A1 grainy picture problem

[James Hooey]

Here's an example of what I get outside sometimes....

Camera about 1800 feet away from trees...cold lake between camera and trees, rocky beach (not warm by any means but sunlit so it could have trapped some heat).

The trees are shimmering like heat waves.

This clip is a cropped portion of the original.

[Paul Cronin]

Hi Piotr it is interesting that the Sony also has the same problem from looking at your pictures.

James I see the same problem with your clip. The trees dancing is the same in my clip. Wonder of these low end 24p and 30p cameras can't handle zoom over 12x with out the problem. Even though I have seen my problem at wide just not as easy to see.

Has anyone had this problem with the H1?

[Piotr Wozniacki]

Originally Posted by Paul Cronin

Hi Piotr it is interesting that the Sony also has the same problem from looking at your pictures.

James I see the same problem with your clip. The trees dancing is the same in my clip. Wonder of these low end 24p and 30p cameras can't handle zoom over 12x with out the problem. Even though I have seen my problem at wide just not as easy to see.

Has anyone had this problem with the H1?

Paul, my "problem" is definitely the hot air convection. Been there again at different time of a day, shot exactly the same place - everything is clear and steady in full telephoto.

[Bill Busby]

My A1 doesn't do this & I'm curious why anyone hasn't done this full zoom test somewhere indoors, to rule out or confirm this hot air theory.

Bill

[Jim Miller]

This discussion had me worried. I just got my A1 on Monday so I did some testing today. This is what I have found.

test 1. Shot at full zoom (20x) at a home about 100 yrds away (60i f6.2). On the roof shingles I saw the moving image that has been described. In fact most of the lighter parts of the shot had some movement. i.e. bricks on the front of the house that were in shade

test 2. Shot at full zoom with flowers and foliage about 10 feet way. No image degradation at all in any part of the picture. White flowers to dark green leaves - tack sharp

test 3. Full zoom at a tree with two squirles about 125 feet away (focused on the tree. A home with trees somewhat out of focus on either side of the tree but at about 300 feet away. Again the tree was tack sharp with no image problems but the same movement/shimmering from test 1 was visible in the home and foliage in the background.

test 4. shot the same scenes but with my HV20 at 10x zoom. No noticible movement or shimmering noted.

Temperature was approximately 78 degrees

These tests have led me to believe that atmospheric interference is the culprit coupled with the 20x zoom. With shorter distances between the camera and subject no problems were encountered. Yet in test three I did see the shimmering movement but only in the distant parts of the scene, not the nearer subject matter.

While heat from road surfaces, roofs or other heat collectors will definately produce distorted images when compressed in a telephoto image I don't believe this was the case here.

Microscopic dust, pollen, and other air particles in the air that can not normally be seen are visible when the mass of those particles are compressed in an image with an HD camera and 20X magnification. With sunlight being very specular it will be reflected at different angles and cause the "shimmering" effect. Because the HV20 only has a 10x zoom the air mass is not compressed as much.

Looking through my 55power telescope I see similar shimmering at angles parralel to the ground. When pointing up the are lessoned.

These are just my observations. This is not to suggest that your camera does not have a problem.

[Petri Kaipiainen]

my 0.02€: I think what we are seeing here is the combined effects of thermal currents and, as the main culprit, the MPEG compression combined with 1440/1920 pixel stretch.

When there is movement in every pixel between every frame of video, MPEG compression can not handle it. It has to record in bigger chunks and you get grainy picture, basically the resolution drops to half. Camera is not at fault, it is the HDV system which is not good enough. This is made worse by the fact that the picture is recalculated from 1440*1080 to 1920*1080 for display further making the horizontal (edit from "vertical", see below) resolution worse (than the native resolution of a HD display).

We have to remember that HDV is not a "real" HDTV system, it is a system that gives a fairly good approximation of HDTV at amazingly cheap price. We can not expect the impossible. For the impossible just invest 10x the money...

Run a test by recording also native SD of the same situation. If that footage is good, within the limitations of less resolution, then in MPEG did it.

[Piotr Wozniacki]

Originally Posted by Petri Kaipiainen

my 0.02€: I think what we are seeing here is the combined effects of thermal currents and, as the main culprit, the MPEG compression combined with 1440/1920 pixel stretch.

I'd tend to agree that the effect is exaggerated by MPEG compression.

