Pages: 1 ... 6 7 [8] 9 10 ... 13   Go Down

Author Topic: The Optimum Digital Exposure  (Read 64523 times)

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: Re: Re: The Optimum Digital Exposure
« Reply #140 on: November 02, 2014, 06:33:55 pm »

I should've noted that when I gain up the sensor with the intent of producing noise it's with JPEG use in mind. Thus the only RAW converter in play is the one on-board the camera.

-Dave-

I'm not so sure I'd want to hire you for a photographic project, Dave.  ;D

Consider the following example. You get a call from a client who wants a noisy, murky, blurry shot for a specific customer who has requested an image that produces a nostaligic effect harking back to the early days of photography, for the purpose of selling a specific product.

You go to some trouble and expense setting up the conditions and hiring suitable models. The action begins and you take numerous shots, deciding to use the camera in jpeg mode and set to ISO 6400, perhaps also using a neutral density filter to slow down the shutter speed.

From the numerous shots you've taken, one is just right. The expression on the model's face is the most appropriate and you've captured the action well. The customer is delighted with the shot.

Some time later you get another call from the same client who has great news for you and excitedly tells you that another very important and wealthy customer has seen your blurry and noisy shot and likes the general theme and composition so much that he would like to have a huge, tack-sharp, noise-free print of the image. He'll pay big  money.

Alas! You inform your client that you only took noisy jpegs on that occasion and that none of the images would be suitable to produce a large, normally sharp, modern-looking print with smooth tones.  >:(
Logged

Wayne Fox

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4237
    • waynefox.com
Re: The Optimum Digital Exposure
« Reply #141 on: November 02, 2014, 11:55:36 pm »

Just finished glancing through this thread, curious if there is a way to see the image that started all of this?  I must of have missed it.

whoops, sorry wrong thread.
« Last Edit: November 03, 2014, 02:21:10 pm by Wayne Fox »
Logged

stamper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5882
Re: The Optimum Digital Exposure
« Reply #142 on: November 03, 2014, 03:30:22 am »

It is an article from the LULA main site.

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: The Optimum Digital Exposure
« Reply #143 on: November 03, 2014, 04:11:09 am »

Just finished glancing through this thread, curious if there is a way to see the image that started all of this?  I must of have missed it.

Hi Wayne,
The image that started it is at Reply #102 where Jeremy posted the famous Henri Cartier Bresson shot of a man jumping a puddle, claiming it was grainy, noisy and blurry, yet lots of people liked it.

Also, in Reply #121 Rand expressed the opinion if the image were grain-free and tack sharp it would lose its tension and mystery and just become another sterile reportage photo. He likes the murky quality.
Logged

stamper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5882
Re: The Optimum Digital Exposure
« Reply #144 on: November 03, 2014, 04:15:09 am »

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: The Optimum Digital Exposure
« Reply #145 on: November 03, 2014, 04:34:33 am »

I doubt there is such camera. RAW noise has a gaussian/Poisson distribution on all 4 channels in any Bayer sensor, and how this noise translates to the final image depends mainly on the RAW development software and processing applied (noise reduction, etc...). Same software and processing, equal noise appearance for the same level of input noise.

The best way to have nice noise in the final image on a digital camera is to capture as noise free as possible, then add noise. Whether these manoeuvres make sense or not is subjective.
Added noise is definitely not the same as noise produced during capture. So if you prefer the noise produced by the camera, then that is the way to go.
Also the lack of detail caused by using a low quality/high ISO image is different from a tack sharp image with added noise. Doing nice lo-fi imagery is as tricky, possibly even harder than high quality photos.
Logged
Tradition is the Backbone of the Spinele

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: Re: Re: The Optimum Digital Exposure
« Reply #146 on: November 03, 2014, 04:39:48 am »

Added noise is definitely not the same as noise produced during capture. So if you prefer the noise produced by the camera, then that is the way to go.
Also the lack of detail caused by using a low quality/high ISO image is different from a tack sharp image with added noise. Doing nice lo-fi imagery is as tricky, possibly even harder than high quality photos.

Of course is not the same, added noise can be MUCH better than camera noise because it can be added to your liking in amount, appearance and even location. That is why I said the best way to have beautiful noisy images starts from a clean RAW capture.

