Topics

[CCML]+CML question of how to confirm ISO rating of digital camera/如何決定ISO?

Tony Ngai
 

All;

This is just re-cap and re-send , using english on following question encountered in the CCML(chinese version of CML):

In first edition of ASC video manual 1972, page 137 by Mr. Harry Mathias, Harry mentioned “How to establish an exposure index for a video camera”

Back then, Video camera are more simple, NTSC or PAL, one fix gamma curve, but nowadays, digital cameras have:

Raw, log, several gamma curve compounded by several color gamut, plus even dual ISO rating, the question from the floor is that :

How should one established , objectively, the ISO rating of nowadays digital cameras?

Pls advise.
Many thanks

Tony Ngai
Main driver of CCML


To: CCML@groups.io
Subject: Re: [CCML] question of how to confirm ISO rating of digital camera/如何決定ISO?
Reply-To: CCML@groups.io,tony@...

各位有一近期常被問到之技術問題, i.e 如何决定現代電影數碼攝影機之ISO rating?

在first edition of ASC video manua 1972l , page 137 by Harry Mathias, Harry 有一頁文章,
“How to establish n exposure index for a video camera”!

當然那是1972年,現在是2018年, 現在的digital camera has:
“Raw, Log, various game and some has dual gain ISO rating”

那麼什様決定(客觀地)camera’s ISO?呢?

Many thanks,

as I mentioned, I will mirror this in CML to gain benefit of both worlds(language),



Tony Ngai 魏天明 
Main driver of CCML

(Mirror of concept of CML)



General Manager - Production & Technical Service
總經理 - 製作及技術服務

  
4/F, Freder Centre, Unit C, 68 Sung Wong Toi Road, Kowloon, Hong Kong
香港九龍宋皇臺道68號飛達工商業中心四字樓C
T:  (852) 2338-6311                      M:  (852) 6191-4222  
M (China)147-1603-4039          QQ nr: 2257169479
Email: tony@...  Website: www.salonmedia.com
Skype: Salon_tony










Tony Ngai 魏天明 
General Manager - Production & Technical Service
總經理 - 製作及技術服務

Signature
4/F, Freder Centre, Unit C, 68 Sung Wong Toi Road, Kowloon, Hong Kong
T:  (852) 2338-6311                     
M (China):  86-147-1603-4039   
QQ nr2257169479
Email: tony@...    
Skype: Salon_tony


JD Houston
 


On May 9, 2018, at 11:13 PM, Tony Ngai <tony@...> wrote:

How should one established , objectively, the ISO rating of nowadays digital cameras?

There is not currently an objective standard for digital motion picture cameras to set ISO ratings.

There is a subjective element in current manufacturers settings of that value which comes from 
evaluation of the usable sensitivity and noise floor of the camera system, since the ultimate question
is how usable images are given a certain amount of exposure at 24fps et.al.

There have been discussions about how to do this scientifically within several organizations but there are no
results at this time.   Ideally, a radiometric specification of incoming light would produce the 
same result on every camera.

Unfortunately, there is enough variance in the actual ISO rating that a camera may be rated 854 ISO
by an official system, but buyers and users prefer to call it an 800 ISO for convenience and easy manipulation
along the exposure scale (ISO 400  ISO1600, etc. )

So no solution in sight at the moment.


Jim Houston
Consultant, Starwatcher Digital, Pasadena, CA

Geoff Boyle
 

Oh yes there is 😊

 

Just download the EXR files from CML and look at the results.

 

The cameras are all exposed at ISO800, with the exception of the Venice which is ISO500, light levels are set using a mixture of exposure meters.

 

Any variation in the EXR’s from “correct” exposure is down to the cameras not being ISO800 as specified by the manufacturer.

 

The amount of variation gives you an idea of the “real “ speed of the camera. Usually pretty small but occasionally a wonderful work of fiction!

 

 

Cheers

 

Geoff Boyle NSC FBKS

Cinematographer

Netherlands

www.gboyle.co.uk

 

 

 

From: cml-general@... <cml-general@...> On Behalf Of JD Houston

 

There is not currently an objective standard for digital motion picture cameras to set ISO ratings.

