Hello all,
One of the most frequent questions I’m emailed about is sensor/film format sizes and 35mm depth of field. Actually what seems to happen more often than not, is that I’m informed that I should build a PL adapter for the Canon 5Dmk2, because it has “35mm depth of field, and the GH1 doesn’t.” This is incorrect. I’m going to repeat this….
THIS IS INCORRECT.
The Canon 5Dmk2 uses a 135 sized sensor which is a 35mm STILL-photo format, the closest equivalent on the motion picture side is VistaVision. This is not the format used predominantly for mainstream motion picture and television production.
The Panasonic Lumix DMC-GH1 sensor, while smaller than the Canon 5Dmk2, is much closer to the field of view/depth of field used in mainstream film and television. Take a look at this chart I created. All sensors/film formats sizes listed are to scale.
One more thing- the 4/3″ format is not roughly equivalent to 16mm. One could say that 2/3″ HD and 16mm film are similar, but it’s not accurate to also include 4/3. 4/3 is FOUR TIMES the size of 2/3″ HD. (22mm hypotenuse vs 11mm hypotenuse). See the chart below for a visual representation.
Film Format and Sensor Size Comparison Chart
I am purchasing a 5D Mark ii – it was my understanding red was a much larger resolution than Red at 4k – is the mark ii better rez?
Thanks and congrats on your rigs!
Cheers. Ben
Excellent chart, Illya, I spotted a similar one with some measurements given in mm on Wikipedia. The 16mm frame you are showing is for Super 16, correct? As regular 16mm is sort of ancient history by now, it having only a 1.37:1 AR. That is a rather weird format size, actually, as is S16 with its non-standard 1.66:1 aspect ratio. The newer Ultra 16 actually makes more sense to me for film use, as Ultra 16 records the image in the 1.85:1 widescreen AR.
I was wondering, when one shotos a video clip with the Canon’s monster sized sensor, is the entire sensor area being used to capture light, and then the iamge is internally downconverted to 1920×1080? Or when there is video capture, only the center section of the sensor measuring only 1920×1080 is even used?
Hi Ben,
I’m sorry I don’t exactly follow your question. If you are asking if the Red cameras has more resolution than the Canon 5Dmk2 is motion images- the answer is yes. If you mean actual resolution on the sensor, the answer is the 5Dmk2, but the Canon doesn’t use all that resolution to full advantage in the video mode. There’s a test going on next week between the Red, 5Dmk2 and GH1. I’ll post more information about this later.
I.
Hi Frank,
Super-16 is 2.26mm wider than regular 16mm. I went back and forth with which 16mm format to included I finally decided that I’ve heard too many people over the years say that 2/3″ HD is the same as Super-16. With this chart I can show that it’s not only smaller than Super 16, it’s actually smaller than regular 16mm. I didn’t want to include both 16mm formats because it just got too cluttered. Ultra 16 is a cool format, but unfortunately it never really caught on, there’s really almost no one set up for Ultra 16, be it cameras or telecines out there. If you compare regular 16mm to Ultra 16, there’s a lot more people shooting regular 16mm.
I’ve been told (and have not confirmed myself) that the Canon uses nearly all of the sensor for light capture (actually a full width 16:9 slice out of the 24×36mm sensor), but it’s not scaling 22MP to 1920. Entire rows of pixels are being dropped and combined before converting to 1920×1080 30p. I’m told this is the primary reason why some 5D images have a problem with aliasing.
Thanks for the response – thats a bit sad that the Mark ii does not utilize its full sensor in motion photography
Would love to see the test footage of the Red, 5Dmk2 and GH1 comparison when it comes out!
Thanks very much,
Ben