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 :)

DR of Scarlet: Only 7.5 usable Stops ?

Sascha Rossier

Sascha Rossier

Well-known member
Joined
Sep 27, 2009
Messages
233
Reaction score
0
Points
0
Location
Zürich, Switzerland
Website
www.sascharossier.com
i can't seem to get my head around the usable/practical stops on the Scarlet ... (maybe its just me ... ;-)) i mean
i always get great footage but i seem to expose my scarlet only around what i've learnt from the usable 6 Stops of my 7D.
On Red.com its been said, that the Epic/Scarlet has a DR of 13.5 Stops (even 18 with HDRx) but this seems
to be strictly technical and may be true in a laboratory-testing-environment.

I bought a Sekonic 478 a few week ago an did some calibrating/testing ... according to the Tests/Calibration
with the Sekonic, the usable DR of the RED is more around 9.5 Stops.

red_dynamikumfang.JPG


and to add to the confusion, working with other "Videocameras" i've learnt, that exposures in the lower 20% and the
upper 20% of a Cameras DR are not giving usable results in Grading. This would leave me with a total of 7.5 Stops.
4.5 above and 3 below my target-Stop. This is cool enough for any lighting-ratio between 1:1 and 16:1, but still ...
somehow i have the bad feeling that there must be more Gaia in my RED than i've been able to pull out yet.

Any Gurus in here that can shed some light into my chaotic mix of thoughts? Some Knowledge in this matter would
be much appreciated ...
 
I did what the Sekonic Users Manual asked me to do ... greycard/chipcard shooting with the RED and then using the included DTS Software to convert the shots into a calibration file for the Lightmeter. Pretty much like the Dude in this Video:
http://www.youtube.com/watch?v=tdyosItw3Mk (exept that i used a Sekonic 478 and a SCARLET ;-)

Then i got the 9.5 Stops and it felt odd, caus e everybody keeps saying Scarlet reaches 13.5 Stops.
So i compared my Calibration Results with The Results of Ryan E Walters who pretty much got the same Values.

So i came over here to ask the big guys, if I am missing something ;-)
 
I saw this in the other thread yesterday. I think that top 0.5 stop is useable; there seems to be definition between +4.5, +5, and clip. Also between -5 and -6 there seems to be signal, which would put you at 11 stops instead of 9.5. If the readings are correct (what ISO were you at?), it could be because 4k uses less sensor, it doesn't have as much usable DR compared to 5k/Full Frame.

Also, you're probably not going to get many responses since I don't think many Scarlet owners have the access to the proper equipment... I will say anecdotally that when I look at footage from other cameras - Alexa frames, DSLR stills, or even BMCC - they seem to have more DR than Scarlet/MX. In the case of the BMCC, it's supposed to have 13 stops of latitude, which is less than MX's 13.5... So who knows.

EDIT: You know what else; maybe take another camera you have (e.g. a DSLR or something with a known, DR -- 11stops or something) and run test on that... If it is only 8 usable stops than you can ascertain that the test might not be the most accurate.
 
Last edited:
The results of your test are correct, but your interpretation of the the results is a little off. If you enter the exposure mode on the scarlet and set your middle grey point properly, and you add light one stop at a time then it will clip 4.5 stops up at 850 ISO, and if you take away one stop at a time from middle grey then there will be 4-5 stops before you get purple, the noise floor. I forget off the top of my head if it is 4 or if it is 5, I can confirm once I get back from a trip. The results from the sekonic seem pretty accurate. Now you have to take a look at what the noise floor really means. The noise floor is just the point of which lifting the darks will add noise to the image, that is all. So there are still a few stops beneath the noise floor where there is information and definition, and in the grade you can still push them up and see more into the darks, the only downside is that there will be noise. Most of the time things that are more than 5 stops beneath middle grey should be dark dark black or true black, usually a regular contrast adjustment will crush blacks and produce a very very clean image if exposed properly. So if you don't wan't to remove noise in your grade then only important parts of the image that you do not want to be dark dark black need to be above the noise floor. My experience with the scarlet tells me that the bottom 20 and top 20 percent rule doesn't really apply with this camera. As long as information has not been clipped there is room to push it around. Obviously overexposed areas close to clipping will not give you as much room to play as middle grey, but as long as bright objects are at least .5 stops away from clipping you have all the room you need. If your following normal exposure rules and trends everything should be in a 3+- stop range anyway unless the style dictates something more extreme, the scarlet gives you and extra wiggle room of 1.5 stops in the highlights and 1-2 stops in the darks to be extra safe in regular cases, and you can get away with more as long as you are willing to denoise.
 
