|
|||||||||
|
Thread Tools | Search this Thread |
October 19th, 2006, 01:58 PM | #1 |
New Boot
Join Date: Oct 2006
Location: Canada
Posts: 5
|
HC1, A1U Rolling Shutter Question
I'm looking to get an HDV camera to film skiing, snowboarding and mountain biking. I really like the small size of the HC1 and A1U but I have been reading a lot about the rolling shutter issue with them. My question is will this always be a problem if the camera is panning fast? Are there any ways to avoid it? And also, has any one had any experience shooting skiing or biking with the HC1 or A1U that has run into any problems?
Thanks, Mike |
October 19th, 2006, 02:08 PM | #2 | |
Join Date: Jan 2004
Location: Stockton, UT
Posts: 5,648
|
Quote:
In still shooting, you'll occasionally see it crop up. At super high shutter speeds with a bouncy camera, you might see it raise its head. It's rarely seen, and more an issue brought up by measurebators vs actual users of the cam. We've mounted them on skateboards, motoX, BMX, skydiving helmets, and other high speed/high action/high risk sports, and not had any issues.
__________________
Douglas Spotted Eagle/Spot Author, producer, composer Certified Sony Vegas Trainer http://www.vasst.com |
|
October 19th, 2006, 02:49 PM | #3 |
Major Player
Join Date: Mar 2006
Location: Helsinki, Finland
Posts: 393
|
I'm am an actual shooter and I do high speed action shots with the hc1. It is a considerable problem and it forces me to do shots differently than I would with a normal camera. It also makes camera shaking unbearable.
Here's a small action test I made with the hc1. You can't really see it because I cut out all of the shots that showed it except for one at 01:05 and at 02:54 (camera shaking makes the screen wobble nastily, had to cut out some frames that were distorted.) Cutting frames out improved it alot though so those distortions aren't almost there anymore. It's about 200 megs. http://hmcindie.pp.fi/suomenlinna/En...%20quality.wmv It amazes me that people don't notice it. It's like a cellphone camera in that regard. Except it produces astonishing video. The great image quality kinda makes up for it. Wouldn't use it for high action sports but some people apparently do and don't notice anything. Weird. There is NO WAY to do anything Michael Bay like camera shaking that's all over the place with a high shutter. It will look horrible. I shot some water skiing stuff from a moving boat and it looks pretty usable, but when you deinterlace it the effect becomes more noticeable. And there's no way to correct them in post except to cut out distorted frames, which I've been doing from time to time. It also causes some funny stuff like people who are falling are stretched. Going up and down stretches/compresses and sideways tilts the moving part. Annoying. There was one producer who complained loudly about the effect as he didn't realize it was there before he actually bought it. And neither did I, it ain't advertised much. Scanning speed (top to bottom) for the hc1 is about 1/60, so using a low shutter does hide the effect somewhat. Maybe that's the reason Sonys new avchd cameras (with the same sensors) don't allow as high a shutter speed as the hc1. There's a good reason the sensor is 4:3 in size and then cropped to 16:9. If you'd see the complete picture then you'd be amazed at the amount of distortion in the lower and upper parts. Just like in cellphone cameras. And you actually WON'T see it crop up in still shooting, the effect disappears when going into camera mode. It was pretty strange to me that people weren't complaining about it. It's a much much much worse problem than the "red inaccuracy". Atleast you can correct that in post. |
October 19th, 2006, 03:02 PM | #4 |
Join Date: Jan 2004
Location: Stockton, UT
Posts: 5,648
|
Yes Mikko, you've been very vocal about the rolling shutter issue.
A few observations; You're shooting PAL. That may make a difference. You're shooting high speed shutter. That definitely makes a difference. It's a tool. Knowing how to actually use the tool makes a difference as to the quality of the image that you end up with. Knowing what situations to avoid with a specific tool is equally part of understanding how to best use it as well. I don't own an HC1, but have several A1U's. Maybe the problem is limited to the HC1 (though I'd doubt it's any different). I too, am an "actual shooter" and I'm happy to challenge anyone to put the camera through the paces I've put them thru. Maybe the engineers at most of the bigger broadcasters just don't see it either; perhaps they need training. ESPN, MTV, CNN, and many other broadcasters use this camcorder regularly. Look closely enough, you'll see it in the background of several Discovery broadcasts.
__________________
Douglas Spotted Eagle/Spot Author, producer, composer Certified Sony Vegas Trainer http://www.vasst.com |
October 19th, 2006, 03:08 PM | #5 | ||
Major Player
Join Date: Mar 2006
Location: Helsinki, Finland
Posts: 393
|
Quote:
Quote:
So don't take my ramblings as the "be all end all", go test the camera out. If you don't notice it, then maybe it's for the better. |
||
October 19th, 2006, 07:37 PM | #6 |
Trustee
|
Well, at least it's a great camera for those of us whom shoot with both feet planted firmly on the ground. I can't emagine jumping out of an airplane with a perfectly good camera.... (or was that a jumping out a perfectly good airplane)?
|
October 21st, 2006, 10:37 AM | #7 |
New Boot
Join Date: Sep 2006
Location: Dallas, TX
Posts: 13
|
The HC3 also shows it but I didn't really take note until someone asked if it had that "problem." Having used all sorts of still and motion picture cameras over the years, I know that various designs have characteristics which can cause artifacts like this.
