Dynamic Range Measurment

[Divo]

Hello. I have a few questions about DR. Any help will be much appreciated.

1. I was wondering is it possible to measure the dynamic range of a scene based on different exposed RAWs and get a number let's say 10 000:1

2. Measure the DR captured in single RAW i.e. 1000:1

3. Measure a tone mapped image of same scene saved in 16-bit tiff and get DR 5000:1


The numbers I wrote are only an example, but they are related based on that DR of scene is greatest and single shot of it is lowest.
I've tried to load HDR (merged from 5) image into Photomatix and it says that DR is 2401:1. But when I've loaded single RAW (middle) it said that DR is 2828:1. How it is possible that single shot contains more dynamic range than merged HDR?

[David]

Hi Divo - Welcome to CiC. I've never worked out how Photomatix comes up with its DR numbers and now tend to ignore them. However, at one time I thought that getting the greatest value in Photomatix terms was the ideal. But, that is not the case. There is software called Picturenaut (Google the name to find it) that is free and arguably as good as Photomatix. However, during its calculations it displays the dynamic range of the HDR image. These calculations, expressed in Ev, do appear to be reliable or at least consistent. One problem with estimating DR in all HDR software is that the calculations are dependent on the precise algorithms used to combine images.

Cheers

David

[alvarito]

I use two methods to measure the dynamic range of a scene:
-Using the program Histogrammar by Guillermo Luijk (photo tiff from a RAW with a neutral revealed , but has the disadvantage of having to use a PC to check, which is not very effective)
-Measure with spot metering the highest light of the scene,
then measure with spot metering to the area of the scene with less light in which we want to get detail.
The difference in EV between the two areas define the dynamic range of our scene

[Divo]

Thank you very much for replies!

@David
Picturenaut is doing well with merged RAWs but in case of single shot it does not show DR at all

@alvarito
Interpreting Histogrammar numbers is little bit incomprehensible for me. I've take my +0EV RAW to DCRAW and processed like C:\dcraw>dcraw -D -d -r 1 1 1 1 -4 -T -t 0 IMG_2563.CR2 Then loaded it into Histogrammar and saw this

How to determinate DR since it says RGB: 65535 (100% of available), range [0..65534] to almost every tiff I load into it?

[alvarito]

Gui told me to reveal a totally neutral RAW file I have to apply the following command:

dcraw-v-r 1 1 1 1-o-g 0 -4 2.2 0-T filename

Moreover, the interpretation of dynamic range in EV measure it, much more intuitive
for the photographer that a relationship of brightness as you said.
In the photo you post, the dynamic range reaches 15 EV as I can see in the screenshot that you have put.
sorry, but my English is very bad

[Divo]

Gui told me to reveal a totally neutral RAW

By neutral RAW you mean normally exposed (+0EV)?

dcraw-v-r 1 1 1 1-o-g 0 -4 2.2 0-T filename

This command returned: Non-numeric argument to "-g".

much more intuitive for the photographer that a relationship of brightness as you said

It doesn't matter for me to get data in EV or contrast ratio, because it can be easily calculated, but exact number would be preferable than interpreting histogram.

[alvarito]

By neutral RAW you mean normally exposed (+0EV)?

histogram a picture attached to the right but not burn

This command returned: Non-numeric argument to "-g"

-V: display info during development
-R 1 1 1 1: white balance canceled
-O : no conversion to output color profile (to open in PS must assign a profile, eg Adobe RGB, the colors really will not make much sense)
-4 -g 2.2: gamma 2.2
-T: TIFF output

It doesn't matter for me to get data in EV or contrast ratio, because it can be easily calculated, but exact number would be preferable than interpreting histogram.

Sorry, but my knowledge of the subject do not give more, maybe if you ask the GUI directly, he'll be able to respond

regards

[Dave Humphries]

Is there a minor typo here that means there should be a space in front of "-g"?

e.g. dcraw-v-r 1 1 1 1-o -g 0 -4 2.2 0-T filename

I'm only guessing though

[alvarito]

then I copied the text as it sent me GUI therefore should be well

[_GUI_]

I'm probably very late to this thread, but anyway wanted to correct the DCRAW commands seen here, none of which seems to be correct. DCRAW is a command line program, so syntax (including spaces and upper/lower case) must be fully respected.

To obtain a EV distribution histogram to find out real scene's dynamic range (only subject to sensor's linearity errors) using Histogrammar, there are two ways:

1. NEUTRAL RAW DEVELOPMENT (this is simpler and probably as good as the other method)

dcraw -v -r 1 1 1 1 -o 0 -4 -T file.cr2

The resulting TIFF will be opened in Histogrammar setting the gamma slider to 1, and switching to Log display mode.


2. RAW EXTRACTION WITH WHITE AND BLACK POINT ADJUSTMENT (this is strictly what the sensor managed to capture)

dcraw -v -d -r 1 1 1 1 -4 -T -t 0 file.cr2

The resulting TIFF will be opened in Histogrammar setting the gamma slider to 1.0, switching to RAW mode, and switching to Log display mode.
If your sensor doesn't have and RGGB bayer pattern, some channels will appear interchanged. Check the other 3 possibilities in the coloured icons.

Regards

[Colin Southern]

Guillermo -- is this really you?

Seems like forever since we've seen you here!

How's everything going?

[_GUI_]

Very busy at work, it was my hard first year in a new job, new company, new sector, new boss. Hope this year things will be easier and I'll have more energies for photographing.

BTW I have bought a M4/3 PEN and cheap vintage lenses on eBay, having a lot of fun with the new stuff (despite its crappy sensor). Some test shots I did this morning in the Reina Sofia museum:











Regards