Is this amount of artifacting normal?
 

I've had my camera for about a month now and am still getting the feel for the right settings, the threads on this sight have been very helpful so thanks everyone for that.

But one thing I'm noticing in my footage is that large areas of dark color (black, dark brown, navy blue) have blocky artifacts within the expanse of the color. Almost like the footage has been compressed for the web. Other colors look fine and my edge detail is great, it's just the dark colors that give me this problem.

heres a screenshot:

http://www.geocities.com/neechay/cat.bmp


Notice the blocky junk within the black of the cat's body?...that's what I'm talking about. I know the light in this shot is pretty low but this sort of thing shows up even in outdoor daylight stuff.This shot was taken with the camera on the default settings at 24p and captured through firewire into premiere cs3.

Also, I do notice it when I view the footage directly through a monitor so I don't think it's a capture problem.


Anyone have any feedback? Is this something I'll just have to get used to and deal with in post or is there something wrong with my camera?

thanks in advance for the help.

Welcome to DVinfo. Unfortunately, Geocities says you've exceeded your data limit and won't display the picture. If you're unable to upload it directly with your post, email the bmp to me and I'll attach it.

Hi Brandon.

It does work for me, but much as I try I don't see any artifacts even if I turn my monitor's brightness up. Possibly it is still too dark to see into cat's fur.
Do you have some other pics showing this problem?


Piotr

Hi Brandon,

The bitmap photo doesn't illustrate your problem but judging by this clip, I'd say it was more a problem with cameraman technique. A black cat is the darkest conceiveable object - filming a black cat inside a room with a large window in the background should be avoided at all costs (unless you are deliberately trying to obsure the cat for psychological effect).

Happy trails, Michael

Adam Grunseth
 

Yes, its normal. DV25 video compresses each of the frames individually- alot like how images are compressed for the web. If the frames were not compressed you would end up with a much higher data rate, it would take more storage, and you would need a more powerful computer to edit.

The quality of the compression though is handled by the camera's DSP. Personally I have always been dissapointed by the quality of Canon's DSPs used in their video gear. Several years ago when purchasing a camera I compared the Canon GL2 and the Sony PDX10- I went with the Sony because it used a 14bit DSP with significantly less visiable compression artifacts when compared to the GL2's 12bit DSP.

However there is good news! NTSC video devices have the black level set at 7.5% IRE setuplevel. That means all detail below this 7.5% IRE level will be lost and it will appear as if it is solid black (at least if the display device handles setup correctly). However, modern digital camcorders will capture a greater range of detail. For example my PDX10 captures detail all the way from 0% IRE to 108% IRE. This gives you a little bit more lattitute in color correcting the image, but some of this detail will be lost when viewing on an NTSC monitor. The contrast between the compression artifacts in your dark areas is below the 7.5% IRE level, and thus it will all appear as solid black in NTSC land. Unless of course your push the brightness up and raise those levels above the 7.5% floor.

Looks like the geocities issue is resolve as I can view that now, but in case of further problems, here are the two files that Brandon emailed me to upload.

Although I'm using my "everyday" computer with monitors that aren't precisely calibrated, I don't see any blockiness. My eye only sees apparently solid black in the underexposed black cat. I wonder if it is more of a monitor calibration issue, or as Adam is referring to, a setup issue?

Thanks to Pete for uploading those images for me.

The catthreshhold.bmp is a b/w reference pick where the shape of the artifacting really stands out (all those nice white rectangles in the cats body) to compare against the regular screenshot if you're having trouble seeing where the artifacting is.


Adam said: "That means all detail below this 7.5% IRE level will be lost and it will appear as if it is solid black"

unfortunately as I see it, the areas of artifacting are actually slightly lighter than the surrounding black, so it's not a case of it being dropped to black.


But it seems like if you all are having trouble even seeing what I'm talking about then maybe I'm being too picky...my only concern is that I'm working on a project that will eventually be compressed onto dvd and then projected onto a movie screen so I just want to make sure my image is the best possible quality.

It's normal.

I took a rectangular section of the black cat and analyzed its content. As gray scale, the average value is 9 and all the values lie between 8 and 14.

When compressing, something has to give. With DV, the amount of compression varies across the image in such a way that a portion of one area (e.g., the cat's left ear) is paired with another that is not near by (e.g., the bright light coming through the blinds). To maximize the quality of the second area (since it has a lot of detail compared to the cat), some of the detail in the black area is sacrificed.

John.