[Jim Miller]

Petri, your explanation make perfect sense

[Richard Hunter]

Originally Posted by Petri Kaipiainen

This is made worse by the fact that the picture is recalculated from 1440*1080 to 1920*1080 for display further making the verical resolution worse (than the native resolution of a HD display).

Hi Petri. Sorry I got lost on this part, how is the vertical resolution affected by the 1440 non-square pixels being stretched to 1920 on the display? I thought the vertical resolution (1080) was unchanged. Or do you mean horizontal resolution?

Richard

[Paul Cronin]

Hi Bill I did do a test inside hooked up to my monitor and not running the tape. Same problem. Test was 4 meters away and the whole picture was moving. Also if I put the lens cap on the information on the screen was moving. This proved it was not the heat. All while locked on a tripod. To see my setting go to page one.

As for HDV limitation I agree we can not expect too much but the Sony Z1 never had this problem so that made me suspect the camera. And I hope Canon will resolve so I can at least get similar restults to the Z1 on the A1 which I know is possible.

I appreciate all the help and will post with the second new A1 and the serviced A1 when they arrive.

[Steven Cowie]

If its an HDV compression or pixel shift related issue (exacerbated by being at that end of the lens) then it would be a constant - Piotrs stable results at his second shoot of same location at full tele suggest it isnt.

Likewise switching to DV from HDV has no impact on the effect on my units. Nor does reframing the shot to cut out all movement from the trees.

If it was a thermals related issue (that does seem to have been involved in Piotrs initial shoot) then Id at least expect a change in the intensity of the effect depending on time of day/night the footage is being taken - whereas the issue on mine (unlike Piotrs) remains the same regardless.

[Piotr Wozniacki]

Steven, if it is like you describe, then you're facing some problems with the camera; I hope Canon will fix it for all of you affected. Yes, I can confirm - when I first saw the shots I posted the 2 grabs from, I almost panicked, because the air movement was not visible to my eyes at all while shooting. I tried to recreate this phenomenon in other surroundings, as well as the original one but at different time on the day after - it's perfectly still and clean (I'm talking about the V1E, of course).

However, since as I said the heat wave was not visible to human eyes (at least I didn't notice it; I wasn't looking for it either), I think the comment above on the HDV compression pronouncing this effect might be quite valid. Human eye (or rather the brain) "knows" what we see looking far ahead is 3D space; to the camera's imagers it's just several kilometers deep amount of air, all compressed into a 2D scene; the probablility nothing moves in it is very small!

[Petri Kaipiainen]

Originally Posted by Richard Hunter

Hi Petri. Sorry I got lost on this part, how is the vertical resolution affected by the 1440 non-square pixels being stretched to 1920 on the display? I thought the vertical resolution (1080) was unchanged. Or do you mean horizontal resolution?

Richard

My mistake, horizontal resolution is ripped apart a bit like you say.

[Paul Cronin]

Well the new A1 arrived today and I went through the following test.

1. Side by side with HV-20 hooked up component to one monitor and A1 on the other with both on tripods. Not running tape just going through presets and added two presets. Zoomed to full even though they are different then 1/2 zoom then full wide all looking out the window with window open on a cool clear day. A1 same problem.

2. Inside test side by side no tape running both zoomed in on US map where each state is diffenent color 3 meters distance. A1 same problem.

3. Running tape out side with just A1 on tripod zoomed at 10 meter, 5 meter, then 1 meter. A1 looked very nice at 1, 5, but at 10 started to see noise/grain/movement.

4. Running tape went to previous locations that I shot with my other A1 which Canon agreed there is a problem. Not as bad as the camera I sent back but not even close to the Z1 crips picture. Over 10 meters the picture is not up to my standards.

All test were manual setting with OIS off, gain off, IAF off.

I am sending the second camera back tomorrow and will replace with ?? in the next few days. Sorry I did not have better news I was excited it would be resolved with the new camera. This is not how I planned on spending this week with two jobs now pushed off two weeks so I can resolve. My clients are asking for 30P which is why I sold the Z1 and went A1. Now I need to decide in the next day which direction to go.

[Piotr Wozniacki]

Paul, for your reference, here is a link to a clip I shot with my V1E (zooming all the way from wide to tele; apologies for the ugly content- some "socialist reality" remnants here in Poland):

http://rapidshare.com/files/30129684/full_tele.m2t