Usage of gst_pad_alloc_buffer in source elements ?
Locked
Usage of gst_pad_alloc_buffer in source elements ?
 
|
Hi gstreamer-devel,
I'm interested in writing a bunch of plugins/elements to get gstreamer perform well enough to display SD/HD videos on a set top box chip. >From my firsts investigations, it seems quite possible and development should start quite soon. However, after reading a few article about gstreamer on TI OMAP chips (this one in particular http://www.eetimes.com/news/design/showArticle.jhtml?articleID=193401461&pgno=2 ), I'm wondering why most of the source elements does not use the gst_pad_alloc_buffer function. In my case, this would be quite interesting to avoid a memory copy and to directly write into a contigeous dma mmapped buffer. Is it something to be done ? Is there a very few interest from the community ? Thanks in advance for your responses. -- Lionel Landwerlin ------------------------------------------------------------------------------ This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
Re: Usage of gst_pad_alloc_buffer in source elements ?
|
|
If your source element directly produce video frames then it makes sense to implement such allocation functions.
As most source elements are producing compressed data (filesrc, demuxers,etc...), it does not really make sense to write into a downstream allocated buffer. Hope this helps, Julien Moutte, FLUENDO S.A. On Tue, Dec 22, 2009 at 10:47 AM, Lionel Landwerlin <[hidden email]> wrote: Hi gstreamer-devel, ------------------------------------------------------------------------------ This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
Re: Usage of gst_pad_alloc_buffer in source elements ?
|
|
Well...
I'm expecting mpegts data from various sources (file/http/udp/etc...) to be transfered through dma to the hardware demuxer. Does it make sense ? -- Lionel Landwerlin On Tue, Dec 22, 2009 at 11:04 AM, Julien Moutte <[hidden email]> wrote: > If your source element directly produce video frames then it makes sense to > implement such allocation functions. > > As most source elements are producing compressed data (filesrc, > demuxers,etc...), it does not really make sense to write into a downstream > allocated buffer. > > Hope this helps, > > Julien Moutte, > FLUENDO S.A. > > > On Tue, Dec 22, 2009 at 10:47 AM, Lionel Landwerlin <[hidden email]> > wrote: >> >> Hi gstreamer-devel, >> >> I'm interested in writing a bunch of plugins/elements to get gstreamer >> perform well enough to display SD/HD videos on a set top box chip. >> >From my firsts investigations, it seems quite possible and development >> should start quite soon. However, after reading a few article about >> gstreamer on TI OMAP chips (this one in particular >> >> http://www.eetimes.com/news/design/showArticle.jhtml?articleID=193401461&pgno=2 >> ), >> I'm wondering why most of the source elements does not use the >> gst_pad_alloc_buffer function. >> >> In my case, this would be quite interesting to avoid a memory copy and >> to directly write into a contigeous dma mmapped buffer. >> >> Is it something to be done ? >> Is there a very few interest from the community ? >> >> Thanks in advance for your responses. >> >> -- >> Lionel Landwerlin >> >> >> ------------------------------------------------------------------------------ >> This SF.Net email is sponsored by the Verizon Developer Community >> Take advantage of Verizon's best-in-class app development support >> A streamlined, 14 day to market process makes app distribution fast and >> easy >> Join now and get one step closer to millions of Verizon customers >> http://p.sf.net/sfu/verizon-dev2dev >> _______________________________________________ >> gstreamer-devel mailing list >> [hidden email] >> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel > > > ------------------------------------------------------------------------------ > This SF.Net email is sponsored by the Verizon Developer Community > Take advantage of Verizon's best-in-class app development support > A streamlined, 14 day to market process makes app distribution fast and easy > Join now and get one step closer to millions of Verizon customers > http://p.sf.net/sfu/verizon-dev2dev > _______________________________________________ > gstreamer-devel mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/gstreamer-devel > > ------------------------------------------------------------------------------ This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
Re: Usage of gst_pad_alloc_buffer in source elements ?
|
|
Potentially yes, you could save one memcpy by doing that. No idea of exactly how much you would gain by doing that for compressed data buffers, might be worth a try :)
Julien Moutte, FLUENDO S.A. On Tue, Dec 22, 2009 at 11:15 AM, Lionel Landwerlin <[hidden email]> wrote: Well... ------------------------------------------------------------------------------ This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
Re: Usage of gst_pad_alloc_buffer in source elements ?
|
|
Ok, thanks you for your quick answer !