Maybe Ryan will chime in. I know he's doing over/under exposure tests of his own currently in his "Some like it Raw" series.
 
Mike P. said:
I saw this in the other thread yesterday. I think that top 0.5 stop is useable; there seems to be definition between +4.5, +5, and clip. Also between -5 and -6 there seems to be signal, which would put you at 11 stops instead of 9.5. If the readings are correct (what ISO were you at?), it could be because 4k uses less sensor, it doesn't have as much usable DR compared to 5k/Full Frame.

Also, you're probably not going to get many responses since I don't think many Scarlet owners have the access to the proper equipment... I will say anecdotally that when I look at footage from other cameras - Alexa frames, DSLR stills, or even BMCC - they seem to have more DR than Scarlet/MX. In the case of the BMCC, it's supposed to have 13 stops of latitude, which is less than MX's 13.5... So who knows.
Click to expand...

At 850 ISO there is a signal at up to 7 stops under, there is no information past 4.5 stops up at 850 ISO, you can underexpose the sensor more to get 5 but you will be introducing more noise, and after normalizing the image you will have equivalently been shooting at 1600 ISO.

4k has no less dynamic range as 5k, its exactly the same. Both the epic and scarlet have 12.5-13 stops of dynamic range depending on subjective opinion. That is calculated by the points of which sensor is no longer collecting information due to under exposure and over exposure. The term you guys are looking for is exposure latitude, how many stops of exposure can be manipulated without degradation of the image. That is different than dynamic range and that is what the sekonic ended up calculating. 9.5 useable stops is amazing guys. That means if your exposed correctly and you keep everything in an 8 stop range then you can manipulate the image pretty freely without degrading the image during the grade. Thats a lot of room, but to put that in perspective daylight is 10000 foot candles, 8 stops beneath that 40 foot candles.
 
Hej Guys ... that sounds awesome! I never was complaing, I just thought i must be doing something wrong. But after your comments, i'm feeling even better with my little red-monster ;-)
And thanx ryan ... i'll adapt to the term 'exposure latitude' then ... Hej Mike ... sounds interesting ... i didn't know Ryan E. was still on his Some Like it Raw Series ... looking forward to a few new infos from the edge of exposure ;-)
 
Yeah, I figured... I was grasping at straws. And yes, exposure latitude is much different than DR.

One thing I'm seeing on those charts is that the 5D seems to have ~8 (-4 to +4) DR, but all of it seems usable (presumably the Mark2 and newer sensors would have more DR, too), which would make a lot of sense based on my anecdotal experience. Alexa is a beast -- I think they mean 14 stops exposure latitude, because it looks to record almost 16 stops of signal.
 
Sascha Rossier said:
And by the Way ... these are Ryans Measurements for the 5D and the Alexa:

5D:
dyn5D.JPG


Arri Alexa:
dynAlexa.JPG
Click to expand...
These results seem really accurate, it shows that you get 1.5 stops extra of exposure latitude on a camera that is marketed as have 1.5 extra stops of dynamic range. It also shows that the scarlet blows the 5d away, having far more exposure latitude. It also shows that underexposed alexa and red should be very similar so if you are underexposing, luminance wise you wont be getting any more with Alexa than Scarlet.
 
What gamma space were you shooting at? Redgamma3 or redlogfilm?
 
You must log in or register to reply here.
Back
Top