ati 8500 dv card capping q's

[tomos]

i use dvd2svcd to make caps from ati's tv proggy into dvd compliant video (i know it can cap straight to it but i need to edit-crop etc).

my question is what are the best settings?

i use visual masking and capture interlaced video since i heard that all ati's deinterlace does it to throw a field away. am i right in capping interlaced video and using smart deinterlace (blend field sin dvd2svcd)? or is there a better way?

ps. i cap to max data rate and full motion estimation.

and i never get any dropped frames.

all in mpeg-2 btw. :)

[cjv]

If you're using a high enough bitrate, always capture it interlaced. I wouldn't trust ATI's on-the-fly deinterlacing to produce optimal results. You asked if there's a better way...if you're making SVCDs, why are you deinterlacing in the first place? You of course should IVTC if applicable, but keep 29.970 video interlaced.

You mention capping at 15mbps...why don't you just capture to MJPEG instead? It may take slightly more hard drive space, but it is easier to work with and edit, the quality will be much better...plus, MJPEG is designed for capturing. That way you don't amplify the MPEG artifacts by encoding to MPEG2 twice?

cjv

[tomos]

should have said actually that i am in the UK so use PAL so dont think ivtc applies does it? :confused:

i tried capping to mjpeg (that's what i used when i had a G400) but for some reason the quality was much lower with mjpeg on the 8500 dv than i had with the g400, or capping to mpeg-2 with the 8500 dv.

i dont really get why that is though. i tried huffyuv as well but same low qual results. :( since then i stuck to mpeg-2

also, i am not capping to svcd, but to DVD (i use maestro to make menus etc :) )

[cjv]

Originally posted by tomos
...i tried huffyuv as well but same low qual results...
This is interesting. But then again, we don't know what you call "low qual". Huffyuv is lossless so intheory it should be the best. Compressing to MPEG2 will give the video a slight blur, maybe this is more pleasing to you.

I know for a fact that MMC does not capture to AVI well...but mostly with regards to dropped frames, and not quality. Maybe try another capture app?

If you stick with MMC/MPEG2, I suggest capping interlaced, and capping only "I-frames." Also try without visual masking. You want your captured file to be as "untouched" as possible. ATI's visual masking is done on-the-fly like the deinterlacing, and its purpose is to conserve bitrate, which you don't need.

Hey, post some raw screenshots in the "Samples" thread if you get a chance.

cjv

[Jeff D]

You can tell MMC to capture AVI and that's an uncompressed format, bigger than HuffYUV.

What do you mean by 'low qual'?

I am having problems with HuffYUV, and I was questioning the bit depth of the capture. HuffYUV states 16 bit, I understand YUV is 16 bit and I understand the ATI's native capture format is a YUV colorspace format. But, even when the decoder displays the huffyuv debug info it states 16 bit output... My results were looking like 16 bit RGB. There color banding, play this back on a DVD and it looks BAD. I understand a bit about colorspace formats, but I'm not expert.
(And my monitor display depth is 32bit)

[tomos]

ahh, i keep forgetting details.

when i capped to avi (mjepg/huffy) i used avi io (which i used to cap to mjpeg with my g400)

as i said before, the video quality with avi isn't as good with this card as i had with my g400. mpeg-2 is the best quality i can get.

i'll try switching off visual masking next time i cap and do 2 encodes. 1 with deinterlace and one without and see what they look like.

on my monitor tho (which i use mostly as a tv) interlaced tends to look choppy or something.

the video basically looks like there's noise in parts of it, and since i cap from satellite i know there's not much noise.

i mean 1 day i was capping with my g400 and the image was almost perfect., then i switch cards over and hey presto, lower quality. :confused: