Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Scarlet VS Ex3 w/HD-SDI out

Subhadip Sen said:
Without even questioning the validity of your original assumption, your 3.5x value is clearly erroneous.
S35 - 5.4µ
2/3" - 3.3µ
Click to expand...
Uups,
I wrongly calculated with Monstros 6µm pixels...

So correction:
5.4x5.4=29.16
3.3x.3.3=10.89
29.16/10.89=2.68

Conclusion: if sensors are same gen S35 should be 2 2/3 stops faster than 2/3"
 
toke lahti said:
Uups,
I wrongly calculated with Monstros 6µm pixels...

So correction:
5.4x5.4=29.16
3.3x.3.3=10.89
29.16/10.89=2.68

Conclusion: if sensors are same gen S35 should be 2 2/3 stops faster than 2/3"
Click to expand...

To suggest that sensitivity is linearly dependant on sensor size and pixel density is ridiculous... There are so many other variables to consider that it's really impossible to calculate how the 2/3" sensor will react.

The only way to test is to point it at a chart and measure...
 
toke lahti said:
If we assume that 2/3" and S35 chips are same generation, their sensivity is straightly proportional to pixel area.
S35's pixels are 3.5 times larger than 2/3", so it should be 3.5 stops faster.
Click to expand...

I asked Jim about this a while ago and his answer was something along the lines of "It's not just about pixel density but about smart pixel design".

I assume that the 2/3 MX has different pixel design (whatever that means) to leverage more juice out of the smaller form factor using methods that don't translate as well to tuning a S35 sized chip. That's just pure speculation on my part though, I don't know how the sensor design team/Graeme's math pixies work.

What I would venture to guess is that the 2/3 chip's over all performance will be lower than the S35 MX, however this decrease in performance will not correlate directly with the smaller physical size.

So, in areas like dynamic range, low light performance etc. expect the 2/3 chip to perform around 1/2 to 3/4 as well as the S35 MX rather than 1/4 as well, simply because it's 1/4 of the S35's size.
 
From what I understand of DR, there should be minimal difference (maybe half a stop?) owing to pixel size at 0 gain. It is only at extreme ISOs that you can observe 2-3 stops of difference, as the greater noise of the smaller pixel intervenes. I am not sure how accurate this is, but I trust Jim's observation that the image quality difference between S35 and 2/3" at 3K is "slight".

Of course, the S35 does get you to 5K... And the image does hold up at far higher ISOs.
 
Dan Hudgins demonstrated that everything else being equal there is no significant difference in noise or sensitivity for pixel sizes between 3 and 7 microns. Sensitivity and noise do change with overall sensor size though.

There was some candle light test footage posted of an M-X RED1 test shot at 4k and 2k crop, 2k being close to the size of the 2/3" sensor. As expected there was a bit more noise or graininess in the 2k footage at the same exposures, but it was not a radical difference. Not as much as the difference between 35mm and 16mm film. 3k Scarlet should do a bit better than S35 cropped to 2k.
 
David Rasberry said:
Dan Hudgins demonstrated that everything else being equal there is no significant difference in noise or sensitivity for pixel sizes between 3 and 7 microns. Sensitivity and noise do change with overall sensor size though.
Click to expand...
Do you have a link to this?
David Rasberry said:
There was some candle light test footage posted of an M-X RED1 test shot at 4k and 2k crop, 2k being close to the size of the 2/3" sensor. As expected there was a bit more noise or graininess in the 2k footage at the same exposures, but it was not a radical difference. Not as much as the difference between 35mm and 16mm film. 3k Scarlet should do a bit better than S35 cropped to 2k.
Click to expand...
Red1's photosites are bigger than Scarlet2/3"'s. Windowing the picture does not change its sensivity.
 
David Rasberry said:
Dan Hudgins demonstrated that everything else being equal there is no significant difference in noise or sensitivity for pixel sizes between 3 and 7 microns. Sensitivity and noise do change with overall sensor size though.
Click to expand...
Could anybody give a link to this. I'd be very happy to refresh my knowledge.
David Rasberry said:
There was some candle light test footage posted of an M-X RED1 test shot at 4k and 2k crop, 2k being close to the size of the 2/3" sensor. As expected there was a bit more noise or graininess in the 2k footage at the same exposures, but it was not a radical difference. Not as much as the difference between 35mm and 16mm film. 3k Scarlet should do a bit better than S35 cropped to 2k.
Click to expand...
Which are the best known explanations for this difference being so small between 2/3" & S35?
Different pixel design? S35 being older (hence not so advanced)?
Does 2/3" use BSI and S35 doesn't?
 
David Rasberry said:
Sorry, wrong name. It was Daniel Browning.
http://reduser.net/forum/showthread.php?t=30076&highlight=myth+busted
Click to expand...
While I've learned a lot from Daniel (ETTR+ITTR=HAMSTTR),
that thread should have been called "Myth busted: sensors with smaller photosites..." like David Mullen commented in that thread.

I have been too focused on single pixels, so maybe I need to step up and see the wider picture of whole sensor.

What Daniel explaned in that thread was that with same sensor size, sensors having smaller & bigger pixels (and of course same pixel design), have equal amount of (perceptual) noise when downsized to same resolution.
And even that is argued over there.

But it doesn't change the fact that S35 sensor has 8 times larger area than 2/3".
So either way if you look at one pixel or the whole sensor, there should be big difference in sensivity.
At least if we think sensivity (=base ISO?) in basis of acceptable and comparable noise level.

So the question is why there is so little difference (at least in many people expectations)? Why 2/3" has so much more advanced pixel design that it performs almost as good as 8 times bigger sensor, which is as new as the other? Why S35 performs so much worse than 2/3" after compensated for their size difference? Is S35's pixel design much older? Would it be too expensive to use 2/3"s' more advanced pixel design in S35?

And if you guys feel that I'm beating a horse that hasn't even born yet, feel free to silence me! :)

ps. Link for the 2k cropped footage? (Can't find it...)
 
You must log in or register to reply here.
Back
Top