AF100 / AF101 - initial report

Barry_Green said:
Nothing has changed. Three chips still deliver superior color to a single-chip system. The reason everyone's going with a single chip is NOT increased color fidelity; a Bayer chip is inferior to three chips for color fidelity, and that's why every broadcast camera on the market has three chips and a prism.
Click to expand...

I don't see any scientific results that prove that bayer chip is inferior to 3 chip system in general.
All common bayer sensor use quite a big oversampling, so all this talks about color fidelity are questionable.
2/3" 3 chip use in broadcast camera came not from color fidelity, but rather simple sensitivity and DR issue.
Having 3 chips without strong color filters and prism is better then even have three times area bayer sensor .
This is as simple as that.
 
Optically splitting the light with an optical prism should result in more accurate pure light than using a colored gel over the pixels. Secondly, the problem with Bayer and color fidelity is that Bayer devotes 50% to green, 25% to blue, and 25% to red. But that's not how video systems work. Video systems generate color by using approximately 60% green, 29% blue, and 11% red. Yes, massive oversampling does help minimize this issue, but it's certainly not an optimal use of chip space. The three-chip system provides huge overkill on red and blue, but by separating them out the way they do, they provide for the optimal combination of mixing the separate chips to create the color signal.
 
What's the native resolution of AF101s senzor? I guess 1,920×1,080? That would mean it has better low-light performance than most of HDSLRs (maybe even better sensitivity than Canon 1D and Nikon D3s)? Am I right?
 
Barry_Green said:
Optically splitting the light with an optical prism should result in more accurate pure light than using a colored gel over the pixels. Secondly, the problem with Bayer and color fidelity is that Bayer devotes 50% to green, 25% to blue, and 25% to red. But that's not how video systems work. Video systems generate color by using approximately 60% green, 29% blue, and 11% red. Yes, massive oversampling does help minimize this issue, but it's certainly not an optimal use of chip space. The three-chip system provides huge overkill on red and blue, but by separating them out the way they do, they provide for the optimal combination of mixing the separate chips to create the color signal.
Click to expand...

Hi Barry
If i'm not mistake it's 29% red and 11% blue

Amos
 
hellboy80 said:
What's the native resolution of AF101s senzor? I guess 1,920×1,080? That would mean it has better low-light performance than most of HDSLRs (maybe even better sensitivity than Canon 1D and Nikon D3s)? Am I right?
Click to expand...


if panasonic really wanted us to bow to them then they might do that, more likely though it will be something like a gh2 sensor modified for video. that being said, if they did go for 1080 res it would still be a single chip bayer pattern, meaning to get 1080 it will likely be a 6-8 megapixel sensor if my debayer math is correct(haven't finished my coffee yet, so could be off by a few megapixels)

on the 3 vs 1 chip debate, what's really the point of debating that? if the color pleases the viewer, then there. some of the highest end cameras are single chip striped pattern, but in our world i think it's more pertinent what just looks good to the eye. i'm sure the color information on the af100 will be better handled than the DSLRs
 
We're not sure what Panasonic has done with the sensor. They could be using a Bayer pattern, they could be using a modified bayer pattern. There has been some assumption that it is the same sensor as the GH2, but I've seen nothing official. The one thing we do know is that a proper OLPF is being used for 1080p resolution.

I wouldn't be surprised if they used a modified bayer pattern where they double up on adjacent pixels with the same color filter, then, through binning, they could possibly use the fact they are the same to help reject noise. Also, not sure what kind of microlensing they are using as well. There are A LOT of variables and different approaches they could make.
 
I had a disastrous experience with Red's Global Shutter flying on a Tyler Nose Mount and had to re-shoot with my trusty VariCam; Tyler reports that the Red on their nose mount is very hit or miss, definitely a gamble.

Jan, is it possible for Panasonic to do a nose mount test when the camera is ready for release? Helicopter shooting is so expensive, I don't want to risk it. If the AF100's Global Shutter can handle the helicopter, it will be quite an achievement.

Good shooting and best regards,

Leo
 
Ticheli said:
If the AF100's Global Shutter can handle the helicopter, it will be quite an achievement.

Leo
Click to expand...

My understanding is that the AF100 has a rolling shutter just like every other cmos camera. So it's probably hit or miss (miss most likely) just like the RED...
 
Red has a rolling shutter as well. Expect the performance to be the same or worse depending on the speed of the AF100 read/reset times as compared to Reds. The M sensor had an 8.9ms read/reset. I know the Epic is 5ms, but Jim stated to me that they were not running it at 5ms on the MX sensor at the time I talked with him, it's possibly they could have sped it up more than 8.9ms.
 
Can't believe I wrote Global Shutter instead of Rolling Shutter! Yes, of course it's a Rolling Shutter problem; I would like to see Panasonic test it on a nose mount.

Good shooting and best regards,

Leo
 
And... one more test I'd like to see is how the AF100 handles a flash in shot; with Red the Rolling Shutter will give you a half exposed frame, so no shooting the "movie star on the red carpet with paparazzi' scenes.

Good shooting and best regards,

Leo
 
yea any rolling shutter camera is going to give you partial frames in flashes. unless they get it down to say under 2ms. but i don't really see it as a huge issue, many people don't notice the partial frames and if you really wanted you can throw a white frame over flashes and be done with it. global shutter cmos is at least a few years away, and manufacturers don't seem to like ccd much anymore, so these are what we have to work with
 
Ticheli said:
And... one more test I'd like to see is how the AF100 handles a flash in shot; with Red the Rolling Shutter will give you a half exposed frame, so no shooting the "movie star on the red carpet with paparazzi' scenes.

Good shooting and best regards,

Leo
Click to expand...

Definitely expect something similar...I doubt the AF-100 has their read-reset time down under 2-3ms.
 
Hi everyone,

here is the Panasonic AG-AF100 first release test (September 2010): http://vimeo.com/15107117
Finally, Panasonic guys gave me the permission to share it.
I shooted this video, based on a short novel written by Paolo Rossini (the actor of this short), on the Panasonic AG-AF 100 in half a day on September 2010 and edited during the night (a very special thanks to Liam Lennon for the precious help) for the Italian official presentation (the morning after). The camera was a pre-release model so was impossible to have an accurate review (menu was partially locked, the sensor was not the definitive one and no recording, I used my great Nanoflash via SDI camera output).
Next week (October 19 to October 21), I will try the new Pany again, so STAY TUNED!

P.S.
The beautiful girl in the photos is Anna Mazza, a very special thanks to her!
 
Looks like you either graded or used some diffusion filters?

also looking at :57 the rolling shutter looks very improved over any DSLR
 
You must log in or register to reply here.
Back
Top