Hi All,
I was browsing through the code and saw line 253 of filedv.C
decoder->quality = DV_QUALITY_BEST;
and wondering what the "non-best" looks like (and considered that for
simple editing, it might be faster to have this turned down, and only
turn it up for the final render). I saw in
/usr/include/libdv/dv_types.h that there were a number of quality settings
possible. I was rather curious and thought I'd benchmark them. Below
are my results for those interested.
The test file is available from
http://cvs.cinelerra.org/~pmdumuid/test* and there are 3 test (made by
enabling / disabling combos of the tracks within). All
samples rates for assets and project were set to 100 (I assumed
nothing was gonna play that fast!) And I had "play every frame"
selected and "cache" deselected. The test were:
(1) track properties match asset. (this should be a direct pass-through
test)
(2) 2 tracks: same source / time, but top track has an 0 < alpha < 100.
(3) 1 track, but 0 < alpha < 100
DV_QUALITY_FASTEST: | DV_QUALITY_COLOR:
X11, YUVA, no alpha 1 track: 63 fps 62.6 fps
X11, YUVA, alpha 2 track: 16 fps 15.9 fps
X11, YUVA, alpha 1 track: 23 fps 23.9 fps
DV_QUALITY_AC_1: | DV_QUALITY_COLOR:
X11, YUVA, no alpha 1 track: 55.6 fps 52.7 fps
X11, YUVA, alpha 2 track: 14 fps 14.5 fps
X11, YUVA, alpha 1 track: 22.7 fps 22. fps
DV_QUALITY_AC_2: | DV_QUALITY_COLOR: (same as
BEST)
X11, YUVA, no alpha 1 track: 52.6 fps 39.2 fps
X11, YUVA, alpha 2 track: 14.27 fps 12.2 fps
X11, YUVA, alpha 1 track: 21.6 fps 19.4 fps
Notes:
DV_QUALITY_FASTEST is somewhat unbearable to work with as an
intermediate format.
It decodes it with roughly 10x10 pixelized blocks..
Possible good for thumbnails though!
DV_QUALITY_AC_1 is not much different to AC_2 in my (color blind
opinion) but there are obvious speed differences! I actually just
double checked the results for DV_QUALITY_AC_1 || DV_QUALITY_COLOR,
(cause I couldn't believe the numbers) and yes, it is much quicker!
I just thought people might be interested in these results. I'm not
using DV material that much at the moment so won't be rushing to code it
up, but I thought someone might, or might be interested to try this out
themselves.
I basically entered the hvirtual/cinelerra/ directory,
# (1)
]$ make; ./cinelerra /tmp/test.cine.xml
# run tests
# edit filedv.C and go back to (1)
Pierre
[EMAIL PROTECTED] ~]$ cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 13
model name : Intel(R) Pentium(R) M processor 1.70GHz
stepping : 6
cpu MHz : 600.000
cache size : 2048 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr mce cx8 mtrr pge mca cmov pat
clflush dts acpi mmx fxsr sse sse2 ss tm pbe est tm2
bogomips : 1201.72
[EMAIL PROTECTED] hvirtual_build]$ more cinelerraconf.sh
#/usr/bin/bash
echo configureing for pentium-m
./hvirtual/configure --prefix=/usr --enable-x86 --enable-mmx
--enable-freetype2 \
--with-buildinfo=svn \
CFLAGS='-O3 -pipe -fomit-frame-pointer -funroll-all-loops
-falign-loops=2 -falign-jumps=2 -falign-functions=2 -ffast-math
-march=pentium-m -mfpmath=sse,387 -mmmx -msse' \
CPPFLAGS='-O3 -pipe -fomit-frame-pointer -funroll-all-loops
-falign-loops=2 -falign-jumps=2 -falign-functions=2 -ffast-math
-march=pentium-m -mfpmath=sse,387 -mmmx -msse' \
CXXFLAGS='-O3 -pipe -fomit-frame-pointer -funroll-all-loops
-falign-loops=2 -falign-jumps=2 -falign-functions=2 -ffast-math
-march=pentium-m -mfpmath=sse,387 -mmmx -msse'
X11 was actually faster than X11-XV, and I'm using an Inspiron 510m
machine with an annoying intel graphics chip..
--
__________________________________________________________________
Get Firefox! http://www.mozilla.org/products/firefox/
Mr Pierre Dumuid
Phd Student
Active Noise and Vibration Control Group
School of Mechanical Engineering
The University of Adelaide
SA 5005
AUSTRALIA
Email: [EMAIL PROTECTED]
[EMAIL PROTECTED]
Work: +61 (0)8 8303 3847
Mobile: 0407 570 263
Web Page : http://www.mecheng.adelaide.edu.au/
__________________________________________________________________
CRICOS Provider Number 00123M
-----------------------------------------------------------
This email message is intended only for the addressee(s) and contains
information that may be confidential and/or copyright. If you are not
the intended recipient please notify the sender by reply email and
immediately delete this email. Use, disclosure or reproduction of this
email by anyone other than the intended recipient(s) is strictly
prohibited. No representation is made that this email or any attachments
are free of viruses. Virus scanning is recommended and is the
responsibility of the recipient.
begin:vcard
fn:Pierre Dumuid
n:Dumuid;Pierre
org:The University of Adelaide;Mechanical Engineering
adr:;;;Adelaide;South Australia;5005;Australia
email;internet:[EMAIL PROTECTED]
title:Postgraduate Student
tel;work:8303 3847
tel;home:8388 5727
tel;cell:0407570263
note:CRICOS Provider Number 00123M
x-mozilla-html:TRUE
url:http://www.adelaide.edu.au
version:2.1
end:vcard
