|
|||||||||
|
Thread Tools | Search this Thread |
January 30th, 2011, 09:45 AM | #1 |
New Boot
Join Date: Dec 2010
Location: Santiago, Chile
Posts: 6
|
F3: 3.36 megapixel 'effective' resolution. Enough for free-aliasing 1000TV lines 444?
A Sony official brochure about the coming PMW F3 has just been released.
http://www.matrixvideo.ca/pdf_folder...3-BROCHURE.pdf I was surpriced when I read that the F3 sensor has an 'effective' resolution of 3.36 megapixels. I wonder what is the meaning of effective resolution...I am assuming that it refers to the number of photosites on the sensor. I remember people speculating that the sensor should be around 3.5K. This horizontal photosites count resolution should translates in a total count near 7 megapixel (3.5x2) for the total sensor. But the number of 3.36 megapixel in the brochure is way below that. What's going on here? I understand that the strategy, when you are delivering HD output and you are using a single Bayer sensor, is oversampling the horizontal photosite resolution to more than the desired output of 1080 (2K). This to prevent aliasing and avoiding chroma subsampling. It seems that a good sweet spot to achieve this is near 4K (the number is 3.5K in the case of the Alexa I believe). By having 4K horizontal photosites you can assign 2K photosites to green, 1K to the blue, and 1K to the red. This will allow you to produce an image with a resolution of about 1K TV line pairs horizontal (TV/ph) on the green channel without aliasing (since you need at least two pixels to capture 1 TV line pair, which is one cycle, we get 1/2*(2K green channel)=1K TV/ph). Also, this horizontal oversampling near the number 4K leaves 1K for each chroma (B and R), minimizing the potential aliasing on those channels too. So if the F3 sensor has only 3.36 megapixels instead of a number like, say 7 megapixels (which is what I was expecting) how can it delivers an image of 1000 TV/ph without aliasing (and without chroma subsampling since we know the sensor can delivers uncompressed 4:4:4)? I would expect that a Bayer sensor that resolves around 1000 TVL/ph (which is about the limit for a 1920x1080 camcorder) should contain something near 3.5K horizontal resolution, which is about 7 megapixel photosites. But 3.36 megapixels is half that amount. Maybe the mismatch comes from the fact that I am equating effective resolution with sensor resolution, when they may refer to different things. Could someone explain to me what 'effective resolution' means? |
January 30th, 2011, 06:13 PM | #2 |
Major Player
Join Date: Jun 2008
Location: melb.vic.au
Posts: 447
|
When I read the the term effective pixel, I take it to mean the final result after processing. Effective being what it's equal too after debayering, and pixel is a pixel, a dot with full luminance and colour information.
I doubt they mean only 3.36 meg photosites.
__________________
www.davidwilliams.com.au |
January 30th, 2011, 07:12 PM | #3 |
Inner Circle
Join Date: Jan 2006
Posts: 2,699
|
No, the chip resolution of the Alexa for recording (it uses more for viewfinding ONLY) is 2880 x 1620, not 3.5K horizontally. (See Camera Specs | ARRI Digital ) Do the maths and that works out to be about 4.6 megapixels.
I wouldn't go too much by pure numbers, it's how they get used that's more important, but something in the region of 3.5 to 4.5 megapixels sounds about optimum. Too few and it impacts on resolution, too many and it makes the processing more complicated, and (if done properly) ups power requirements, heat etc. It's true it won't give "true" 4:4;4 (equal R,G,B resolution), but 4:4;4 recording will make the most of what it is giving. |
February 4th, 2011, 01:11 PM | #4 |
Inner Circle
Join Date: Dec 2005
Location: Bracknell, Berkshire, UK
Posts: 4,957
|
I wonder if they have used a 45 degree twisted bayer array as on their clear-vid sensors. It makes interpolation more accurate.
__________________
Alister Chapman, Film-Maker/Stormchaser http://www.xdcam-user.com/alisters-blog/ My XDCAM site and blog. http://www.hurricane-rig.com |
February 4th, 2011, 04:49 PM | #5 |
Major Player
Join Date: Jan 2006
Location: Victoria Australia
Posts: 246
|
Maybe the sensor is not a straight bayer? This blog seems to indicate the sensor may be something we haven't seen before.
Notes On Video: Sony F35 : Another unusual sensor Edit: Oops my bad. It's the F35. Why do they make all these names sound the same! |
February 4th, 2011, 06:01 PM | #6 |
Major Player
Join Date: Jun 2008
Location: melb.vic.au
Posts: 447
|
At the Melbourne F3 demo I asked about the sensor pattern. Bayer was the response. The guy was technically knowledgeable, so I have no reason to doubt him. Not the Q67 I asked? No, Bayer.
__________________
www.davidwilliams.com.au |
February 8th, 2011, 12:24 PM | #7 |
New Boot
Join Date: Dec 2010
Location: Santiago, Chile
Posts: 6
|
resolution info from the original brochure was deleted
Interestingly, the two lines about sensor 'effective' resolution, and camera measured resolution have been deleted from the last version of the brochure.
These two lines don't appear any more: Effective Sensor Resolution (HxV): Approx. 3,36 Mega Pixels Horizontal Resolution (1920x1080/59.94i) 1000 TV lines or more link to the old version of the brochure: http://www.cameradepartment.tv/files...ads/PMW-F3.pdf link to the more recent version of the brochure: http://www.matrixvideo.ca/pdf_folder...3-BROCHURE.pdf |
February 8th, 2011, 01:28 PM | #8 |
Inner Circle
Join Date: Dec 2005
Location: Bracknell, Berkshire, UK
Posts: 4,957
|
They have also removed the reference to Hypergammas from the old brochure and replaced them with Cinegammas in the new on. It's been common knowledge that the F3 would have Cinegammas as standard for some time.
__________________
Alister Chapman, Film-Maker/Stormchaser http://www.xdcam-user.com/alisters-blog/ My XDCAM site and blog. http://www.hurricane-rig.com |
| ||||||
|
|