|
|||||||||
|
Thread Tools | Search this Thread |
April 5th, 2005, 06:44 PM | #1 |
Tourist
Join Date: Apr 2005
Location: oakland, CA
Posts: 3
|
verticle line problem
I've been shooting a talking head on green screen for about a week - the lighting is good - I'm getting a good key... the problem is, every time the subject moves across the screen, i get funny lines showing up vertically . They don't move with the subject - it's like he is moving through something. I should be able to pst an image of it soon
Anyone have any insights into this problem? This project is for the web so i'm shooting 30p, 16.9... Thanks for any help |
April 5th, 2005, 09:24 PM | #2 |
Wrangler
Join Date: May 2003
Location: Eagle River, AK
Posts: 4,100
|
Hi Eric,
Any chance that you can post a short clip on a web site for us to look at? The discription isn't ringing any bells yet for me.
__________________
Pete Bauer The most beautiful thing we can experience is the mysterious. It is the source of all true art and science. Albert Einstein Trying to solve a DV mystery? You may find the answer behind the SEARCH function ... or be able to join a discussion already in progress! |
April 5th, 2005, 09:34 PM | #3 |
Trustee
Join Date: Jul 2004
Location: Canton, Ohio
Posts: 1,771
|
It sounds similar to an issue that Canon had on the original XL1 in early 1998 on the initial production run. Chris has it documented well on the XL1 Watchdog page. It was caused by a chip that was picking up interference from a neighboring chip oe something of that nature. Canon recalled the first 500 or so and added a noise cancelling ciruit or something like that to these.
All the cameras produced after that point were fixed. Never heard of this with any other camera. Surely not the XL2 but there could be a defect or something. Some screen grabs would help. Heres a link to the original XL1 issue...but it is probably not really related to your current issue. http://www.dvinfo.net/canon/articles/lines.php |
April 6th, 2005, 04:17 PM | #4 |
Tourist
Join Date: Apr 2005
Location: oakland, CA
Posts: 3
|
here are some screens
http://img164.exs.cx/img164/2654/clipboard018de.jpg http://img164.exs.cx/img164/8721/clipboard024ig.jpg it's really hard to see what i'm seeing without the motion - basically the lines that I'm pointing to dont move - it's almost like the smear and/or sharpen the image on those vertical lines. most annoying when the subject moves |
April 6th, 2005, 08:15 PM | #5 |
Tourist
Join Date: Apr 2005
Posts: 3
|
45 degree fringes
I just bought an XL2 morning and sunday afternoon left it running on a tripod on a wedding gown show phot shoot. I didnt play with the settings except to turn off the camera shake option. It was put it into spotlight mode and 25P.
When i look at the footage the models when they walk seem to have lines running about 45 degrees up where they contrast against the back ground. They all show it but the green evening dresses show it worst The still parts look great but i have nikon for stills !. I realise i'm brand new to the camera but i have a $500 single CCD sony that produces better movement video and at this point i'm a little concerned. |
April 7th, 2005, 07:15 AM | #6 |
Major Player
Join Date: Feb 2005
Location: Toronto
Posts: 917
|
Well, let us be totally fair here...
Green always provides a problems for digital cameras. For those who do not know, green is not only the colour, but also luminance. The luminance chanel in all digital media green, so all the brightness stems from this colour. Bright greens are knowl to cause some problems, and heavy aliasing in Digital Cameras... This is why you use a Green Screen for digital compositing. Green shows up brighter and more evenly than any other colour. The trick is in the compositing. Getting around that nasty aliaising. You need to learn ot use After ffects or Combustion quite well... Granted, it does seem like your problems a little worse than that. I'll have my XL2 in a month, and will be shooting a feature with lots of green screen so I will post my findings then Matt |
April 7th, 2005, 07:19 AM | #7 |
Major Player
Join Date: Feb 2005
Location: Toronto
Posts: 917
|
Also, consider turning your sharpness down a bit.
|
April 7th, 2005, 08:09 AM | #8 |
Regular Crew
Join Date: Feb 2005
Location: Rainham, Kent, UK
Posts: 69
|
Could you post a full size picture, with less jpeg compression?
