Did you get the distortions with CVS version too ? If yes, then
please give me a link to a short *.dv file (2 frames) to check this.
Which render mode did you use YUV2, RGB, mmx or c version?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I reproduced the problem in non-asm mode too because I was
debugging an issue with some non-asm code. Also, it
reproduced using both the shared lib encode functions as
well as statically linked encodedv utility. Looks like a
general problem...
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Peter Schlaile came through again and sent me a patch. I
applied it to my working copy, and I have been testing it.
There were a few bugs I fixed. So far, I have working good
the asm and non-asm versions of the ppm and pgm input
filters as well as the ppm-based library encode interface.
If I recompress your test image 5 times, there is still some
very noticable artifacting in the rightmost 16 pixels, but
when encoding the test image once, it is hardly noticeable.
So, the improvement is huge even though there is always room
for improvement.
I need to further test some new encoder functions that write
metadata/timecode, and I plan to commit these changes to CVS
tonight.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Logged In: YES
user_id=38475
Did you get the distortions with CVS version too ? If yes, then
please give me a link to a short *.dv file (2 frames) to check this.
Which render mode did you use YUV2, RGB, mmx or c version?
Logged In: YES
user_id=83549
http://heroinewarrior.com/test.dv is a frame encoded with
the cvs. Compiled with ARCH_X86 and ran on Pentium 4. All
options to dv_encoder_new are 0.
Logged In: YES
user_id=78682
Ow, that test frame hurt my eyes, dude!
I reproduced the problem in non-asm mode too because I was
debugging an issue with some non-asm code. Also, it
reproduced using both the shared lib encode functions as
well as statically linked encodedv utility. Looks like a
general problem...
Logged In: YES
user_id=78682
Peter Schlaile came through again and sent me a patch. I
applied it to my working copy, and I have been testing it.
There were a few bugs I fixed. So far, I have working good
the asm and non-asm versions of the ppm and pgm input
filters as well as the ppm-based library encode interface.
If I recompress your test image 5 times, there is still some
very noticable artifacting in the rightmost 16 pixels, but
when encoding the test image once, it is hardly noticeable.
So, the improvement is huge even though there is always room
for improvement.
I need to further test some new encoder functions that write
metadata/timecode, and I plan to commit these changes to CVS
tonight.