So now the 100$ question ;) If there is a reasonnable gain, would it be considered for inclusion in the official gstreamer elements ? That's just to save me some time in posting patches whether nobody is interested. -- Lionel Landwerlin On Tue, Dec 22, 2009 at 11:19 AM, Julien Moutte <[hidden email]> wrote: > Potentially yes, you could save one memcpy by doing that. No idea of exactly > how much you would gain by doing that for compressed data buffers, might be > worth a try :) > > Julien Moutte, > FLUENDO S.A. > > > On Tue, Dec 22, 2009 at 11:15 AM, Lionel Landwerlin <[hidden email]> > wrote: >> >> Well... >> >> I'm expecting mpegts data from various sources (file/http/udp/etc...) >> to be transfered through dma to the hardware demuxer. >> Does it make sense ? >> >> -- >> Lionel Landwerlin >> >> On Tue, Dec 22, 2009 at 11:04 AM, Julien Moutte <[hidden email]> wrote: >> > If your source element directly produce video frames then it makes sense >> > to >> > implement such allocation functions. >> > >> > As most source elements are producing compressed data (filesrc, >> > demuxers,etc...), it does not really make sense to write into a >> > downstream >> > allocated buffer. >> > >> > Hope this helps, >> > >> > Julien Moutte, >> > FLUENDO S.A. >> > >> > >> > On Tue, Dec 22, 2009 at 10:47 AM, Lionel Landwerlin >> > <[hidden email]> >> > wrote: >> >> >> >> Hi gstreamer-devel, >> >> >> >> I'm interested in writing a bunch of plugins/elements to get gstreamer >> >> perform well enough to display SD/HD videos on a set top box chip. >> >> >From my firsts investigations, it seems quite possible and development >> >> should start quite soon. However, after reading a few article about >> >> gstreamer on TI OMAP chips (this one in particular >> >> >> >> >> >> http://www.eetimes.com/news/design/showArticle.jhtml?articleID=193401461&pgno=2 >> >> ), >> >> I'm wondering why most of the source elements does not use the >> >> gst_pad_alloc_buffer function. >> >> >> >> In my case, this would be quite interesting to avoid a memory copy and >> >> to directly write into a contigeous dma mmapped buffer. >> >> >> >> Is it something to be done ? >> >> Is there a very few interest from the community ? >> >> >> >> Thanks in advance for your responses. >> >> >> >> -- >> >> Lionel Landwerlin >> >> >> >> >> >> >> >> ------------------------------------------------------------------------------ >> >> This SF.Net email is sponsored by the Verizon Developer Community >> >> Take advantage of Verizon's best-in-class app development support >> >> A streamlined, 14 day to market process makes app distribution fast and >> >> easy >> >> Join now and get one step closer to millions of Verizon customers >> >> http://p.sf.net/sfu/verizon-dev2dev >> >> _______________________________________________ >> >> gstreamer-devel mailing list >> >> [hidden email] >> >> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel >> > >> > >> > >> > ------------------------------------------------------------------------------ >> > This SF.Net email is sponsored by the Verizon Developer Community >> > Take advantage of Verizon's best-in-class app development support >> > A streamlined, 14 day to market process makes app distribution fast and >> > easy >> > Join now and get one step closer to millions of Verizon customers >> > http://p.sf.net/sfu/verizon-dev2dev >> > _______________________________________________ >> > gstreamer-devel mailing list >> > [hidden email] >> > https://lists.sourceforge.net/lists/listinfo/gstreamer-devel >> > >> > >> >> >> ------------------------------------------------------------------------------ >> This SF.Net email is sponsored by the Verizon Developer Community >> Take advantage of Verizon's best-in-class app development support >> A streamlined, 14 day to market process makes app distribution fast and >> easy >> Join now and get one step closer to millions of Verizon customers >> http://p.sf.net/sfu/verizon-dev2dev >> _______________________________________________ >> gstreamer-devel mailing list >> [hidden email] >> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel > > > ------------------------------------------------------------------------------ > This SF.Net email is sponsored by the Verizon Developer Community > Take advantage of Verizon's best-in-class app development support > A streamlined, 14 day to market process makes app distribution fast and easy > Join now and get one step closer to millions of Verizon customers > http://p.sf.net/sfu/verizon-dev2dev > _______________________________________________ > gstreamer-devel mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/gstreamer-devel > > ------------------------------------------------------------------------------ This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
Re: Usage of gst_pad_alloc_buffer in source elements ?