 

 

JD Houston
 


On May 9, 2018, at 11:13 PM, Tony Ngai <tony@...> wrote:

How should one established , objectively, the ISO rating of nowadays digital cameras?

There is not currently an objective standard for digital motion picture cameras to set ISO ratings.

There is a subjective element in current manufacturers settings of that value which comes from 
evaluation of the usable sensitivity and noise floor of the camera system, since the ultimate question
is how usable images are given a certain amount of exposure at 24fps et.al.

There have been discussions about how to do this scientifically within several organizations but there are no
results at this time.   Ideally, a radiometric specification of incoming light would produce the 
same result on every camera.

Unfortunately, there is enough variance in the actual ISO rating that a camera may be rated 854 ISO
by an official system, but buyers and users prefer to call it an 800 ISO for convenience and easy manipulation
along the exposure scale (ISO 400  ISO1600, etc. )

So no solution in sight at the moment.


Jim Houston
Consultant, Starwatcher Digital, Pasadena, CA


JD Houston
 


Yes, that tells you the practical of how are the cameras different?  The IDTs sometimes hide details in converting to EXR.
Can you tell me which one is the correct one?  There should only be one answer.  It would be an 0.18 EXR if an exact method
was followed, but then how many lux on set turned into that.

:-)

Jim

On May 10, 2018, at 2:08 AM, Geoff Boyle <geoff.cml@...> wrote:

Oh yes there is 😊
 
Just download the EXR files from CML and look at the results.
 
...
 
Any variation in the EXR’s from “correct” exposure is down to the cameras not being ISO800 as specified by the manufacturer.
 
The amount of variation gives you an idea of the “real “ speed of the camera. Usually pretty small but occasionally a wonderful work of fiction!

Jim Houston
Consultant, Starwatcher Digital, Pasadena, CA

Geoff Boyle
 

That's why I use the manufacturers software for creating the EXR's.

The light level on set is constant as is the aperture used so any variations from "Normal" on the 0 stop frames is a variation away from ISO 800.

Light levels are checked with a Pentax Digital Spot, Minolta 6 and my venerable Minolta 4 flash with the tiny exposure probe attached.

Geoff

Tony Ngai
 


Many thanks to all!
多謝答覆


Tony Ngai 魏天明 
Main Driver of CCML


Geoff Boyle <geoff@...> 於 2018年5月10日 下午6:30 寫道:

That's why I use the manufacturers software for creating the EXR's.

The light level on set is constant as is the aperture used so any variations from "Normal" on the 0 stop frames is a variation away from ISO 800.

Light levels are checked with a Pentax Digital Spot, Minolta 6 and my venerable Minolta 4 flash with the tiny exposure probe attached.

Geoff


______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please visit http://www.symanteccloud.com
______________________________________________________________________

Art Adams
 

Light levels are checked with a Pentax Digital Spot, Minolta 6 and my venerable Minolta 4 flash with the tiny exposure probe attached.

Weren't you putting middle gray at code value 512? Was that on set or in post?

--
Art Adams
Director of Photography
San Francisco Bay Area

Art Adams
 

My method is to look at the image and see where the noise becomes bothersome, but then cap the lens and look at a waveform to see how thick the noise floor is at various EIs. That way I have a semi-objective measure of how much noise there is in the image, and in which channel, as some monitors are very good at hiding noise and others exaggerate it.

From there on, it's just an intuitive process of matching the waveform results with what I see by eye and figuring out where I feel comfortable exposing the camera. I haven't really found an objective noise floor trace thickness to go by, it's just instinct and experience.

I did this the first time I shot a project with an Alexa. It was a VFX project, and I went through this process at the rental house and came up with EI 400 as a very clean EI for VFX. The post house loved me. (I now rate Alexa at 800 for most things as I've loosened up a bit on how much noise I'm willing to tolerate, and I've realized I'm hypersensitive to such things, but there's always a difference in what one can get away with in live action vs. VFX.)

By the way, I love that Arri refers to those settings as EI (exposure index) which basically means "If you set your meter this way you'll get predictable results." Other manufacturers use ISO, and in theory a camera only has one ISO setting. As soon as you go away from that, you're no longer shooting at an ISO standard. For example, if I set an F55 to ISO 640, that's not really ISO 640, because Sony says the native ISO is 1250, and you can't have more than one ISO. Any deviation from the native ISO is actually EI.

