Cmos sensor of the AF100

Katty Galloo

Katty Galloo

Active member
According to the specifications of the AF 100 cmos sensor 12.2 megapixel he captures his raw footage at 4K
so its capaple of producing 4.032x 3024 Raw images who are been downconverted , that's verry close to the 'R' sensor 12.6 megapixel who produces a 4.900x 2.580 but of course uses another compression codec who compresses the image and blows it back up to its original size. Kinda like a zip. file compression.

An't that good news to know that the AF100 actualy shoots 4k
 
Last edited:
That's wright Taylor, but it was rather about some personal statement I made before about a specific ... , now I will behave myself and just make a desent comparisation. There is nothing wrong with that isn't it? or am I wrong, again, hmm...
 
No, the AF100 does not shoot 4k. I own one, and the resolution it resolves is close to 800 lines. The sensor might be that dense,, but the way they are being used is obviously different than a full raster 12MP bayer from the results. Panasonic has not released how they are down-sampling the sensor.
 
Hey grimepoch , I did not say he shoots 4k , I just say the sensor would be able of shooting 4k. but they compress the image to mpeg.
Besides the compression H.264 isnt so bad at all. Read more about raw wavelet compression and find out it isnt that perfect at all, but rather a commercial advantage.
 
Last edited:
To the footage I see and the low light ISO possibily settings this cmos is verry simular to the Mark II that means they use just like Mark II the full power of their sensor
but the downscaling resolves in 800 lines. a 12 Mega pixel sensor uses 12.000.000 million pixelsensors to capture his image, why would they ban fore example 6.000.000 million of it.
Canon doesn't either.
 
969638-cool_story__bro_super.jpg
 
Katty Galloo said:
Hey grimepoch , I did not say he shoots 4k , I just say the sensor would be able of shooting 4k. but they compress the image to mpeg
Click to expand...

Hey Katty, The problem is that you are counting pixels and not the actual performance of the camera. We use a low pass optical filter(grimepoch this one is for you) to bring the 12.2M pixels into a 800 lines of resolution so that it can be recorded as an HD signal without the moire and aliasing that the DSLRs have. There is no means of gaining access to the higher pixel density.

Secondly you should be very clear in your reference of MPEG, as there is a world of difference between MPEG2 and MPEG4. The AF100 records to MPEG4. Take a read under the articles section for the AVCHD vs XDCAM EX, you may begin to understand the differences.

And lastly, be happy for what the AF100 is, and worry not about what it isn't. It isn't a RED.

Best regards,

Jan
 
Hi jan,

Thank you for your technical explenation and your wright I should have mentioned mpeg4
wich is a mutch higer quality. I don't want the AF100 even to be a Red. I want it just the way it is. And am so happy with is release. It's a wonderfull cam. But I am a curious person to technical stuff and learning. But I have a question you might be able to give me an answer. Can you tell me the huge diverence between DVCPRO100 and mpeg4 please? Why the choise of Panasonic to go for mpeg4. Just out of curiousity , I still have some P2 cards here from the HVX can I still use them on the AF100 with an external device and have 4.2.2 ?

Explenation about the low pass filter;
http://www.newton.dep.anl.gov/askasci/eng99/eng99312.htm
 
Last edited:
Jan_Crittenden said:
Hey Katty, The problem is that you are counting pixels and not the actual performance of the camera. We use a low pass optical filter(grimepoch this one is for you) to bring the 12.2M pixels into a 800 lines of resolution so that it can be recorded as an HD signal without the moire and aliasing that the DSLRs have. There is no means of gaining access to the higher pixel density.

Secondly you should be very clear in your reference of MPEG, as there is a world of difference between MPEG2 and MPEG4. The AF100 records to MPEG4. Take a read under the articles section for the AVCHD vs XDCAM EX, you may begin to understand the differences.

And lastly, be happy for what the AF100 is, and worry not about what it isn't. It isn't a RED.

Best regards,

Jan
Click to expand...

Are any pixels skipped on the way, or is it downsampling from the full 17.8x10mm and how many pixels fit in there ( all 12.2Mpix?)...
Why wasn't the GH2 teleconverter feature implemented in the Af100?
TIA
 
It is obvious english is not her native language, be nice, not all of us may have great language skills.
 
nugat said:
Are any pixels skipped on the way, or is it downsampling from the full 17.8x10mm and how many pixels fit in there ( all 12.2Mpix?)...
Why wasn't the GH2 teleconverter feature implemented in the Af100?
TIA
Click to expand...

No pixels are skipped.

It isn't the same camera and the deployment of the imager is not the same.
Thanks,

Jan
 
Last edited:
What happens to those original pixels I cannot say. What we do know is that there is a plethora of methods than can be used to down sample and a lot of that depends on the color filters being used, circuitry that could bin different pixels, etc. I don't know how much detail say a GH1 or a GH2 resolves, but the AF100 does resolve more detail than my 7D, and that is without question. I have both and can compare the results.

Regardless though, the OLPF (which probably uses birefringent material) to essentially blur the image down to reduce the spacial frequency of incoming light to be less than the resolution capability of the sensor.

http://en.wikipedia.org/wiki/Birefringence

The problem with saying "Sensor X is Y megapixel" is there is SO MUCH more information required to know what is happening with the camera and how much real resolution it has. A 4 pixel camera could be blown up to 4K resolution, but clearly, it wont resolve that in tests. The important thing to know when comparing camera resolution is how much it resolves, which people test with a test chart. 800 lines is good, I am MORE than happy with that. Sure, who wouldn't want as much detail as possible, but that's not what $5k is going to get you. There is a reason the Red One costs $17.5k for the brain alone. I think it's an amazing tool, and have worked on footage that is amazing from it. But, for $5k, the AF100 is an amazing tool, and so far, I'm thrilled with the results.

Sure, you are right that codec and bit depth is important as well after the sensor, the thing is, people try and focus on ONE parameter of a camera, say resolution, but these are multi-variant systems. Resolution is ONE issue. When you are shooting, you care about them all.
 
Jan_Crittenden said:
No pixels are skipped.

It isn't the same camera and the depolyment of the imager is not the same.
Thanks,

Jan
Click to expand...

Can you tell engineering, good job on the read/reset. I measure about 13.4ms and that is pretty damn good! About the same speed as the Red One in independent tests.
 
Rick Burnett said:
Can you tell engineering, good job on the read/reset. I measure about 13.4ms and that is pretty damn good! About the same speed as the Red One in independent tests.
Click to expand...

Hi, I don't wana make a comparisation with RED and not beeing a pain in the but either. It was just my goal to learn something about cmos by easing some tension here. It results in some intresting discussions. So if I get it wright now the FA100 uses full capacity of the sensor to capture as a manner of speaking, and is then been downscaled by the low pass filter and compressed after to mpeg4.
Well then we can say we have a super great image as result !! and it makes me happy :-) Its a kick ass camera :-)
 
However, the low pass filter is optical, not electronic, so it happens before the photons even reach the sensor to be captured. Panasonic is not telling so far how they are capturing the data, or how they are down-sampling. All we know is they are actually using MOS and not CMOS technology (which is different) and that they say they are not skipping lines. Sure, given they know the cutoff of the OLPF, they probably tuned the sensor workings to deal with that, but regardless, you can avoid mpeg4 completely by taking the signal out of the HDSDI port. What's important though, is it is NOT a 4k camera. Like I said, the only way to really judge the resolution of a camera is not by the pixels it outputs, but by the lines it resolves in a proper test, which I find really good.
 
You must log in or register to reply here.
Back
Top