XviD setting for Analog TV capturing

[ookzDVD]

Sorry for asking this in this forum.

I really need suggestion using XviD for analog TV capturing.

My config. is :
WinTV Hauppauge, Brooktree 878 Chip.
Athlon XP 1700, DDR PC2100 256 MB.

I'll plan to capture analog TV program,
PAL B/G @25fps at 352x288.

So... which setting of XviD you prefer due of my hardware above,
1-pass CBR ?

Thank you.

[rui]

Why do you capture at a such low resolution?
With the cpu you have, you shouldn't have any problems capturing at a much higher resolution.

Regarding xvid config, i don't capture myself, but i plan to in the future (i don't have yet a tv card), so i am interested in this thread :)

Maybe for starters, i would try to capture in quantizer mode. But this way would be more dificult to have a predicted filesize.:rolleyes:

Or, maybe, you could try to capture using huffyuv to get a perfect avi, and then compress that avi using xvid 2-pass mode? Would this work? What the experts say? :D

[amirlsm]

Originally posted by ookzDVD
So... which setting of XviD you prefer due of my hardware above,
1-pass CBR ?


I am capturing PAL TV at 384*288 on a P3-750 using virtual dub internal capture mode (vfw). Video hardware: ATI All-In-Wander.
Xvid CBR at 1000/motion-search level 6/H263/ gives resonable quality (better then my philips VCR) at 25 fps with no dropped frames.

On my P4-1500, I'm using IULAB Iuvcr (a good application for WDM capture) to capture at 640x480 resolution. Video hardware: ATI Radeon 7200 + ATI TV wander.
I use Xvid at 3000/H263/motion search level 4-5. I get 25 fps, no dropped frames with great quality.

For bset quality I use the last configuration with PICVIDEO MJPEG codec. this gives the BEST results but it uses about 700 MB/Min.

amirlsm

[unplugged]

Analog TV capture has the best fit with 704x576 resolution (PAL), because of the noise.
It's advisable to use first a temporarily fast encoder lossless or little lossy.
HuffyYUV as lossless, MJPEG codecs as fast lossy.

Then use the desired potent encoder (2pass).
Here XviD has a perfect rule, because of the interlaced support, Great!
We don't need to deinterlace frames, thus *without* losing half of detail in moving scenes.

Yes, each frame is encoded in two separate fields, first derived from odd horiz. lines the second derived from even lines... (as happen with MPEG2)

[wing1]

You guys are lucky with video feed at 25fps :D because 29.97fps is a b!#@h to capture without frame drop at high resolution frame size.

ATI AIW128 16Mb BT8x9 chipset
AMD 1.466Mhz
1Gb SDRAM PC133
30Gig 7200RPM UDMA133 IBM HD
WIN98

xvid settings
CBR @ 2000KBps/16 ( M.S.5/H263/30&6 for editing reason)
I-frame quant 1/4
P-frame quant 1/4
alt-curve
200/120/35
( Don't ask why here...My results say that it is effective. )

544x480 @ 29.97fps ( xvid codec can't handle 640x480 with my current CPU, but my end result would be 512x384 so it would not matter ).

The capture clips still have combing artifacts in fast motion area, but that is not a problem for decomb filter plugin to resolve afterward.

LoadPlugIn("<drive>:\<folder>\decomb.dll")
AviSource("<drive>:\capture_edited.avi")
FieldDeinterlace()
BicubicResize(512,384,0.00,0.75)
Sharpen(-0.20,0.25)

[ookzDVD]

@wing1,

thank you,
I think I should enlarge the resolution. ;)