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: The Optimum Digital Exposure
« Reply #147 on: November 03, 2014, 04:55:37 am »

Of course is not the same, added noise can be MUCH better than camera noise because it can be added to your liking in amount, appearance and even location. That is why I said the best way to have beautiful noisy images starts from a clean RAW capture.
Not necessarily the best way for the reasons I pointed out in my previous reply. I don't necessarily want 'beautiful' noisy images, I may want crappy noisy images.  :P 
Post production cannot always replicate camera or film flaws as you may want them.
Logged
Tradition is the Backbone of the Spinele

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re:
« Reply #148 on: November 03, 2014, 05:08:15 am »

The point is camera noise is not pleasant, but some statistical knowkedge is needed to understand why.

Film is a totally different story and can be very beautiful because of its analogue nature, don't mix it with digtal noise.

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: The Optimum Digital Exposure
« Reply #149 on: November 03, 2014, 05:24:26 am »

The point is camera noise is not pleasant, but some statistical knowkedge is needed to understand why.
Don't confuse your preferences with other people's likes/dislikes and knowledge of statistics is definitely not necessary to decide what you like.

Quote
Film is a totally different story and can be very beautiful because of its analogue nature, don't mix it with digtal noise.
I wasn't, but digital camera noise can also look good in my view.
Colour noise however I completely loathe, but as that is very easily removed [and part of my default settings] it never bothers me. At times this luminance noise can can be very film like and therefore pleasing to me, though it all depends on your particular sensor, ISO/exposure combination.
Logged
Tradition is the Backbone of the Spinele

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8914
Re: The Optimum Digital Exposure
« Reply #150 on: November 03, 2014, 05:47:40 am »

I don't necessarily want 'beautiful' noisy images, I may want crappy noisy images.  :P

Practice makes perfect crap.

Quote
Post production cannot always replicate camera or film flaws as you may want them.

There are tools available that (besides improving images) can also be used to assist (simulate Lomography and light leaks, add grain, etc.) in the process of crapifying ...

Cheers,
Bart
« Last Edit: November 03, 2014, 08:40:16 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: Re: Re: The Optimum Digital Exposure
« Reply #151 on: November 03, 2014, 05:49:52 am »

digital camera noise can also look good in my view.

You are right, a matter of having good taste :P

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: The Optimum Digital Exposure
« Reply #152 on: November 03, 2014, 07:35:50 am »

You are right, a matter of having good taste :P
Well let me know if you get some then. ;)
Logged
Tradition is the Backbone of the Spinele

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: Re: Re: The Optimum Digital Exposure
« Reply #153 on: November 03, 2014, 08:52:43 am »

Colour noise however I completely loathe, but as that is very easily removed [and part of my default settings] it never bothers me. At times this luminance noise can can be very film like and therefore pleasing to me, though it all depends on your particular sensor, ISO/exposure combination.

It does not depend on the sensor, all sensors will provide the same noise appearance as long the same software is used to develop and process the RAW files. Some statistical knowledge is needed to understand why though.

In fact it doesn't make sense to talk about colour noise when talking about sensors since there is no colour noise in a RAW file. What you call colour noise is the result of a particular RAW demosaicing algorithm (software) propagating the gaussian/Poisson noise in captured RAW data and giving it a shape in the RGB rendered image.

williamchutton

  • Newbie
  • *
  • Offline Offline
  • Posts: 7
Re: The Optimum Digital Exposure
« Reply #154 on: November 03, 2014, 10:14:28 am »

Added noise is definitely not the same as noise produced during capture. So if you prefer the noise produced by the camera, then that is the way to go.
Also the lack of detail caused by using a low quality/high ISO image is different from a tack sharp image with added noise. Doing nice lo-fi imagery is as tricky, possibly even harder than high quality photos.

Photon-noise statistics are well-predicted by Q.M. Adding noise in post-production to simulate increased photon noise would be straightforward. Several camera brands offer an option to record lossy, compressed, raw data.  The more clever compression algorithms act as noise filters that selectively affect the brightest regions in the image. While irreversably manipulating original data is always a bad idea, in this case most of the information loss (photon noise) can be added back to the image during post production.