The reality is that the ISO standard is so malleable you can do whatever you want, or so I'm told. That's why lots of different cameras set to the same ISO can show different amounts and types of noise. I like EI more, because that's just a way to calibrate your meter. 

Arri suggests a "sweet spot" EI 800, but it's not the same as saying "This is what the ISO standard dictates." Rather, it's "This is where we think you get the most bang for your buck."

--
Art Adams
Director of Photography
San Francisco Bay Area

Geoff Boyle
 

Post, when I shoot it falls where it falls.

It's the only way.

It's only the rendered sequences that are tweaked in any way, if you want to see exactly what the camera does then the EXR files are there.

cheers

Geoff Boyle NSC
cinematographer
travelling


From: cml-general@... <cml-general@...> on behalf of Art Adams <art.cml.only@...>
Sent: Thursday, May 10, 2018 5:56:44 PM
To: Cml-General
Subject: Re: [cml-general] [CCML]+CML question of how to confirm ISO rating of digital camera/如何決定ISO?
 
Light levels are checked with a Pentax Digital Spot, Minolta 6 and my venerable Minolta 4 flash with the tiny exposure probe attached.

Weren't you putting middle gray at code value 512? Was that on set or in post?

--
Art Adams
Director of Photography
San Francisco Bay Area

Art Adams
 

Perfect. Thanks, Geoff.

--
Art Adams
Director of Photography
San Francisco Bay Area

Tony Ngai
 

Dear Geoff and Art:

Many thanks for the answer, one more question, why nowadays digital camera donot have spec as:

noise : xx db

as if conventional video camera in the old days?


Tony Ngai 魏天明 





Geoff Boyle <geoff.cml@...> 於 2018年5月11日 上午12:15 寫道:

Post, when I shoot it falls where it falls.

It's the only way.

It's only the rendered sequences that are tweaked in any way, if you want to see exactly what the camera does then the EXR files are there.

cheers

Geoff Boyle NSC
cinematographer
travelling


From: cml-general@... <cml-general@...> on behalf of Art Adams <art.cml.only@...>
Sent: Thursday, May 10, 2018 5:56:44 PM
To: Cml-General
Subject: Re: [cml-general] [CCML]+CML question of how to confirm ISO rating of digital camera/如何決定ISO?
 
Light levels are checked with a Pentax Digital Spot, Minolta 6 and my venerable Minolta 4 flash with the tiny exposure probe attached.

Weren't you putting middle gray at code value 512? Was that on set or in post?

--
Art Adams
Director of Photography
San Francisco Bay Area



______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please visit http://www.symanteccloud.com
______________________________________________________________________

Art Adams
 

>as if conventional video camera in the old days?

That I don't know. My guess is that because the look is rarely created in the camera these days, and most cameras have a variety of gamma curves to choose from anyway. But I could be wrong.

--
Art Adams
Director of Photography
San Francisco Bay Area

Argyris_Theos_cml
 

HI Tony,
I believe the reason has to do with the target group these cameras are aiming at.
In the past tv stations would buy cameras based on the advice of video engineers and specs were built to be understood by them.
Today most cameras are evaluated/chosen by people who care more about overal picture quality (including subjective issues) and less about numbers.
So manufactures spec accordingly, I guess.
One more thing about your questions:
3 chip prism cameras did not have a constant sensitivity. This would vary according to the aperture on the lens. At wide apertures sensitivity would fall. Slightly, I must add, but it would fall. That's why they would declare (e.g) 2000lux at f11. This did not mean 125lux at f2.8. It would mean <125lx@...
Best

Argyris Theos, gsc
DoP, Athens Greece,
theos@...
+306944725315
Skype Argyris.Theos
www.vimeo.com/argyristheos
via iPhone

11 Μαΐ 2018, 2:43 πμ, ο/η Tony Ngai <tony@...> έγραψε:

why nowadays digital camera donot have spec as:

noise : xx db

as if conventional video camera in the old days?

Tony Ngai
 

Dear Argyris:

Many thanks for your feedback, loud and clear!



