ok,I see. And probably the difference between 4:4:4 and 4:2:2 in 10bit @ 60 frames/s is undetectable for the naked eye. Thanks Steve.
Sony PWW-F3 - $13,300 USD
- Thread starter Matthew Bennett
- Start date
ok,I see. And probably the difference between 4:4:4 and 4:2:2 in 10bit @ 60 frames/s is undetectable for the naked eye. Thanks Steve.
I am sorry it seems the camera won't be DUAL-LINK from release, only in April. The firmware update will enable the Dual-Link (4:4:4 Full RGB and S-Log) and will be released together with the 1TB SxS Card for the new HDCAM SR portable deck.
More details here:
http://www.xdcam-user.com/?page_id=1182
Again, sorry for the wrong info!
More details here:
http://www.xdcam-user.com/?page_id=1182
Again, sorry for the wrong info!
basspig
Veteran
If this were a 3-chip camera and had really good color (as opposed to RED-like color cast under some lighting), I'd be really excited. But I think it's a Beyer filtered single chip, so there's no compelling reason to buy this.
Oh, and 4:4:4 out of a Beyer filtered chip is an oxymoron.
Oh, and 4:4:4 out of a Beyer filtered chip is an oxymoron.
Steve Castle
Well-known member
Sony's approach to 4:4:4 seems to be identical to Arri's.
The Arri Alexa is a 3.5k sensor, the F3 is a 3.43k sensor (based on Phillip Johnstons' interview with Sony). What Arri & Sony are doing is downscaling a 3.4-3.5k sensor into a 1920x1080 output resolution.
Keep in mind, there is a general misunderstanding of the differentiation of a 'pixel' relative to a photosite. So if one group of a bayer pixels equals one output pixel, then yes, it will not be 444. But its not a 1:1 relationship in Alexa or the F3.
So in the case of Arri's Alev III sensor in the Alexa, it has a real resolution of 3392 x 2200 pixels which then downsized to 1920x1080. So by increasing the ratio of photosites relative to output resolution you are increasing the RGB resolution. It won't be perfectly RGB, but you can come incredibly close. Sony's F3 sensor is supposedly 3.43k so its taking a fairly similar route.
For RED, they are taking a similar approach but with a different implementation, by packing in more photosites you can compensate for a lack of RGB resolution in bayer. Red's 4k sensor isn't really 4k, its closer to 3.2k, hence the existence of the Epic 5k. Unlike, Arri and Sony, Red leaves most of the work up to the user to convert that into a final product. This introduces workflow issues but allows more creative freedom; pick your poison.
Sony and Arri seem to expect the final target resolution to be 2k rather then 4k. Hence, a ~3.5k sensor to be the right balance between resolution and the size of the photosites compared with the 5k of the Epic. So if your target resolution is 2k, then you benefit from larger photosites compared with the Epic.
But in short, the Sony's F3 should come close to 444...
I can say something about Sony, that is based in my experience. I've shot features in Sony Z1, Sony EX3, F900, F900 with SR Deck, F23 and F35. None of them failed. Not even 1 time. And the result of all of them was beautiful (considering the limits of each format).
From the F3, I don't expect less.
From the F3, I don't expect less.
basspig
Veteran
Thanks for that clarification, Steve. I was under the (apparently incorrect) assumption that the sensor was 1920x1080, out of which, a Beyer filtered color image would be pretty poor. A 3.5K sensor sounds about right though. Now if they could just improve the color differentiation. Single-chip still has difficulty handling subjects lit by colored lighting, rendering the unnatural color cast. Good color is very important to me.
Nice personal blog basspig.
I now this is offtopic, but I heard somewhere Panasonic was going to release a 35mm CCD camera, for about 20-30K, next year, can anybody confirm that?
