Few questions regarding confusing info from a would-be DSMC'er

[Xavo DP]

It is interesting to follow what RED has to offer, even if you are like me not sure whether you will be able to afford starting a new system built around one of the upcoming Scarlet brains. At least it's good to know what top technology can achieve, if one can afford it. While reading about RED intensively these days, I came across some conflicting or hard to understand info, here is some of that:

1) I read that one can use 35mm still lenses on RED-one, but that you should watch out for still lenses labeled as "digital-only" which means they are designed to cover APS-C sensors only and hence produce pronounced vignetting if used on RED-one! Well this sounds wrong to me because the diagonal of the RED-one sensor (28mm) is actually slightly smaller than the diagonal of Nikon DX/APS-C sensor (28.4mm).

2) Some of the specifications about sensors of upcoming RED Scarlet & Epic here:

, some of the info above seems to be unclear and may include some typos or such:
2a) S35mm (Scarlet or Epic) sensor is said to be 30x15mm (2:1 aspect ratio) and in the same time 5120x2700 pixels (1.89:1 aspect), how come?! It is also said to have pixel size 5.4 micron :
W dimension = 5.4 micron/pxl x 5120 pxl = 27648 micron = ~27.6mm (not 30mm)
H dimension = 5.4 micron/pxl x 2700 pxl = 14580 micron = ~14.6mm (not 15mm)
* A side question here is, does this S35 format require larger lenses than the RED-one, the sensor of the former is at least slightly larger than the latter (31.25mm vs 28mm in diagonal respectively)?
2b) Similar to #2a, the sensor of the 617 format (Epic) is said to be 186x56mm (3.3:1) and 28000x9334pxl (3:1), again conflicting aspect ratios between mm and pxl dimensions! I think the 186mm is a typo and it should have been 168mm based on this:
W dimension = 6 micron/pxl x 28000 pxl = 168000 micron = 168mm
H dimension = 6 micron/pxl x 9334 pxl = 56004 micron = ~56mm
2c) 2/3" format (Scarlet) 10.1x5.35mm, 3072x1620pxl, no aspect ratio conflict here, either way you calculate, it's very close to 1.89:1, but the given pixel size of 3.2 micron, can't be exactly 3.2 micron, should be more like 3.3micron (for the above mm & pixel dimensions to be right), or 3156x1664 (for 3.2micron & 10.1x5.35 to be right), or sensor size is 9.8x5.2mm (for 3.2micron & 3072x1620 to be right), this is the most probable as it makes the 2/3" closer to its standard 11mm diagonal.

3) I read that it's not difficult to manufacture mount adapters that allow use of still lenses from other systems (eg Canon or Nikon) on RED camera bodies because the adapters needed are mainly mechanical not requiring much electronic or software development. I also read that such adapters will make the RED camera able to control the lens aperture and AF, IMO this sounds in conflict with not much electronics or software being involved in the adapters' manufacturing.

4) Finally what is going to happen to the RED-one (or at least its sensor size) after the upcoming models become available? Will it continue with the same $17,000+ price (if so it'll look too expensive compared to the Scarlet S35 & FF35 which have larger sensors and cost considerably less, am I missing something here)? OR will it be discontinued (hopefully not, because its sensor size is strategically similar to APS-C presumably allowing use of the latter's lenses)?

EDITs: Funny I am talking about typos and made some of my own, have corrected them above 32.25mm crorrected to 31.25mm, 54mm corrected to 56mm.

 

[Felix K.]

2) Some of the specifications about sensors of upcoming RED Scarlet & Epic here:

, some of the info above seems to be unclear and may include some typos or such:
2a) S35mm (Scarlet or Epic) sensor is said to be 30x15mm (2:1 aspect ratio) and in the same time 5120x2700 pixels (1.89:1 aspect), how come?! It is also said to have pixel size 5.4 micron :
W dimension = 5.4 micron/pxl x 5120 pxl = 27648 micron = ~27.6mm (not 30mm)
H dimension = 5.4 micron/pxl x 2700 pxl = 14580 micron = ~14.6mm (not 15mm)

There are always a few spare pixels. Seldomly the whole sensor is used.

 

[Dave Gray]

1.) Why would you want to use that glass anyway. I think the main problem is that the lenses generally aren't as high quality in the first place and vignette even on the correct sensor size. Could be wrong though.

2) There is also an gap around each photosite where electronics sit.

3) Third parties are already manufacturing Canon mounts to allow for electronic control of lenses. Check out The Birger Mount. Red will be making Nikon/Canon mounts but I don't think they're including the electronics.

4) You'll be able to upgrade the sensor or trade it in and put the full $17k towards an Epic. There is a thread with the options somewhere. If you want to use less of the sensor you'll be able to just shoot 4k (speculation) or a custom size.