Tony Ngai 魏天明 
Main Driver for CCML


Argyris_Theos_cml <cml@...> 於 2018年5月11日 下午12:18 寫道:

HI Tony,
I believe the reason has to do with the target group these cameras are aiming at.
In the past tv stations would buy cameras based on the advice of video engineers and specs were built to be understood by them.
Today most cameras are evaluated/chosen by people who care more about overal picture quality (including subjective issues) and less about numbers.
So manufactures spec accordingly, I guess.
One more thing about your questions:
3 chip prism cameras did not have a constant sensitivity. This would vary according to the aperture on the lens. At wide apertures sensitivity would fall. Slightly, I must add, but it would fall. That's why they would declare (e.g) 2000lux at f11. This did not mean 125lux at f2.8. It would mean <125lx@...
Best

Argyris Theos, gsc
DoP, Athens Greece,
theos@...
+306944725315
Skype Argyris.Theos
www.vimeo.com/argyristheos
via iPhone

11 Μαΐ 2018, 2:43 πμ, ο/η Tony Ngai <tony@...> έγραψε:

why nowadays digital camera donot have spec as:

noise : xx db

as if conventional video camera in the old days?


______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please visit http://www.symanteccloud.com
______________________________________________________________________

Blain Brown
 


Many thanks for the answer, one more question, why nowadays digital camera donot have spec as:

noise : xx db

That’s a great idea. Let’s hope manufacturers pick up on it. 

Blain Brown
DP
LA

Geoff Boyle
 

I know that Canon do, it’s in their charts showing dynamic range V EI.

 

It would be useful if everyone else published charts like these.

 

 

 

Cheers

 

Geoff Boyle NSC FBKS

Cinematographer

Netherlands

www.gboyle.co.uk

 

 

From: cml-general@... <cml-general@...> On Behalf Of Blain Brown

 

Many thanks for the answer, one more question, why nowadays digital camera donot have spec as:

 

noise : xx db

 

That’s a great idea. Let’s hope manufacturers pick up on it. 

 

 

Tony Ngai
 

Geoff:

多謝指教!

Great info
Thank You


On 12 May 2018, at 1:51 PM, Geoff Boyle <geoff@...> wrote:

I know that Canon do, it’s in their charts showing dynamic range V EI.

 

It would be useful if everyone else published charts like these.

 

 

 

Cheers

 

Geoff Boyle NSC FBKS

Cinematographer

Netherlands

www.gboyle.co.uk

 

 

From: cml-general@... <cml-general@...> On Behalf Of Blain Brown

 

Many thanks for the answer, one more question, why nowadays digital camera donot have spec as:

 

noise : xx db

 

That’s a great idea. Let’s hope manufacturers pick up on it. 

 

 

<canonc300markiic-log2schematic1.jpg>

______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please visit http://www.symanteccloud.com
______________________________________________________________________

Jeff Kreines
 

I suspect each manufacturer will use their own standards, that perhaps might favor their cameras. 

Nah, they’d never do that!

Jeff Kreines

Sent from iPhone. 

On May 12, 2018, at 12:51 AM, Geoff Boyle <geoff@...> wrote:

It would be useful if everyone else published charts like these.

 

Daniel Rozsnyó
 

If we incorporate the SNR into the chart, it looks much clearer. Notice how the gray (noise) component gets more dominant.
Of course Canon do not want you to spot that easily, even that everybody knows high iso is high noise.

They try to claim 15 stops, but the best you get is 11.17 and worse is 7.83 stops (linear bits) SNR wise.

Daniel Rozsnyo
camera developer
Prague, Czech Republic



On 05/12/2018 07:51 AM, Geoff Boyle wrote:

I know that Canon do, it’s in their charts showing dynamic range V EI.

 

It would be useful if everyone else published charts like these.

 

 

 

Cheers

 

Geoff Boyle NSC FBKS

Cinematographer

Netherlands

www.gboyle.co.uk

 

 

From: cml-general@... <cml-general@...> On Behalf Of Blain Brown

 

Many thanks for the answer, one more question, why nowadays digital camera donot have spec as:

 

noise : xx db

 

That’s a great idea. Let’s hope manufacturers pick up on it.