|
|
Am 22.12.2009 12:38, schrieb Lionel Landwerlin:
> Ok, thanks you for your quick answer ! > > So now the 100$ question ;) > If there is a reasonnable gain, would it be considered for inclusion > in the official gstreamer elements ? > That's just to save me some time in posting patches whether nobody is > interested. If you get it working, it saves CPU and it does not make the coder very ugly, it will be considered :) Stefan > > -- > Lionel Landwerlin > > On Tue, Dec 22, 2009 at 11:19 AM, Julien Moutte <[hidden email]> wrote: >> Potentially yes, you could save one memcpy by doing that. No idea of exactly >> how much you would gain by doing that for compressed data buffers, might be >> worth a try :) >> >> Julien Moutte, >> FLUENDO S.A. >> >> >> On Tue, Dec 22, 2009 at 11:15 AM, Lionel Landwerlin <[hidden email]> >> wrote: >>> >>> Well... >>> >>> I'm expecting mpegts data from various sources (file/http/udp/etc...) >>> to be transfered through dma to the hardware demuxer. >>> Does it make sense ? >>> >>> -- >>> Lionel Landwerlin >>> >>> On Tue, Dec 22, 2009 at 11:04 AM, Julien Moutte <[hidden email]> wrote: >>>> If your source element directly produce video frames then it makes sense >>>> to >>>> implement such allocation functions. >>>> >>>> As most source elements are producing compressed data (filesrc, >>>> demuxers,etc...), it does not really make sense to write into a >>>> downstream >>>> allocated buffer. >>>> >>>> Hope this helps, >>>> >>>> Julien Moutte, >>>> FLUENDO S.A. >>>> >>>> >>>> On Tue, Dec 22, 2009 at 10:47 AM, Lionel Landwerlin >>>> <[hidden email]> >>>> wrote: >>>>> >>>>> Hi gstreamer-devel, >>>>> >>>>> I'm interested in writing a bunch of plugins/elements to get gstreamer >>>>> perform well enough to display SD/HD videos on a set top box chip. >>>>> >From my firsts investigations, it seems quite possible and development >>>>> should start quite soon. However, after reading a few article about >>>>> gstreamer on TI OMAP chips (this one in particular >>>>> >>>>> >>>>> http://www.eetimes.com/news/design/showArticle.jhtml?articleID=193401461&pgno=2 >>>>> ), >>>>> I'm wondering why most of the source elements does not use the >>>>> gst_pad_alloc_buffer function. >>>>> >>>>> In my case, this would be quite interesting to avoid a memory copy and >>>>> to directly write into a contigeous dma mmapped buffer. >>>>> >>>>> Is it something to be done ? >>>>> Is there a very few interest from the community ? >>>>> >>>>> Thanks in advance for your responses. >>>>> >>>>> -- >>>>> Lionel Landwerlin >>>>> >>>>> >>>>> >>>>> ------------------------------------------------------------------------------ >>>>> This SF.Net email is sponsored by the Verizon Developer Community >>>>> Take advantage of Verizon's best-in-class app development support >>>>> A streamlined, 14 day to market process makes app distribution fast and >>>>> easy >>>>> Join now and get one step closer to millions of Verizon customers >>>>> http://p.sf.net/sfu/verizon-dev2dev >>>>> _______________________________________________ >>>>> gstreamer-devel mailing list >>>>> [hidden email] >>>>> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel >>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> This SF.Net email is sponsored by the Verizon Developer Community >>>> Take advantage of Verizon's best-in-class app development support >>>> A streamlined, 14 day to market process makes app distribution fast and >>>> easy >>>> Join now and get one step closer to millions of Verizon customers >>>> http://p.sf.net/sfu/verizon-dev2dev >>>> _______________________________________________ >>>> gstreamer-devel mailing list >>>> [hidden email] >>>> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel >>>> >>>> >>> >>> >>> ------------------------------------------------------------------------------ >>> This SF.Net email is sponsored by the Verizon Developer Community >>> Take advantage of Verizon's best-in-class app development support >>> A streamlined, 14 day to market process makes app distribution fast and >>> easy >>> Join now and get one step closer to millions of Verizon customers >>> http://p.sf.net/sfu/verizon-dev2dev >>> _______________________________________________ >>> gstreamer-devel mailing list >>> [hidden email] >>> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel >> >> >> ------------------------------------------------------------------------------ >> This SF.Net email is sponsored by the Verizon Developer Community >> Take advantage of Verizon's best-in-class app development support >> A streamlined, 14 day to market process makes app distribution fast and easy >> Join now and get one step closer to millions of Verizon customers >> http://p.sf.net/sfu/verizon-dev2dev >> _______________________________________________ >> gstreamer-devel mailing list >> [hidden email] >> https://lists.sourceforge.net/lists/listinfo/gstreamer-devel >> >> > > ------------------------------------------------------------------------------ > This SF.Net email is sponsored by the Verizon Developer Community > Take advantage of Verizon's best-in-class app development support > A streamlined, 14 day to market process makes app distribution fast and easy > Join now and get one step closer to millions of Verizon customers > http://p.sf.net/sfu/verizon-dev2dev > _______________________________________________ > gstreamer-devel mailing list > [hidden email] > https://lists.sourceforge.net/lists/listinfo/gstreamer-devel ------------------------------------------------------------------------------ This SF.Net email is sponsored by the Verizon Developer Community Take advantage of Verizon's best-in-class app development support A streamlined, 14 day to market process makes app distribution fast and easy Join now and get one step closer to millions of Verizon customers http://p.sf.net/sfu/verizon-dev2dev _______________________________________________ gstreamer-devel mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/gstreamer-devel |
«
Return to GStreamer-devel
|
1 view|%1 views
| Free forum by Nabble | Edit this page |