream~ externals.... ;-)
yep, it seems more clever and i just did it.
the hardest part in mutual developpement is NOT to have a CVS server up and running but rather to agree on all features and to communicate properly. i won't participate in such an effort here again ..... of course, at work, i HAVE TO.
cheers,
sevy/yves
thanks anyway,
Olaf
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
-
august -
Kjetil S. Matheussen -
Olaf Matthes -
Yves Degoyon