View Full Version : Optical image stabilisation not 100%??


Gabriel_Knight
November 22nd, 2002, 12:54 AM
Hi everyone,

Today a shopkeeper told me that the optical image stabilisation of the GL2 is nice, but a normal not optical image stabilisation is far more stabel.

Normal camcorders uses image stabilisation in both vertical and horizontal directions.
The optical image stabilisation, gives a much better image, but it only uses vertical image stabilisation.

So the image on an ordinary cam is far more stable (he says)

I found 2 websites with information.

http://www.canon.com/technology/optics/vap/index.html
http://www.canon.com/technology/optics/shakecorrect_shift/index.html

The main question is: Wich stablisation is the GL2 using (VAP or ISS). And do you have problems to get a stable image

Gabriel

Peter Moore
November 22nd, 2002, 08:05 AM
The GL2's stabilization to me works best when you're zoomed in fairly close. You don't notice huge jitters that you would without it. A different DV camera that uses digital stabilization is not going to have as good of a picture, as you said.

If it's true the GL2 only uses vertical that makes sense - gravity is going to cause you to be more unstable vertically than horizontally.

Jeff Donald
November 22nd, 2002, 08:13 AM
I think your shopkeeper is misinformed. The GL2 uses VAP to the best of my knowledge. Canon has the majority of the patents on Optical Image Stabilization. Sony licenses the technology from Canon to use OIS in their higher end cameras.

I have Canon OIS on my Pro Canon 35mm lenses and they stabilize in both directions. They also have a switch to disable the horizontal correction so that I can pan with a subject. Video may require a lot of horizontal pans and the inclusion of it would add needlessly to the price. Why? Most videographers would have it disabled so as not to interfere.

The only other common IS is Electronic Image Stabilization. It is typically used in lower end cameras and it reduces picture quality by lowering resolution.

Jeff

Andre De Clercq
November 22nd, 2002, 08:41 AM
Gabriel, I am not familiar with this camera. Alllmost for shure it will be a variable prism system. If you want to verify this just look into your lens while switching on yr camera. You will see the prism settling on its ref position. With the other three concepts lens shift, CCD shift (patented but not commercialised as far as I know) and electronic stabilisation, you need another "trick"to verify the concept: try to shut off a corner of the picture by putting a tape on the lenshood so that you get vignetting in that corner. Slightly shake yr can and if the area changes it's an optical one, if it doesn't change, or OIS is OFF or it's an electronic one. Like Jeff wrote the salesman misinformed you about OIS and shake compensation.

Gabriel_Knight
November 22nd, 2002, 09:33 AM
Because I also don't have the cam yet, I would ask if someone can verify this trick, and also give me more information about the image stabilistation that has been used in the camcorder.

Gabriel

Jeff Donald
November 22nd, 2002, 10:21 AM
Dre... is correct and the article you referenced on Canon site is about as good as it gets. OIS works great and Canon's implementation of it (since their the originator) is the best in my opinion. I have used Canon's OIS in video, 35mm film and I have it in a pair of binoculars. It works as described.

Jeff

Chris Hurd
November 22nd, 2002, 02:13 PM
Just so there's no further confusion, the OIS on the GL2 does indeed employ a VAP (vari-angle prism), as do the XL1S, XL1 and GL1.

Refer to http://www.canondv.com/gl2/f_image_stabilization.html -- a graphic there illustrates how the VAP works in the GL2. Hope this helps,

Andre De Clercq
November 24th, 2002, 10:48 AM
Chris, FYI
Image stabilisation involves a lot of trade-offs. One of them is the fact that optical stabilizers get their data from gyroscopic sensors . Such sensors are dificult to use for "slow" shakes. Therefore the advanced OIS systems use additional image analisys based info, to correct for these low shake frequencies. The analisys methods is the same as these used for electronic stabilisation (autocorrelation), but the CCD readout position is (almost) not changed, because the info goes straight into the VAP actuators. The result is better stabilisation. The XL1S apparently uses such a supper bandwidth stabilizer..

Chris Hurd
November 24th, 2002, 03:54 PM
Perhaps we should have made an effort earlier to point out these differences between the OIS system in the XL1 & XL1S as compared to the GL1 & GL2 (the GL2 utilizes the same lens and OIS as the GL1).

There are two major trade-offs to watch out for when using OIS. First, it's important to switch OIS "off" when shooting from a tripod. Otherwise the OIS will try to fight your intended pans and tilts, resulting in choppy movements. Second, beware of the "dancing fireflies" effect recorded with the image when shooting around bright lights at night. These are tiny reflections from the VAP. Switch the OIS "off" when shooting around bright lights at night in order to avoid this.

Gabriel_Knight
November 24th, 2002, 04:08 PM
Correct me if I am wrong, but after reading this thread I come to one conclusion for the GL2:

Electronic stabilisation gives a more stable image.
Optical stabilisation gives a better image, but not as stable as electronic stabilisation.

Chris Hurd
November 24th, 2002, 06:31 PM
I would have to disagree. Optical is in most respects far preferable to Electronic. Electronic (EIS) will noticeably degrade the quality of the video. Optical (OIS) is a mechanical process which doesn't directly affect image quality. You'll want Optical over Electronic. Hope this helps,

Don Palomaki
November 24th, 2002, 06:57 PM
With EIS you still will get the blur that goes along with the shake.

And because EIS is readingt pixels, it can be fooled.

Andre De Clercq
November 25th, 2002, 05:42 AM
You are right Don, but there are a couple of things which you maybe didn't considder. First of all, whether or not stabilisation is involved, once the image moves over the CCD structure there is blurr. If OIS can't compensate shakes having frequency components below 20Hz (for different reasons) , you get the image motion on the CCD anyway, The slower the image motion on the CCD (pixels/field), the less blurr occurs. So OIS still results in a small amount of blurr for slow shakes. By just freezing in playback a slowly shaked OIS camera picture vs one from a stable camera (resolution test pattern) you will see the reduced resolution which is the result of blurr. The extra EIS is not ment to compensate for this slow shakes' residual blurr, but for extracting image information for VAP based correction, which finally gives an extra amount of picture stabilisation where the OIS can no longer garantee stabilization.