Read noise is different matter. But modeling read noise is possible. People who are cursed with making parameter estimated from data with oppressive signal-to-nose ratios (astronomers for instance) are forced to include a model for the noise in their parameter-estimation calculations. It is common to include several terms to the properly model the noise. Including a model for the noise reduces the uncertainty in the signal parameter estimates of interest. Sometimes adding noise computed from candidate models to high SNR data is an efficient way to empirically refine the noise model.
Logged

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: The Optimum Digital Exposure
« Reply #155 on: November 03, 2014, 11:04:41 am »

It does not depend on the sensor, all sensors will provide the same noise appearance as long the same software is used to develop and process the RAW files. Some statistical knowledge is needed to understand why though.
Funny that as different cameras/sensors I've used give different noise characteristics. Some were quite crappy, others OK - with the same software.

Quote
In fact it doesn't make sense to talk about colour noise when talking about sensors since there is no colour noise in a RAW file. What you call colour noise is the result of a particular RAW demosaicing algorithm (software) propagating the gaussian/Poisson noise in captured RAW data and giving it a shape in the RGB rendered image.
I understand how raw files are made usable, but it still makes perfect sense to refer to colour noise as colour noise, because that is exactly how it looks. We don't look at the raw data, we look at images. after demosaicing. You may as well argue the entire raw file is B+W despite the fact it looks pretty colourful when viewed by humans or that say a camera isn't really a camera, it's actually metal, glass and plastic.

What you appear to be doing is arguing that the word 'colour' is not actually colour or the concept [English speaking] people understand by colour, but is six individual letters that have no meaning.
The phrase reductio ad absurdum was invented for such ways of arguing.

BTW I studied stats at uni and school and do not really care about what distributions are used in the data, I look at the images not the physics and maths that produced them.
Logged
Tradition is the Backbone of the Spinele

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: The Optimum Digital Exposure
« Reply #156 on: November 03, 2014, 11:12:44 am »

Photon-noise statistics are well-predicted by Q.M. Adding noise in post-production to simulate increased photon noise would be straightforward. Several camera brands offer an option to record lossy, compressed, raw data.  The more clever compression algorithms act as noise filters that selectively affect the brightest regions in the image. While irreversably manipulating original data is always a bad idea, in this case most of the information loss (photon noise) can be added back to the image during post production.

Read noise is different matter. But modeling read noise is possible. People who are cursed with making parameter estimated from data with oppressive signal-to-nose ratios (astronomers for instance) are forced to include a model for the noise in their parameter-estimation calculations. It is common to include several terms to the properly model the noise. Including a model for the noise reduces the uncertainty in the signal parameter estimates of interest. Sometimes adding noise computed from candidate models to high SNR data is an efficient way to empirically refine the noise model.
There seems to be a lot of not seeing the wood for the trees going on here and with Guillermo.
Rather than faffing around in post trying to simulate an effect, why not just do it? Sometimes that is the better, simpler and easier solution.
Logged
Tradition is the Backbone of the Spinele

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re:
« Reply #157 on: November 03, 2014, 12:01:47 pm »

I never said it is not simpler or easier, of course it is. I simply said camera digital noise is ugly, and if you like it is only because of a lack of good taste.

Telecaster

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3686
Re: The Optimum Digital Exposure
« Reply #158 on: November 03, 2014, 04:33:39 pm »

Consider the following example… (see above for full post)

It's all about context. As I'm pic-taking only for my own enjoyment your example has no relevance. If I were a pro, OTOH, I'd be doing everything I could to satisfy my clients…and that would include trying to anticipate their future wants & needs. But I'm not a pro and have zero interest in emulating that approach.

I'm not precious about my own photos. They remind me of experiences I've had & things I've seen but otherwise aren't that important in-and-of themselves. My favorites end up on a wall for awhile, then get replaced by others, which will in turn be replaced, etc. I get far more pleasure out of continually experimenting than from sticking with any particular photographing style, processing approach…or subject matter.

-Dave-
« Last Edit: November 03, 2014, 05:03:56 pm by Telecaster »
Logged

Telecaster

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3686
Re:
« Reply #159 on: November 03, 2014, 04:58:14 pm »

I simply said camera digital noise is ugly, and if you like it is only because of a lack of good taste.

Now there's an example of thoughtful humility. ::)

-Dave-
Logged
Pages: 1 ... 6 7 [8] 9 10 ... 13   Go Up