No it can not do that. If you are getting dropped frames, you should work on your hardware side first (both analog w/ TBC, and digital w/ capping subsystem).
I routinely NEVER get dropped frames. At all. It would be a big surprise it it happened. So if your are getting them, something isn't set up right or up-to-par performance-wise.
No, median filter is similar to average, but statistically refers to different things (average = mean, median = middle item, mode = most frequent item).
Here's the link to a doom9 discussion on median filter use with multiple caps: http://forum.doom9.org/archive/index.php/t-159780.html, and here's a link regarding the use of the terms median, average, mean , mode, etc.: http://www.purplemath.com/modules/meanmode.htm
It keeps the common pixels and removes the outliers/spikes/dropouts. On a pixel-by-pixel basis, not the whole frame.
I think you would use this after you know to have 2 (or more, usually 3 or other larger odd number) passes of the same clip, with the identical start & stop frames and frame lengths. Which means either no dropped frames or those frames conciding with the dropped ones edited out of each pass whether it needed to or not. I would of course choose the former.
Scott
<edit>Doh! too slow for jagabo!
+ Reply to Thread
Results 31 to 38 of 38
-
-
but he said - "-Have no frame drops in the captures"
i think that when he said " frame drops", he meant something similar to what i meaning.
" frame drops" by what he said means that those drops make the video to be out of sync with other caps?
what i calling "frame drops" is a duplicated frames but it is not change the length of the video.
it is more like lost frames that didn't captured during the recording process.
is that a problem for me? -
For dropped (missing) frames you have to work in sections. Ie, sync the three videos for one section, run the median filter. Sync the video for the next section (after a dropped frame) and run the median.
If your capture card inserted a frame to replace a dropped frame a median filter should still recover the correct frame pretty well. Say, you have frame 1000 in two caps but the third cap has a copy of frame 999 instead (for some reason frame 1000 was missed and a second copy of frame 999 was substituted). A median filter should still give pretty good results. You might get some pixels from frame 999 but only in the case where that pixel falls between the values of the pixels in the other two caps. So even if it's technically from the wrong frame it's similar to the pixels from the correct frames.
Here's an example. I don't have 3 VHS caps right now so I used two videos. In the top left corner is one of the videos. In the top right is the same video, just darker to simulate a second capture that's a little different. In the bottom left is a completely different video. In the bottom right is the median of those three videos:
So even with a completely unrelated third video the median filter gives pretty good results. With a substituted prior frame you should get even better results.Last edited by jagabo; 4th Apr 2013 at 19:02.
-
Yes, and I think gil900 is under the impression that his caps mainly work that way. However, I know plenty of people who, when capping, have "dropped frames" where because the card doesn't have something ready when the buffer is expecting it (because of bad input - no TBC, or because of PC overhead problems) will just skip to the next frame.
So you end up with:
00,01,02,03...66,67,69,70,72,73,74...93,95,96,97,9 8,99. Clearly here frames 68, 71 and 94 are missing. This wouldn't be bad if it was just 3 single frame spots, but I've seen people run into multiple-frame sections and hundreds of them! No reasonable way to work on sections there.
That's why I was saying median works well given mainly good, similar captures that have a consistent timebase. And that if gil900 is having more difficult problems with timebase, hardware solutions are more appropriate.
Scott -
Thanks for the responses and help.
According to two recent comments, I understand that the game is about sync.
As I thought .. frames is not the issue. this as only makes is the problem.
It's not so bad as you might think ..
If you know how to use sony vegas.
It is very easy to make Raw caps Without sync problems.
So that's the plan of my work:
1) record 3 times
2) create a 3 row video files Without sync issues (by manual correction if needed in vegas)
From here comes AVISynth part:
3) Use this filter ..
4) Maybe to use more filters in VirtualDub..
After all, I do not think I need to TBC.
But it's thanks to sony vegas ..
Because part of patching the sync issue is unbearable with AVISynth
EDIT:
Cornucopia, The phenomenon you described, does not happen to me at all. -
-
i just wanted to share useful information about sony vegas.
I always compresses my caps the x264 encoder.
When I do the syncing repair, then sony vegas does not compress the video again.
So there is no disadvantage with syncing in sony vegas
Similar Threads
-
Change Reference Frames
By Manigorudo in forum Video ConversionReplies: 1Last Post: 22nd Jan 2013, 16:12 -
How to compare and edit individual frames of two videos side by side
By SushiSky in forum Newbie / General discussionsReplies: 8Last Post: 24th Jul 2012, 08:48 -
Stripping out no-change frames
By swkenney in forum Newbie / General discussionsReplies: 3Last Post: 2nd Aug 2011, 10:23 -
compare compression of two frames
By mathmax in forum Newbie / General discussionsReplies: 21Last Post: 27th Apr 2011, 23:19 -
Cannot compare lossless with MPEG-2 encodes: Frames aren't lining up!
By Mini-Me in forum Video ConversionReplies: 15Last Post: 5th Dec 2010, 01:08