|
|||||||||
|
Thread Tools | Search this Thread |
June 13th, 2010, 05:48 PM | #16 | |
Trustee
Join Date: Jan 2009
Location: Manchester UK
Posts: 1,212
|
Quote:
The only other logical method would be to devise a way for the colours of each part of the "offending" frame to be captured from the previous/following frame - in much the same way as MPEG compression works. This method would maintain the movement continuity and avoid the stutter but would involve more processing and thus possibly take longer. Thinking along the lines of MPEG compression, I suppose it is possible to conceive a method by which the programme would create a replacement frame based on the data and colours in the preceding and following frame (and which could be called "blending") but that sounds like even more work for the computer to do The fact that New Blue's filter doesn't remove all the flashes suggests a simple frame doubling because correction of too many near-simultaneous flashes would result in too many frame doubles and thus a major stutter. Although, as I've indicated previously, our cameras don't have rolling shutters, but I'd be interested in how effective this filter is with flashes recorded via and affected by rolling shutters. Finally, since post-pro slo-mo'd material has frames doubled anyway, any risk of stuttering caused by further frame doubling to avoid camera flashes would be increased. I doubt if anyone in this field has a genuine slo-mo camera (ie recording at a faster than normal rate) but obviously any flash recorded in this way would affect proportionately more frames and thus risk proportionately more obvious stutters. In any case it seems to me that Dave Blackhurst has it right when he says that flashes in this type of programme are part of the ambiance anyway so why risk degrading the material to make it look like something it wasn't? |
|
June 14th, 2010, 01:19 AM | #17 |
Inner Circle
Join Date: Mar 2003
Location: Billericay, England UK
Posts: 4,711
|
Hi Philip, my Canopus Speed Control program doesn't do simple frame doubling, it really does appear to blend and smooth fields together. So of course it works best with interlaced footage. And I should say my NX5 does indeed have a genuine slo-mo facility (ie records at a faster than normal rate). I must fire off some electronic flash while it's shooting (MOS) slo-mo and see what the rolling shutter makes of that.
So much to do; so little time. tom. |
November 23rd, 2010, 10:06 AM | #18 |
Major Player
Join Date: Nov 2010
Location: Santa Ana, CA
Posts: 499
|
Been playing with this with 5d mark II footage. No problems with rolling shutter (one frame and flash fills half of the frame). When it doesn't work it doubles/stutters the footage. I tried to animate the effect to ignore the flash's it has problems with, but even at 0 the effect is still being used. Tonight I'm going to try to great a subclip and use the effect only on areas where it excels at. Having one flash go off in a clip compared to five is a lot less annoying.
|
November 23rd, 2010, 02:34 PM | #19 |
Major Player
Join Date: Aug 2006
Location: Larkspur, CA
Posts: 378
|
Greg, are you saying the software does work with 5d footage (flash banding) or just some of the time?
|
November 23rd, 2010, 03:19 PM | #20 |
Major Player
Join Date: Nov 2010
Location: Santa Ana, CA
Posts: 499
|
It does work. Depends. Things like multiple flashes right after another trip it up, flash during out of focus to focus pulls, or fast pans. The way I do it, scrub to the frame with the flash showing, then try one of the three presets (you will see the frame going from half of the frame having a white streak, to it clearing up). Usually one of the presets will get rid of the flash, and the other two will cause some kind of ghosting on the footage (which you will either notice sitting on the current frame, or when you scrub the footage if it gets jerky). Worst case scenario non of the presets work, so in that case, which I'm going to try tonight, create a sub-clip without the flash removal effect and just leave that stubborn instance of the flash in. My success rate of wedding footage last night was about 75% of the time it works.
|
November 23rd, 2010, 03:49 PM | #21 |
Major Player
Join Date: Aug 2006
Location: Larkspur, CA
Posts: 378
|
Thanks for the reply. I'll give it a try out : )
J. |
November 24th, 2010, 08:58 AM | #22 |
Sponsor: JET DV
Join Date: Dec 2001
Location: Southern Illinois
Posts: 7,953
|
If you have some footage you'd like us to test with, feel free to start a help ticket and I can give you an upload location. We can then test with your footage and see if there's anything that can be done to improve results in those situations.
__________________
Edward Troxel [SCVU] JETDV Scripts/Scripting Tutorials/Excalibur/Montage Magic/Newsletters |
January 5th, 2011, 08:14 AM | #23 |
Regular Crew
Join Date: Oct 2007
Location: Harpers Ferry, WV USA
Posts: 164
|
Edward -- I use the VASST quad cam script in Vegas 10 Pro with ice skating video, 2 cameras. I have tried using the flash remover on camera tracks and it has no effect. I also tried it on the master track that quad cam creates and again no joy. It does make the video 1/4 size in the preview (with the cue marks) although the render comes out full size. Any tips on this situation?
[32 bit Vegas, OpenFX plug] Last edited by Chip Gallo; January 5th, 2011 at 08:21 AM. Reason: add environment info |
January 5th, 2011, 02:08 PM | #24 |
Sponsor: JET DV
Join Date: Dec 2001
Location: Southern Illinois
Posts: 7,953
|
I would apply the effect to the actual event that needs the flash removed after the Master track is created. Then you can apply it to only the specific events that need the effect. Then it should just be a matter of adjusting the settings necessary for that specific event.
__________________
Edward Troxel [SCVU] JETDV Scripts/Scripting Tutorials/Excalibur/Montage Magic/Newsletters |
January 10th, 2011, 09:07 AM | #25 |
Regular Crew
Join Date: Oct 2007
Location: Harpers Ferry, WV USA
Posts: 164
|
I tried applying the plug on select events in the master track and it didn't do anything. I'm thinking that the product install didn't work correctly so my next step will be to try it on an event in a test veg that doesn't have the multi cam and see if that works.
|
January 13th, 2011, 10:14 AM | #26 |
Regular Crew
Join Date: Oct 2007
Location: Harpers Ferry, WV USA
Posts: 164
|
I tried this on a single video event and it did not remove flashes. I will open a service request with the vendor to see what I may be doing wrong.
|
January 13th, 2011, 11:14 AM | #27 |
Inner Circle
Join Date: Mar 2003
Location: Billericay, England UK
Posts: 4,711
|
I think if it removed flash frames that would be overkill. At any given cake-cutting I can count 60 or 70 flashes on the timeline. Each camera person may only take a couple of photos, but many cameras give pre multi-flashes to set their camera's exposure and focus. Consequently my CMOS chipped timeline is a barrage of quarter, third and half exposed frames. Nasty. But it's a passing phase.
tom. |
January 13th, 2011, 01:10 PM | #28 |
Regular Crew
Join Date: Oct 2007
Location: Harpers Ferry, WV USA
Posts: 164
|
My video is of ice skaters in a rink. No flash photography was allowed but for some reason there was a sole photog banging away with pre-flash and flash throughout the entire show. The sales literature doesn't say how they achieve flash removal but if it is by removing the entire frame then that could be a problem.
|
| ||||||
|
|