 

[Xavo DP]

Thanks for the replies, Felix and Dave. You both suggest that there may be something not accounted for in the specs that explains the sensor size discrepancies I mentioned in #2, but thats unlikely given two things:
1) that doing the same calculations on the sensors of RED-One, FF35mm & 645 fromats will give you exactly the same sensor size you have in the specs (just 1 quick example ff35 sensor is 6000 pixel wide, each pixel 6micron => 6x6000=36000micron=36mm same width as in specs) , if there was something like gap or spare pixels, you'd expect to find consistent discrepancies in all formats.
2) Gaps or spare pixels do not explain the discrepancies in aspect ratio of a sensor when deduced from pixel dimensions vs when from the mm dimensions.

Anyway not terribly important, we have at least an approximate idea of the sensor size of each format. Thanks for reading and replying when you can. I hope I can still get one or two replies on the other questions of the first post.

 

[guiziliu]

it also has an gap around each photosite where electronics sit.

 

[Fritz M.]

1) The APS-C has a 1.5:1 aspect ratio while the Red One (I'm guessing here) is closer to 2.33:1. So they may have the same diagonal, but the dimensions would be different. The Nikon/Canon would be more square while the Red would be more elongated. As such, on the Red, the outer left and right edges would be obscured. [this is mostly guess as I'm not an R1 expert by any sense]

2b) That would appear to be a typo. If you look at the epic/scarlet page on the Red.Com site (http://www.red.com/epic_scarlet/) it lists the size of the 617 sensor as being 168mm wide. This is to match the 6x17 film cameras which put a little bit of a gap (2mm) between pictures on the negatives.

Others have already mentioned that there's also space around each photosite. The pixel size (e.g. 5.4 micron) is given as that is used to determine how much light each pixel is able to collect. I'm not sure how much those kinds of calculations are done with these types of cameras. I do a lot of work with Infra-red cameras and our optics guys and physicists do all kinds of math calculating how many photons would be collected by each pixel when we're evaluating different cameras for the work that we do. And to do that, we would need the actual pixel size.

 

[Xavo DP]

1) The APS-C has a 1.5:1 aspect ratio while the Red One (I'm guessing here) is closer to 2.33:1. So they may have the same diagonal, but the dimensions would be different. The Nikon/Canon would be more square while the Red would be more elongated. As such, on the Red, the outer left and right edges would be obscured. [this is mostly guess as I'm not an R1 expert by any sense]

the lenses in our cameras "do not know" what shape is our sensor, they just project an image circle which has a certain diameter depending on the lens format, within that circle you can draw any sensor shape, be it square, rectangle or even triangle, the only thing you need to be carefule of is that the sensor should not protrude outside the image circle, if a 28mm-in-diagonal square doesn't protrude then a rectangle with an equale diagonal won't protrude either. try drawing different aspect ratio rectangles inside the same circle to see how it goes practically.

2b) That would appear to be a typo. If you look at the epic/scarlet page on the Red.Com site (http://www.red.com/epic_scarlet/) it lists the size of the 617 sensor as being 168mm wide. This is to match the 6x17 film cameras which put a little bit of a gap (2mm) between pictures on the negatives.

thanks a lot for the link.

Others have already mentioned that there's also space around each photosite. The pixel size (e.g. 5.4 micron) is given as that is used to determine how much light each pixel is able to collect. I'm not sure how much those kinds of calculations are done with these types of cameras. I do a lot of work with Infra-red cameras and our optics guys and physicists do all kinds of math calculating how many photons would be collected by each pixel when we're evaluating different cameras for the work that we do. And to do that, we would need the actual pixel size.

as i said in the last post, it is not very important to know what is the exact size of the sensors. but honestly i didn't buy the pixel-gap or spare-pixel theories for a second because my method of calculations gave exactly correct results with many of the other formats included in the same table, how and why would the author of the table give the pixel size as the space between two neighboring pixels for one format and then on the same table for another format give the actual size of the light-collecting photo-diode area instead?! or how could he or she include the peripheral non-active spare pixels in the sensor size of some formats then exclude that in other formats?! that would be a more gruesome mistake more than having typos, it makes the table pointless and misleading for comparisons!!!

 

[Stephen Gentle]

1) The APS-C has a 1.5:1 aspect ratio while the Red One (I'm guessing here) is closer to 2.33:1. So they may have the same diagonal, but the dimensions would be different. The Nikon/Canon would be more square while the Red would be more elongated. As such, on the Red, the outer left and right edges would be obscured. [this is mostly guess as I'm not an R1 expert by any sense]

It's not 2.33:1 - that is the ratio at 4.5K (4480x1920) but at 4K 16:9 the frame is a lot taller (4096x2304). According to the RED site, the ratio of the active pixel array is more like 1.78:1, but that's not that relevant, as Xavo said.