[FFmpeg-devel] Intel IPP H264 encoder

Robert Swain robert.swain
Fri Dec 4 13:50:28 CET 2009


On Fri, 2009-12-04 at 10:57 +0000, M?ns Rullg?rd wrote:
> Robert Swain <robert.swain at gmail.com> writes:
> 
> > On Thu, 2009-12-03 at 18:48 -0800, Jason Garrett-Glaser wrote:
> >> > Now, I suppose I could run everything through some carefully crafted
> >> > scripts, making a couple of metrics and distributions on residues, packet
> >> > sizes, processor loads, cache usage and everything you can throw in there.
> >> 
> >> Or you could, you know, post a screenshot or two and give your
> >> encoding commandlines.  But this is apparently too hard for you.
> >> 
> >> > No worries, I wont bother anyone with the results, or the code, since
> >> > obviously it has been developed on crack, with no idea whats going on,
> >> > resting in a legal swamp.
> >> >
> >> > Have a nice day all
> >> 
> >> Thanks.  Don't let the door hit you on the way out.
> >
> > Yey! Another miscommunication or breakdown of communication, due to lack
> > of humility and/or willingness to explain what someone from outside the
> > 'community' may not know, leads to another contributor being driven
> > away.
> 
> I think we'll do just fine without his "contributions".

We probably will, but his (and other people's) contributions could
improve the project directly or indirectly.

> > All that was needed was "Could you please provide information about the
> > sample used, the settings used to encode the sample with each encoder
> > and any other useful information? e.g. x264 defaults with no b-frames
> > and IPP with this option and that option."
> 
> That's exactly what Jason did, and he replied with a snort.

Not really. Jason denounced the IPP h.264 encoder as being
embarrassingly crap.

Erik presented a sub-par (for us) quick test to see if he thought IPP
was in the same ballpark as x264. He seemed to conclude that it did plus
he and Frank thought it had some other benefits versus x264. He did this
without 'snorting'.

Jason then 'snorted'. (See below.)

> > But instead there's rudeness, which gets met with defensiveness to
> > rudeness and it snowballs from there...
> 
> Look back at the thread.  The last few messages aside, Erik was the
> rude one.

The rudeness starts with Jason's response of:

"So you do an encoding quality test by using a highly lossy source?  Is
this some kind of joke?"

Not polite, and Erik's response reflects this. And then you proceeded to
jump on Erik yourself. Also not polite, nor presenting a welcoming forum
for discussion and development.

Regards,
Rob




More information about the ffmpeg-devel mailing list