Thin line Problems
1 Attachment(s)
I was out shooting this morning with an EX3 and the Canon 5DmkII.
Most of my MkII video is very good, but this was very disappointing, I am attaching a still from the video. I have noticed problems with horizontal lines before but never this bad. In the video the banding moves around. It makes this shot unusable. I guess the moral is don't shoot buildings with clapboards or anything else with strong horizontal lines in the distance. The file upload does not seem to work, I will retry later. |
The only practical solution to this is to use a softening filter in front of the sensor/lens. I've done some initial testing with the Tiffen Soft/fx 1, and plan to order more filters soon.
To solve it in post requires a creative, rather than mathematical approach, given that with aliasing artifacts, the required information is already lost. |
I'm doing a shoot this weekend with my 5DM2 and was checking out wardrobe. I went down to a local fabric store and shot several samples of "thin lined" cloth.
The results were predictably sad. Some shots simply had moire and others downright jumped and danced. Amusing, but not usable. Filtering will be necessary in many cases... Wish I had more filters - and understood them better. |
My quick recommendation is a Tiffen Soft/fx 1 on an 85 or 100mm lens, fx 2 on a 50mm and fx 3 on a 24. For a 200, try a 1/2.
Bear in mind that I've only done tests with the fx 1 and extrapolated the rest. It does nothing on a 24mm, doesn't do enough on a 50mm, and is just a tad soft on a 105. In any case, you'll be balancing moire with softness. One strategy would be to run a "safety filter", such as an fx 1 on a 50mm in general, and to only run on the soft side (fx1 on a 105) when you know that you are likely to alias. Actually, on single lines, you might get some improvement by softening and then sharpening in post. With surface moire, like on fabric, you really need an optical filter. |
I just ordered the Tiffen soft fx1, I will try the same shot in a few days to see the result. I will post it as well.
Below is an attached .mov (ProRes) that shows 2 problems. From the same shoot. I am just starting to use the MkII seriously, so I am shooting with the EX3 and MkII whenever possible to get a real idea of what it can do. It compliments video cams but in my opinion is not a replacement, just another great tool in the kit. 1.) The above mentioned moire seen as focus is sharpened. 2.) The way still camera lenses breath during rack focus. This is the 24-105 F4 L is lens with a .9 ND filter. A good video lens does not do this. This is a great still lens but problematic for video rack focus work, something I do a lot. Ofcourse this can be fixed in Post, I use AE for those touches. The upload tool seems real funky today. Here is a link to the video on my website: http://westsideav.com/Problems.MOV |
I'm looking at getting the 5D Mark II do do some various film-looking projects.
As to your ultra-sharp issues, I would check my 40D, but Canon has it at the moment. Isn't there a setting as to how sharp the image is that you can set globally in the camera that can translate into the video mode. Either that, or an unsharp mask effect in FCP or AE? I guess I will be running into these issues soon. Saw some of the same thing on ESPN the other day! |
Let us know your results. I have a self-made test chart that I can upload soon, but I don't have access to it right now. Once available, just print it on a standard printer, mount and shoot.
There are some variables that I'm not sure about... 1) Does the needed filter vary based on the physical length of the lens? My guess is that it does to some degree, especially on wide lenses. What works on a physically long wide lens might be too soft on a stubby lens. 2) Does the needed filter vary depending on where you focus? My guess is no, but I haven't tested it. It seems like a large pattern at a long distance would present itself the same as a small pattern at a close distance. Lens breathing would only have a small, negligible effect. I'm hoping that there's nothing more fundamental that could affect this. FWIW, my tests were with a Nikon 105mm f/2.8 Micro AF (the older, discontinued version, not the new VR that lacks an aperture ring) with the target at about twenty feet. |
1 Attachment(s)
Here is the same shot from my EX3.
Quite a difference it was shot 1080 24p HQ. |
I know the EX3 is way more money than the 5dmkII...
But besides the aliasing, was the 5d video underexposed? The EX3 video still looks 1000x better. Look at the details in the trees on the hill behind the station... Bob C |
I had those same observations between the EX1 and 5DMkII (with pictures) in this thread.
In that kind of shot the EX series is not very far from the top of the HDTV food chain especially at 24/30p The pictured scene doesn't play to the strength of 5DMkII, but it should be much better with tight shots in low light. |
I got the FX1 filter a few days ago. I am now waiting for a sunny morning to test it on the mk2. I will post result.
I am much better at judging exposure with the EX3, I use it more and it has zebra and a histogram as well as % spotmeter readout. I am learning the mk2 though. I love the 5D but it is limited in what it can do. The new firmware is a huge help and it will inspire me to use it more. I agree in low light it is fantastic. And with my 5 L lenses, there are shots I could not get with the EX3 w/o a 35mm ground glass adapter, also a very good choice for filmic look. If I need low light, wide angle or shallow DOF this is the cam I grab. It is also great to just throw in the car along with a set of good sticks or steady-bag (I use those a lot for low angle and when traveling light. It is cheap light and inconspicuous. And the somewhat lacking codec does well if you don't have to much detail in the shot. Blurry BG and or FG helps codec tremendously. |
Have you tried to play with a cheaper lens? The EF L lenses have a great resolving power which is much higher than needed for video.
Have you tried to filters? |
This is interesting.
The gamma on the 5D is very different from the EX with the EX showing a lot more shadow detail, even with any exposure differences. The highlights on the 5D are not that far from proper exposure. I wonder if you took a still photo with the 5D if that photo would show the same result? I have always thought of my still cameras as have greater dynamic range than my video cameras. Does the video mode on the 5D put out a lesser image that if one were to shoot a still frame? Or are we used to processing the still shots and ending up with a better result? I do prefer the color of the reddish-orange from the 5D compared to the Ex. |
The EX3 has dynamic range of up to 11 stops depending on "picture profile" about 30 different settings to play with. This is truly remarkable, I think it actually has the mkII beat.
I have a few different pp's depending on light conditions. I have been shooting with the Sony for about a year now so I am pretty good at setting it up and getting proper exposure, there is a live histogram and 2 levels of zebra in the view finder. The shots of the girl in the field are both raw, no correction. I could probably pull some more detail in the shadows in Color from the mkII. As a matter of fact I can actually match those two shots so they are almost the same in Color. See this link for Images: Quote:
I have the EX3 set up so I usually need very little CC. I have not been very successful with the mkII, I just accept it needs more CC work. I may have to shoot at "6 Gun City" a wild west town in Jefferson NH. |
All times are GMT -6. The time now is 11:40 AM. |
DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network