Mercurial > pidgin
view src/mediastreamer/msAlawenc.c @ 12360:3153661f4d5c
[gaim-migrate @ 14664]
Faceprint is concerned about 2 things:
1)some of the random colors are very close together.
as best we can tell, there are two ways to fix this
1a) for each proposed color, iterate the entire list of selected
colors, looking to ensure that it is not too close to any of them.
this is an O(n^2) operation, with n >= 220 (the current number of
colors we look for)
1b) iterate the entire set of possible colors, skipping ahead by some
guess (rather than iterating by 1). this is an O(n^3) operation,
where n is 65535/(whatever we skip ahead by). This is not only a
more expensive operation, but because of the nature of the color
list, it is not _necessarily_ going to yield more predictable
results, skipping ahead 5 (or any other number) does not necessarily
guarantee that you've skipped 5 very similar colors.
2) as you can see, either solution to #1 is potentially a resource hog.
#1a is a random delay, #1b is inherently expensive. How often #1a will
exceed the bound #1b, if ever, is unknown.
rather than doing either of these, we settled on a middle course: a .h
file has been created containing a set of colors. currently the set we
were previously hard coded to. Gaim will search that list for usable
colors and start randomly looking only if that list does not contain
sufficient usable colors. ideally this list would be generated to have
colors that are known to be a "safe" distance appart, that is colors that
you can tell appart. and Ideally it would have a (small) multiple of the
number of colors we are searching for. This should ensure that IF we go
to randomly searching, we need do so only for a few colors.
Right now I have no good way to generate a "safe" list of colors though.
committer: Tailor Script <tailor@pidgin.im>
author | Luke Schierer <lschiere@pidgin.im> |
---|---|
date | Mon, 05 Dec 2005 21:46:47 +0000 |
parents | e67993da8a22 |
children |
line wrap: on
line source
/* The mediastreamer library aims at providing modular media processing and I/O for linphone, but also for any telephony application. Copyright (C) 2001 Simon MORLAT simon.morlat@linphone.org This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ #include "msAlawenc.h" #include "g711common.h" extern MSCodecInfo ALAWinfo; static MSALAWEncoderClass *ms_ALAWencoder_class=NULL; MSFilter * ms_ALAWencoder_new(void) { MSALAWEncoder *r; r=g_new(MSALAWEncoder,1); ms_ALAWencoder_init(r); if (ms_ALAWencoder_class==NULL) { ms_ALAWencoder_class=g_new(MSALAWEncoderClass,1); ms_ALAWencoder_class_init(ms_ALAWencoder_class); } MS_FILTER(r)->klass=MS_FILTER_CLASS(ms_ALAWencoder_class); return(MS_FILTER(r)); } /* FOR INTERNAL USE*/ void ms_ALAWencoder_init(MSALAWEncoder *r) { ms_filter_init(MS_FILTER(r)); MS_FILTER(r)->infifos=r->f_inputs; MS_FILTER(r)->outfifos=r->f_outputs; MS_FILTER(r)->r_mingran=ALAW_ENCODER_RMAXGRAN; /* the filter can be called as soon as there is something to process */ memset(r->f_inputs,0,sizeof(MSFifo*)*MSALAWENCODER_MAX_INPUTS); memset(r->f_outputs,0,sizeof(MSFifo*)*MSALAWENCODER_MAX_INPUTS); } void ms_ALAWencoder_class_init(MSALAWEncoderClass *klass) { ms_filter_class_init(MS_FILTER_CLASS(klass)); ms_filter_class_set_name(MS_FILTER_CLASS(klass),"ALAWEncoder"); MS_FILTER_CLASS(klass)->info=(MSFilterInfo*)&ALAWinfo; MS_FILTER_CLASS(klass)->max_finputs=MSALAWENCODER_MAX_INPUTS; MS_FILTER_CLASS(klass)->max_foutputs=MSALAWENCODER_MAX_INPUTS; MS_FILTER_CLASS(klass)->r_maxgran=ALAW_ENCODER_RMAXGRAN; MS_FILTER_CLASS(klass)->w_maxgran=ALAW_ENCODER_WMAXGRAN; MS_FILTER_CLASS(klass)->destroy=(MSFilterDestroyFunc)ms_ALAWencoder_destroy; MS_FILTER_CLASS(klass)->process=(MSFilterProcessFunc)ms_ALAWencoder_process; } void ms_ALAWencoder_process(MSALAWEncoder *r) { MSFifo *fi,*fo; int inlen,outlen; gchar *s,*d; int i; /* process output fifos, but there is only one for this class of filter*/ /* this is the sophisticated design of the process function: Here the filter declares that it can be called as soon as there is something to read on the input fifo by setting r_mingran=0. Then it ask for the fifo to get as many data as possible by calling: inlen=ms_fifo_get_read_ptr(fi,0,(void**)&s); This avoid multiple call to the process function to process all data available on the input fifo... but the writing of the process function is a bit more difficult, because althoug ms_fifo_get_read_ptr() returns N bytes, we cannot ask ms_fifo_get_write_ptr to return N bytes if N>MS_FILTER_CLASS(klass)->w_maxgran. This is forbidden by the MSFifo mechanism. This is an open issue. For the moment what is done here is that ms_fifo_get_write_ptr() is called several time with its maximum granularity in order to try to write the output. ... One solution: -create a new function ms_fifo_get_rw_ptr(fifo1,p1, fifo2,p2) to return the number of bytes able to being processed according to the input and output fifo, and their respective data pointers */ fi=r->f_inputs[0]; fo=r->f_outputs[0]; inlen=ms_fifo_get_read_ptr(fi,ALAW_ENCODER_RMAXGRAN,(void**)&s); if (s==NULL) return; outlen=ms_fifo_get_write_ptr(fo,ALAW_ENCODER_WMAXGRAN,(void**)&d); if (d!=NULL) { for(i=0;i<ALAW_ENCODER_WMAXGRAN;i++) { d[i]=s16_to_alaw( *((gint16*)s) ); s+=2; } } else g_warning("MSALAWDecoder: Discarding samples !!"); } void ms_ALAWencoder_destroy( MSALAWEncoder *obj) { g_free(obj); }