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 :)

Hannover Hands On comparisons of RED F35 & D21

Status
Not open for further replies.
Well, it's 3 smaller sensors rather than 1 large one. I'm assuming the sensor driving circuitry would be different? Who knows?

Graeme
 
I thought the F35 and Genesis were 3 sensor as well. I need to do some reading.
 
geoffboyle said:
I just had a very quick look at that, there is no exposure compensation in the SG settings.

Where using 320 IS0 results in an 8 bit clip of 255, setting it to just 250 ISO give a maximum value of 242.

At that difference I'm not sure that you're losing anything.
Click to expand...

ISO 250 in the Red SDK makes sure you aren't losing anything. Maybe the ISO 320 puts pure white at 1023 but I don't have the possibilities to confirm this. Hence I use ISO 250 as many others do as well, as it seems.

Hans
 
Filip Vandoorne said:
when we compare the dpx files from the different camera's, say at 'normal',
did all camera's had the same T-stop on the lens?
or was the 'normal' point for each camera determined with a graychart and spotmeter, iso value and set the each camera accordingly?
Click to expand...


We believed the manufacturers :-)

Well, actually we probably didn't but that's one of the things that comes out in a test like this!

Cameras were rated as recommended and exposed by taking spot readings from the grey chart, we also checked with incident readings.

Trying all the time to narrow down the variables.

Yes, that does mean different apertures, I'd like to have avoided that but we would have needed custom ND's to have done so and then I'd get people saying that the ND's had an effect.

Can't win :-)
 
Irri said:
Aren't F23/35/Genesis virtually the same cameras?
Click to expand...

F35 and Genesis are very similar if not identical, there are all kinds of rumours and silly stories about how they have been matched or not.

I think that the general opinion is that they are interchangeable.

The F23 is a very different animal as far as sensors go, 3 individual R, G, B sensors and a prism block.
 
geoffboyle said:
I just had a very quick look at that, there is no exposure compensation in the SG settings.

Where using 320 IS0 results in an 8 bit clip of 255, setting it to just 250 ISO give a maximum value of 242.

At that difference I'm not sure that you're losing anything.
Click to expand...

you definitely lose something on my tests using 320.. If you processed with no gamma correction at all (linear) then the 8bit value would mean more to me.

Good luck

Michael L
 
8bit linear light export would loose you around 4 stops DR. Not recommended.

Graeme
 
Graeme Nattress said:
8bit linear light export would loose you around 4 stops DR. Not recommended.

Graeme
Click to expand...

Well it could represent more linear stops but you would lose precision within those stops. 8Bit log files don't have any less dynamic range in them do they? They're just more bandy.
 
Sure, it would help you see the brightest stop. But if you did a 16bit linear light tiff, you'd see everything.

Graeme
 
Jeff Coatney said:
If this wasn't a comparison, why did you title the thread thusly:

"Hannover Hands On comparisons of RED F35 & D21" ?
Click to expand...

While it may have been a poor word choice, I believe what he means is he didn't intend it to be a competitive comparison, but rather an informative comparison to help make choices between different tools for different jobs. i.e. not an absolute better-than/worse-than comparison.
 
Gabe Conroy said:
While it may have been a poor word choice, I believe what he means is he didn't intend it to be a competitive comparison, but rather an informative comparison to help make choices between different tools for different jobs. i.e. not an absolute better-than/worse-than comparison.
Click to expand...


Thank you, somebody understands :-)
 
But that doesn't put butts on chairs Geoff, we need controversy, drama. What if the D21 and the F35 were in a long term relationship, but it was soon revealed that Red seduced the D21, and they had been seeing each other for months! Then maybe, the F35 could retaliate and try to make the D21 jealous with the much younger Scarlet! Yes! But of course... Epic would then come into the picture, and all bets would be off. Big cat fight! Epic cat fight! Everyone could embarrass themselves and after a while they would all realize how similar they really are, how we are all digital in the end, 0's and 1's you know, really deep stuff. Then out of nowhere, BAM! 645 rolls out in Q1 of 2010!! Well, that's all I got, but it could be something. Just needs a little work. We should shoot it on film though, just to piss everyone off, we'd get some free attention on the forums.
 
I can't stay away too much from this kind of discussions...

I have done a series of OBJECTIVE measurements of DR using the well known method of the Stouffer transimitive chart.

Taken a flat 16bit tiff image and put it in the Imatest noise analysis tool.

This software takes every step of the chart and analyzes the noise deviation in that specific step. It measures lets say in the -9th stop how much the noise deviate. If the noise deviate in a range of 1/10 of a stop it puts that step of the chart to the high quality level. If the noise is 1/25 of a stop it puts it in the mid high level and so on... there are five levels of noise:

_____________Noise levels in stop's_____________1/10 stop____1/4 stop___1/2 stop____1 stop_____5 stop

RED ONE 14/3/2008 b14 my test_________________6,69 stops__ 8,17 stops__8,9 stops__9,97 stops__9.98 stops
RED ONE 24/11/2008 b17 Alfonso Parra DoP test ___6,66 stops__ 8,16 stops__9,68 stops__9,97 stops__10 stops
Alfonso Parra DoP test F23______________________6,31 stops__ 8,38 stops__9,5 stops__10,6 stops__12,6 stops

For me the F23 has less good quality range than RED because it has a bit less range in the 1/10 noise range... that means RED can have brighter midtones with sufficient blacks... but someone could argue that the F23 can go up to 12,6 stops... I have to say the 11th and 12th stop of F23 have 2-3 stops noise... so they are absolutely unusable...

All the above is added to what Jim was trying to say about the M-X sensor... that its not what is the roof of the sensor in terms of noise but instead is how it performs in the whole range...

Additionally a very good proof, that this kind of objective DR measurement are very accurate and valid as a comparative method, is that my measurement of RED ONE at March of 2008 with b14 and eight months later Alfonso Parra measurement with b17 have in all steps apart from the 1/2 noise level a difference of 0,02stops... You have to note that I don't know the guy and I wasn't there when he was doing the tests... he just did them without me knowing him and our results are almost identical except of the Greeme magic that added with b17 firmware a good 2/3 stop more DR in the mid-low range...

According to Alfonso Parra that did also a multi exposure with Mcbeth and faces his findings was that the two methods are matching if you count the stops that are in the 1/3 noise level. We have to ask Norman Korren that is the author of Imatest to add this level also on his software to get an immediate result that will match the empirical method of multi exposures...

It would be very informative If I got a chance to measure a RED with b20 and see how it compares with the other builds...
 
Status
Not open for further replies.
Back
Top