|
|||||||||
|
Thread Tools | Search this Thread |
December 5th, 2006, 09:49 AM | #1 |
New Boot
Join Date: May 2006
Location: Salt Lake City
Posts: 11
|
Severe blocking artifacts in red colors
Hey all,
This is a problem I have been noticing recently on several projects I have done. If you refer to the attached image, you can see that all the colors in the image appear clear and correct (ignoring any small jpeg artifacts) but that the reds on the woman's shawl are very blocky and messed up. Like I said, this has happened on several different occasions, and only occours with the color red. I am shooting in 24pa, and no artifacts are visible at the the time of recording, only during playback either on the XL2, in FCP, or on one of my broadcast monitors. I have been searching the forums for any answers to no avail. Perhaps something to do with DCT compression???? Any ideas would be appreciated!! Thanks!! Last edited by Ryan Black; December 5th, 2006 at 11:36 AM. |
December 5th, 2006, 10:45 AM | #2 |
Wrangler
Join Date: Jan 2005
Location: Boulder, CO
Posts: 3,015
|
reds are notoriously challenging to DV. if you're working with clients, advise them not to wear it for the shoot. no stripes either. no stripes, no reds. some things that can help mitigate the effects of bouncing reds include de-interlacing and flicker filters....
|
December 5th, 2006, 11:06 AM | #3 | |
Regular Crew
Join Date: Jun 2006
Location: Paisley, Scotland UK
Posts: 99
|
Quote:
__________________
http://www.myspace.com/theeejit |
|
December 5th, 2006, 11:20 AM | #4 |
Regular Crew
Join Date: Jul 2006
Location: Tampa, Florida
Posts: 261
|
I have wondered this too. Why are they a problem? Its seems to me that because of the 4:1:1 compression that both blue and red should be equality as bad but blue seems to be okay most times. I have only had problems with blue when I tried blue screening stuff and learned that it bleeds on some occasions. So I use green now. But why is read a hard color to capture and how can we correct the problem either in camera (maybe some kinda modification) or in post?
|
December 6th, 2006, 10:42 AM | #5 |
Major Player
Join Date: Jun 2006
Location: chattanooga, tn
Posts: 721
|
If your NLE has a chroma blur effect, try using it when you see chroma blocking like this. It will blur only the color information in the clip (therefore not blurring the b&w luma part at all), effectively "smoothing" out those blocky edges in your reds. Depending on the severity of the problem, which depends mainly on the shade of red in question and the level at which the red is exposed, this approach is often just the little magic trick that everyone is looking for.
If your NLE doesn't have chroma blur--Vegas is the only one I know really well--then you can try the Magic Bullet deartifactor, which is just a fancy name for chroma blur. It comes with Magic Bullet. It's nowhere near as adjustable as Vegas' chroma blur, for example, but it's more or less the same thing. |
December 6th, 2006, 11:07 AM | #6 |
Trustee
Join Date: Feb 2005
Location: Worldwide
Posts: 1,589
|
The problem is that red is a very important colour to the human eye and it can be very difficult to avoid filming a subject that has some red in it, and more often, especially in nature, it is a vitally important colour to include in some scenes or landscapes.
Yes, deep or bright reds have always been the most difficult to capture cleanly, especially with sunsets over water. The older XL1s was far worse in this respect and I think that the XL2 copes far better with red subjects. |
December 6th, 2006, 12:35 PM | #7 |
New Boot
Join Date: May 2006
Location: Salt Lake City
Posts: 11
|
Thanks for your input guys. It is weird how it only seems to happen with specific hues of red. And I know that it is not a defect with my specific camera, because I am running a multicam XL2 setup, and it is affecting all of them. Maybe I just found the "compression sweet spot" with red, where the in-camera compression freaks out.
Jarrod, I tried out the "chroma blur" filter (in FCP 5 its the "Channel Blur" filter) and while smoothing the reds a little, it seemed to just add in more fringing and blocking artifacts around the red on the subject. The other thing I thought of, is that perhaps it is the contrasting colors of red on blue? They are on opposite ends of the light spectrum if I remember correctly, and that may contribute to the issue. But then again, the blocking is prevalant throughout the red tones, not just the edges.......I give up!!! |
December 6th, 2006, 12:38 PM | #8 | |
Major Player
Join Date: Jun 2006
Location: chattanooga, tn
Posts: 721
|
Quote:
|
|
December 9th, 2006, 02:57 PM | #9 |
Regular Crew
Join Date: Jul 2006
Location: Tampa, Florida
Posts: 261
|
So can someone please explain scientifically why this happens? Like I said above it seems because of the 4:1:1 compression that it would happen equally in the blues, but I don’t seem to notice.
|
December 9th, 2006, 05:15 PM | #10 |
Major Player
Join Date: Jun 2006
Location: chattanooga, tn
Posts: 721
|
I've always thought that it happens with all colors, but that it's just more noticeable with red. If you look very carefully, you can often see the same kinds of blocking on greens and blues--it just doesn't stick out as much.
That's not very scientific, and might possibly be a load of crap. Sorry. :) |
December 10th, 2006, 02:02 PM | #11 |
Trustee
Join Date: May 2003
Location: Tulsa, OK
Posts: 1,689
|
Green is the cleanest channel, blue is the sharpest, red is somewhere in the middle. The result, blues get noisy due to the sharpening effect, greens are VERY clean and reds are neither! Chroma blur will only help if you are not recompressing to DV. This is generally a bigger problem when you have warm white balances, the saturation turned up or if the the reds are a little overexposed.
When I shoot something with lots of red I turned down the red gain, turn UP the coring and down the sharpness a little. Master pedestal going down a bit helps as well. I generally am able to avoid this almost every time, it just takes some practice. ash =o) |
December 12th, 2006, 08:52 PM | #12 |
New Boot
Join Date: May 2006
Location: Salt Lake City
Posts: 11
|
Thanks Ash I will try those settings next time I am faced with this problem.
|
January 6th, 2007, 08:10 AM | #13 |
Regular Crew
Join Date: Apr 2006
Location: Malmo Sweden
Posts: 103
|
Also can be because of the fact that the ccd´s have a very non linear spectral sensitivity much higher in the red area (for example look at http://www.ph.tn.tudelft.nl/Courses/...-Spectral.html )
The camera manufacturers try to compensate that curve electronically but i suspect that many times the reds get saturated so the ccd "pass" a clipped information to the electronics. I´m a PAL user and I saw many times the blocks in saturated reds. |
| ||||||
|
|