pannyfreak
Active member
A lot of comments on sample footage seen around the net seem to praise the footage immensely in terms of sharpness, color richness, but one critique crops up repeatedly: Image noise. Particularly in the blacks.
BUT - Are we forgetting nearly all footage on the web from this camera is coming from two websites: Vimeo or YouTube. YouTube, if you right-click on a video during playback and select "Show Video Info" reports 1080p streaming through about 6Mbps peak. And that's using H.264 compression. Vimeo, appears to use around 5Mbps (based on their upload compression recommendation page).
So, could it very well be a lot of the noise in the blacks we're seeing is from watching 24Mbps H.264 recompressed for online distribution into another H.264 6Mbps stream? Shadow detail and dark areas will be the first to be compressed heavily when bitrates lower-- especially when throwing away 75% of the image data!
I propose the definitive test:
When someone gets their AF100 (or even has one now), put the body cap on with no lens. No ND filter engaged. Set ISO to 100, 200, 320, 400, 800, 1600, 3200 and record 5 seconds of each at 1/48 (180 degree shutter). Pull a TIFF from each setting and post as a zip file. Then we'll see inherent noise.
P.S. - Also post the video, something like this: http://vimeo.com/10473734
I haven't seen anyone do anything like this yet with the AF100 and I think would be a great thing to see.
BUT - Are we forgetting nearly all footage on the web from this camera is coming from two websites: Vimeo or YouTube. YouTube, if you right-click on a video during playback and select "Show Video Info" reports 1080p streaming through about 6Mbps peak. And that's using H.264 compression. Vimeo, appears to use around 5Mbps (based on their upload compression recommendation page).
So, could it very well be a lot of the noise in the blacks we're seeing is from watching 24Mbps H.264 recompressed for online distribution into another H.264 6Mbps stream? Shadow detail and dark areas will be the first to be compressed heavily when bitrates lower-- especially when throwing away 75% of the image data!
I propose the definitive test:
When someone gets their AF100 (or even has one now), put the body cap on with no lens. No ND filter engaged. Set ISO to 100, 200, 320, 400, 800, 1600, 3200 and record 5 seconds of each at 1/48 (180 degree shutter). Pull a TIFF from each setting and post as a zip file. Then we'll see inherent noise.
P.S. - Also post the video, something like this: http://vimeo.com/10473734
I haven't seen anyone do anything like this yet with the AF100 and I think would be a great thing to see.