Also, what software are you using? This is just a wild guess at the moment, but it looks a little like the image is being resized horizontally, and it's being done in "draft" mode. |
April 7th, 2005, 08:20 AM | #9 |
Trustee
Join Date: Jul 2004
Location: Canton, Ohio
Posts: 1,771
|
It's very hard to tell from these images but this looks like DV compression issues. Particularly DV 4:1:1 color sampling. IF you can post some more larger images I can better guess. But this looks a lot like issues I have seen doing bluescreen work with DV.Do these lines show up anywhere else except at the edges where the greeen meets the talent?
|
April 7th, 2005, 11:41 AM | #10 |
Tourist
Join Date: Apr 2005
Location: oakland, CA
Posts: 3
|
Thanks for all your replies. The company is throwing a big party this evening so all the equipment is down until monday. I'll be trying to get a short video clip up somwhere then. Thanks again - Eric
|
May 19th, 2005, 02:51 PM | #11 |
Regular Crew
Join Date: Aug 2004
Location: Gloucestershire, UK
Posts: 187
|
Verticle Lines
I have just shot two clips, one after the other in order to do a video morph test. The camera was placed in record mode, first clip recorded, pause, record second clip. Stopped. None of the camera settings changed between clips and the camera was in full manual, manual focus, 25p 16:9.
I captured the clips into the PC ready for a combustion morph. The first clip has 'fixed' verticle bars (when motion is occuring) and the second clip does not. Looking at the clip Zoomed in, it appears that every few lines, a verticle bar has slipped down (as if shifted). The clip is unusable. |
May 19th, 2005, 03:05 PM | #12 |
Wrangler
Join Date: May 2003
Location: Eagle River, AK
Posts: 4,100
|
Declan,
Could you possibly post a couple of frame grabs so we can see the problem directly?
__________________
Pete Bauer The most beautiful thing we can experience is the mysterious. It is the source of all true art and science. Albert Einstein Trying to solve a DV mystery? You may find the answer behind the SEARCH function ... or be able to join a discussion already in progress! |
May 19th, 2005, 11:58 PM | #13 |
Regular Crew
Join Date: Aug 2004
Location: Gloucestershire, UK
Posts: 187
|
I have put some images up on the web.
These are 4 consecutive frames (each one is 1.2M uncompressed TIFF) http://www.xlr.demon.co.uk/img/vline1.tif http://www.xlr.demon.co.uk/img/vline2.tif http://www.xlr.demon.co.uk/img/vline3.tif http://www.xlr.demon.co.uk/img/vline4.tif For a quick look (zoomed in) http://www.xlr.demon.co.uk/img/slip.bmp Looking at them again, I am wondering if this is a DV issue as opposed to an XL2 issue. Even so, how do I stop it (or minimise it). The problem is more apparent when zooming in (which I need to do to animate masks etc). I have only viewed the footage on PC not on PAL TV screen yet. I checked some footage from my XM-2 and the same applies to that which is why I now think it's a DV compression issue. Last edited by Declan Smith; May 20th, 2005 at 12:02 AM. Reason: URL's were incorrect |
May 20th, 2005, 05:58 AM | #14 |
Wrangler
Join Date: May 2003
Location: Eagle River, AK
Posts: 4,100
|
Declan,
There are folks who know a whole lot more than me about compression and hope they'll comment. But to my eye that looks like an unfortunate coincidence: three 8x8 pixel DCT compression blocks, one above the other. In other words, yup, an artifact of DV compression. Don't know that there's anything you can do about it except re-shoot and hope that slightly differences in the re-shoot will avoid such a visible artifact.
__________________
Pete Bauer The most beautiful thing we can experience is the mysterious. It is the source of all true art and science. Albert Einstein Trying to solve a DV mystery? You may find the answer behind the SEARCH function ... or be able to join a discussion already in progress! |
May 20th, 2005, 06:23 AM | #15 |
Regular Crew
Join Date: Aug 2004
Location: Gloucestershire, UK
Posts: 187
|
Thanks Pete
In this instance, it was a test shot anyway (so I can setup a morphing sequence in combustion), however, the real shoot in a few weeks time will not be so easy to re-shoot as we only have the actors for a limited period. It is a shame that there is not a way of outputting the raw CCD image (without compression) or even having a different standard to write to tape (even if that means less run time per tape and custom driver to read into the computer). At least that way we could have got full resolution from the camera (needed for much compositing / computer minipulation). It's done with standard 35mm Digital stills cameras! If there is a way to minimise this effect I would be interested. The footage was shot in 25P 16:9 shutter speed 25, and aperture 7.2 if that helps. Cine gamma, black stretch, so don't know if any in camera settings would make a difference. |
| ||||||
|
|