creating profile with expodisc cap on lens

look for ready-made device noise profiles and offer your own ones here
Post Reply
pip boogaloo
Posts: 9
Joined: Sun Jun 15, 2008 10:05 am

creating profile with expodisc cap on lens

Post by pip boogaloo »

hi all (using canon 1ds)
i have the excellent expodisc 18% highly accurate white balanced cap
this is a high technology device that basically clips onto the end of the lens and perfectly diffuses and completely blurs incoming ambient light entering the lens ordinarily for purposes of setting camera white balance

but because of the action of this device in how it gives a perfect 18% diffused incoming light - it seems that i have been getting good results when i have profiled experimented and created profiles with that on the lens noise - and then done the same for process other ISO images !!!

is there any reason why this is a bad idea and i should be using colour targets instead that would lead to better results

thanks in anticipation
NITeam
Posts: 3173
Joined: Sat Feb 01, 2003 4:43 pm
Contact:

Post by NITeam »

Our regular test target would be better because it provides patches of different color/brightness and all of them are required for accurate noise profiling. That is better than using just one brightness area even if it is perfect 18%.

Hope this helps,
Vlad
pip boogaloo
Posts: 9
Joined: Sun Jun 15, 2008 10:05 am

Post by pip boogaloo »

so from that can i take it neat image is active not only across the colour channels , but it is proportionate according to the intensity of the colour varying across the image

i.e noise is more likely in dark areas, so neat image applies more noise suppression there while leaving less prone highlights untouched ???

or

does it formulate one overall average noise profile, and then globally apply that to the image in its entirety ???
NITeam
Posts: 3173
Joined: Sat Feb 01, 2003 4:43 pm
Contact:

Post by NITeam »

It applies different treatment to different parts of image. That's why it can achieve high-quality results. With just one "average" profile the results would be worse.

Vlad
pip boogaloo
Posts: 9
Joined: Sun Jun 15, 2008 10:05 am

Post by pip boogaloo »

this is interesting , as i have had neat image for a while now - and i like it alot - its good software

but

i have had a chance to do alot of testing and my average grey is working very good indeed, unlike a computer screen of even the highest quality, they are still poor compared to the quality of the light hitting the sensor and other inconsistencies screens have

but i fully understand the logic in what you have said, is stands to reason that more samples or varying luminosity/colour will produces more random noise for sampling based on the mechanics of neat image.

based on what i have experimented with so far, is it possible to sample noise from different images and formulate the results into one profile.

what im thinking is that to sample many scenes through this filter at different brightnesses ranging from bright to dark and to dump the rusults into one profile.

it doesnt matter if is grey or colour surely - as the cameras sensor only operates in rgb anyway , so even photographing a bright grey scene the values would be just for examples sake 200,200,200, and eventually going down to 30,30,30 for a darker grey (but still rgb inside camera for sampling different channels)...........
pip boogaloo
Posts: 9
Joined: Sun Jun 15, 2008 10:05 am

Post by pip boogaloo »

i have thought of a way how to do it while pondering.........

but please respong to my rgb v grey sampling
NITeam
Posts: 3173
Joined: Sat Feb 01, 2003 4:43 pm
Contact:

Post by NITeam »

Yes, any image, including grey, contains noise in all of its r, g, b channels so analysing a grey target (the grey version of Neat Image calibration target) produces good results too.

Vlad
Post Reply