Log in

View Full Version : why no intrlcng artfcts when watching, but appear after capture?


grigory
19th March 2002, 05:52
Please forgive a possibly stupid question, but I am new to video capturing, and although I have spent quite a few hours researching and testing, I am still not clear on this:

1. Why do I get severe interlacing artifacts after capturing a video from VCR, but no such problems appear when I simply WATCH a tape through the computer via any capture software? If interlacing problem does not appear during real time playback, shouldn't there be a way to completely do away with it on actual capture without all the filters?

2. This may sound really stupid, but again, if these problems don't appear during playback, wouldn't it be possible to "capture" the movie directly from the display screen, rather then from the capture card? I know it doesn't sound technically correct, but wouldn't it solve the problem?

Guest
19th March 2002, 06:12
1. Capture software, such as AVI_IO, typical shows only one field on the overlay or preview during capture. But when both fields are *captured* and frames are assembled from them, you'll get interlacing artifacts. Yes, most capture software will allow you to capture only one field and thereby avoid interlacing, but many people prefer an adaptive approach that performs deinterlacing only in "moving" areas. And if you have telecined material, throwing away a field amounts to unnecessarily discarding one half of your vertical resolution.

2. It's not possible when the overlay is used, but in any case, you get the same effect by configuring your capture software to capture only one field.

grigory
19th March 2002, 06:31
Thanks for your answer!

Just to confirm your point - so, when I simply watch a VHS movie through let's say ATI Multimedia Center, I don't get the same result as when I watch it through a TV set? Please, say it's not true - I threw out my TV after a bought the capture card! :)

And one more question on deinterlacing - I have tried using the Telecine filter and it really works great, but I would like to know: if I mistakingly use Telecine filter on a VHS source that was NOT made from a film (like a TV show), is the filter "smart enough" to know that and not drop neccessary frames, or do I have to be sure of my source? Also, is it helpful to use FieldDeinterlace after Telecine to completely smooth out the combing, or is that redundant?

Thanks again!
Grigory

Guest
19th March 2002, 23:48
Just to confirm your point - so, when I simply watch a VHS movie through let's say ATI Multimedia Center, I don't get the same result as when I watch it through a TV set? Please, say it's not true - I threw out my TV after a bought the capture card!
I don't know what ATI does in that SW, but knowing them, they are just discarding a field. There *is* software that works with some chipsets (DScaler) that can do adaptive deinterlacing on the fly. But even with that, you are not going to get the equivalent of your TV, because the computer monitor is a progressive display and your TV is interlaced.

if I mistakingly use Telecine filter on a VHS source that was NOT made from a film (like a TV show), is the filter "smart enough" to know that and not drop neccessary frames
First, if you are referring to my Telecide() filter, it does not drop frames. Assuming you are talking about Telecide()... If you erroneously pass pure interlaced material through it, it will still attempt to field match and will usually match the fields in the current frame and so pass the frames through. This is useful for dealing with hybrid film/nonfilm material when used in conjunction with Decimate(mode=1). Decimate() is the guy that either will or will not drop frames, depending on the setting of the mode parameter. It can't intelligently decide; you have to tell it.

is it helpful to use FieldDeinterlace after Telecine to completely smooth out the combing
RTFM! (hint: look at the post parameter)

grigory
20th March 2002, 01:07
Thank you!