Canon XF100 Tests & Practical Limits

[Philip Lipetz]

I will be slightly delayed in posting new material, took my mother to hospital. First, will be shoot at the low light limits, then audio and new settings tests.

[Geoff Tompkinson]

Hi Philip,

Could you post some real world tests on how good the OIS is on this camera. And how much cmos Jello effect is produced when moving the camera. This is a major problem with dslrs like the 5dmk2 and is supposed to have been much improved in this camera.

I am hoping to use the camera for a lot of pov style shooting from cars and aircraft and need to know if this is going to be a big problem.

Thanks,

Geoff

[Charles W. Hull]

Originally Posted by Geoff Tompkinson

Hi Philip,

Could you post some real world tests on how good the OIS is on this camera. And how much cmos Jello effect is produced when moving the camera. This is a major problem with dslrs like the 5dmk2 and is supposed to have been much improved in this camera.

I am hoping to use the camera for a lot of pov style shooting from cars and aircraft and need to know if this is going to be a big problem.

Thanks,

Geoff

Geoff, I measured rolling shutter with Gunnar Thalin's method and got 31%. This is pretty low for a CMOS camera, obviously much lower than a 5DMk2. I also shoot from aircraft and expect the XF100 will do a good job, but I have not had a chance to try it from the air yet. In just trying some shots with the camera moving around, there is a little jello but it corrects nicely with deshaker set at 31%. My experience is that it should tolerate some mild turbulence with that correction. Chuck

[Stephen Boss]

Thanks Philip, I was waiting for someone to finally put up a test like this. Unfortunately, this cams performance appears to be a bit less than what I had hoped. I still think the infared mode is cool though... if only I had stuff to shoot in infared.

[Erik Norgaard]

Hi Philip:

Nice systematic testing. I noted too when I was on the street significant noice towards the useless with 12dB gain. Possibly usable footage w. 6dB gain. However, on Gran Via, the main street of Madrid, I shot with 0dB gain at night. I guess cars and street lights all add up to a good lighting.

I don't know if this is a fair test, I suppose most cameras will have this issue, anyway, autofocus wobbling:

XF100 Autofocus wobbling

Erik

[Lou Bruno]

Take the IAF off and do a re-test. You should see a difference in focus searching.

[Geoff Tompkinson]

Originally Posted by Charles W. Hull

Geoff, I measured rolling shutter with Gunnar Thalin's method and got 31%. This is pretty low for a CMOS camera, obviously much lower than a 5DMk2. I also shoot from aircraft and expect the XF100 will do a good job, but I have not had a chance to try it from the air yet. In just trying some shots with the camera moving around, there is a little jello but it corrects nicely with deshaker set at 31%. My experience is that it should tolerate some mild turbulence with that correction. Chuck

Thanks for getting back to me on this Charles. I work in after effects and am not familiar with Gunnar Thalin's method. As a comparison what does the 5Dmk2 rate on this scale?

Could you post some footage from the XF100 with the camera moving around so I can try to stabilize with Prodad Mercali in AE ?

It seems almost impossible at the moment to get any sample footage from this camera.

Thanks

Geoff

[Charles W. Hull]

Originally Posted by Geoff Tompkinson

Thanks for getting back to me on this Charles. I work in after effects and am not familiar with Gunnar Thalin's method. As a comparison what does the 5Dmk2 rate on this scale?

Could you post some footage from the XF100 with the camera moving around so I can try to stabilize with Prodad Mercali in AE ?

Geoff, there is a very capable rolling shutter plug-in for After Effects, it is Rollingshutter, see www.thefoundry.co.UK/products/Rollingshutter . It sells for around £300. Prodad Mercali cannot handle rolling shutter issues.

The 5Dmk2 has a rolling shutter of around 70-75% on the scale I mentioned, so it is much worse than the XF100. Whenever you have any pointing shake with the 5D you are in trouble. There is a program that runs on virtualdub that does a truly amazing job with this problem, it's called Deshaker, and Thalin is the author. Check his write up and instructions at www.guthspot.se/video/Deshaker.htm. Both virtualdub and Deshaker are freeware.

Deshaker has the other unique characteristic that it can stabilize without changing the video size - there is no magnification or loss of resolution. This is done by borrowing from the previous and following frames. So it is a fascinating program for pov shooters.

I'm traveling - when I get back I'll try to post an example. (edit -those links were hand entered on an iPad so you may need to Google to find Rollingshutter and Deshaker)

[Erik Norgaard]

Originally Posted by Erik Norgaard

I don't know if this is a fair test, I suppose most cameras will have this issue, anyway, autofocus wobbling:

XF100 Autofocus wobbling

Originally Posted by Lou Bruno

Take the IAF off and do a re-test. You should see a difference in focus searching.

I've done some more testing on this and it is most definitely a low light issue. It happens with all three auto-focus speeds, I.AF, M.AF and AF. It's not related to aperture and it's not related to gain.

I did a test of all AF speeds with F 4.8, F4.0 and F3.2. When the wobbling effect happened it happened for all AF speeds. In this case, at F3.2 it disappeared. So to test that it was not related to aperture I used an ND4 filter and the wobbling came back. I also tried different gain settings, same thing.

Now, I don't find the video I posted particularly low lit, but the lesson is that if you want to use AF you need lots of light.

BR, Erik

[Lou Bruno]

This issue has been around for some time as it relates to auto-focus and lowlit situations. This has occurred in my XL-2 and both of my XH HD cameras. I just tested the XF-300 and YES...it is a focus hunting issue. As I mentioned, when the IAF is set to OFF, there is somewhat less hunting. I have also observed this in some earlier JVC SD cameras, namely the 5100 camera. I will test the XF-100 as well.

[Erik Norgaard]

Originally Posted by Erik Norgaard

Now, I don't find the video I posted particularly low lit, but the lesson is that if you want to use AF you need lots of light.

I'll admit though that the oranges lack clear contrast lines, though the green/orange should do - maybe the AF works on a luminance/BW, and the green/orange does not produce enough contrast in BW?? So, maybe this issue appears in any low contrast situation, be it high key or low key. I guess I need to do more testing.

BR, Erik

[Philip Lipetz]

Today the. 6dB low light video goes up. One reason that it took so long was I used AF and there was subtle focus shifting at low light, rounds shots. Really subtle, not easily visible in small edit window but clear in full screen. Seems that it happens most with horizontal pans, not vertical. Was in instant AF mode. Will have to redo with instant off.

[Deborah Gallegos]

Originally Posted by Philip Lipetz

... I used AF and there was subtle focus shifting at low light, rounds shots. Really subtle, not easily visible in small edit window but clear in full screen.
...
Was in instant AF mode. Will have to redo with instant off.

Saw your latest Vimeo test, Philip...great stuff!

Sooo... do you recommend AF at all in low light or does it have to be MF? I understand you all are saying IAF does NOT do well in low light.

Thanks!
Deborah