I've had really distorted images from some motion picture cameras under certain conditions but that didn't keep me from using them. I've seen photos from still cameras which showed huge distortions depending on the circumstance. I don't see this as a problem but only as an idiosyncracy of this camera. That's my opinion and I know that others' opinions are much different. |
November 19th, 2006, 05:28 PM | #8 |
Regular Crew
Join Date: Dec 2005
Location: Helsinki/Finland
Posts: 131
|
Hi,
Due to my official daytime work I have been playing around with different CMOS sensor chips -with rolling shutters - and without. Typically the aterfacts, due to the rolling shutter, are clearly visible at the slower shutter speeds, not the faster ones. I think that this effect is very prominent in the HC1, especially when you know how it shows up in the image. First I did not understand what was wrong with some video images of moving objects. Specially when I shot large objects that moved at a reasonable speed. The I read about that the sensor has a rolling shutter. That explained everything. Try panning with a wide angle lens at a reasonable speed using a tripod (and without the image stabilisator) - I think you will see the artefact. If not - compare with another cam with an ordinary shutter - and you will notice the difference. Knowing this behaviour - or limitation - directs you to use this tool for what it is best for - shooting in good light - and objects that moves only at moderate speeds - and preferrably not red ones ;) My feeling is that the red color rendering is a bigger problem than the artefacts in the video due to the rolling shutter. Electronically this is, an elegant and brilliant solution... But as we all know, there is nothing like a free lunch... I just compiled some HC1 HDV footage directly in Vegas 7 - in it's native format - and I must say that I am very pleased with the results. I just got the cam back from Sony and it seems that the reds are more balanced after the service tour... that is my first impression anyhow. I am going to shoot some color tests to really compare before/after, probably this week. Best regards, Christian
__________________
The future does not exist - we must invent it ! |
November 22nd, 2006, 01:32 AM | #9 | |
Major Player
Join Date: Mar 2006
Location: Helsinki, Finland
Posts: 393
|
Quote:
|
|
November 22nd, 2006, 01:15 PM | #10 |
Regular Crew
Join Date: Dec 2005
Location: Helsinki/Finland
Posts: 131
|
Hi Mikko & co,
I think we both are right. You in practive, and I in theory... ;) After testing this phenomenon more thoroughly I found out that the rolling shutter "picture bending" artefact is there equally strong at all shutter speeds! You are righ that it is more visible at fast shutter speeds since there is less motion blur in the actual image. Nevertheless, it is always there. The CMOS chips that I have tested behaved differently. That mislead me to make to early assumptions. The difference is that in HC1 the actual frame rate is always the same (25 interlaced frames/s) - independent of the shutter speed (as long as it is shorter than the main frame rate). In the CMOS sensors that I tested the frame rate increased in relation with the shutter speed reduction. That caused LESS picture bending with higher shutter speeds since the time to read out one frame also decreased. Since HC1 keeps the frame rate constant - the rolling shutter effect is always there, and more prominent less motion blur you have. Right again Mikko! The only thing that determines its magnitude is how much the picture has moved in a relation to a certain pixel location since the last frame. A fast pan causes the image to slant in either direction, depending of the pan direction. You need a quite fast pan to see this. Fast pans causes such motion blur that scrambles the picture so badly that you do not notice this picture bending very easily. It seems that this "feature" is there to stay in the CMOS sensors, that is just the way how they are constructed. Probably all of us that notice this effect just have to learn how to live with it. It is not a very big issue after all... Regards, Christian
__________________
The future does not exist - we must invent it ! |
November 24th, 2006, 06:18 PM | #11 | |
Major Player
Join Date: Mar 2006
Location: Helsinki, Finland
Posts: 393
|
Quote:
|
|
November 27th, 2006, 12:31 PM | #12 |
Regular Crew
Join Date: Dec 2005
Location: Helsinki/Finland
Posts: 131
|
Hi,
The most probable explanation is that the HC1 CMOS imager chip has also a global electronic shutter. Actually, it MUST have one in the normal video mode (when the rolling shutter is in action), othervise there would not be a possibility to stop incoming photons to intergrate charge in the pixels during very fast shutter speeds. No mechanical shutter could do that. A still picture is imaged by opening the global shutter for a certain period, and then reading the resulting image out - with the global shutter closed. There is a rolling shutter in action during the readout but there is no more light integrating into the pixels. Therefore you cannot see the rolling shutter artefact in stills, just a global blur of the whole image, if your cam moves during the time the global shutter is open. Actually, the rolling shutter is not a shutter at all, it is a rolling reset, where the pixels are cleared after the intergration period during readout - row by row - and the global shutter is the shutter that gates (within the pixels) the accumulation of charge, and is controlled by the shutter time setting. This all makes sense, at least compared to other CMOS imagers I have studied, I do not think that Sony's differs that much. I might have a look at some documents (patents) to find out more. I have never tested it myself - is there a mode where you can shoot stills during recording video? If there is, I think that you might be able to see the rolling shutter artefact during these circumstances since recording video simultaneously requires that the rolling shutter (or reset) is active. Best regards, Christian
__________________
The future does not exist - we must invent it ! |
| ||||||
|
|