How Do I Measure Noise Using Photoshop? And What Is A Significant Change?

Started 2 months ago | Discussions thread
bclaff Veteran Member • Posts: 9,004
Re: Science To The Rescue Again
1

MrBrightSide wrote:

That is a great point about raw vs. processed; I would love to be able to do what you do and look at the issue from the larger, global perspective.

But unlike you, I'm old and don't have that kind of time left, so I need to focus on real world results from my particular cameras in the places I normally shoot and with the software I use to process my shots.

bclaff wrote:

MrBrightSide wrote:

You are a lifesaver. I shoot in problematic light quite a bit and the hard part for me is figuring out whether I'm making intelligent tradeoffs in shutter/aperture/light temp/ISO/gear. Being able to shoot a ColorChecker and compare the deviation between squares will go a long way to answering the questions I run into.

Then the question turns back to how many standard deviations is significant? Seems like we humans use 1 as the yardstick and that will do until experience proves otherwise.

knickerhawk wrote:

MrBrightSide wrote:

Question 1: Long ago I remember reading that you could use the statistics tools in Photoshop to measure the noise in a photograph. There must be something in the program because it has noise reduction so it must be able to identify noise. How does it work?

There is no direct, objective/statistical measurement of noise in PS. However, there is an indirect objective measurement that provides some useful insight into noise. ....

The answer from knickerhawk is excellent.

The next step is to convert those standard deviations to a Signal to Noise Ratio (SNR) by dividing by the average signal.

If the SNR is above a certain value then you'll find the result acceptable, otherwise you will not. Once the SNR is acceptable higher won't make much if any difference,
The threshold at which you find results acceptable will vary on image size, viewing distance, and personal taste.

That's about as scientific as you can get with non-raw image data because the tone curve can be used to arbitrarily change SNR.

My suggestion is that if you shoot raw or raw+jpeg then examining the raw files (with something like RawDigger) is the real "scientific" approach, otherwise the tone curve hopelessly muddles the results.

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow