• Members 546 posts
    June 3, 2023, 1:47 a.m.

    Is there a way to see how much noise will be in your photo before you take your shot?

    I know that I can see the effects of changing your white balance ahead of time by putting your camera in Live View, hitting your Q button and changing your WB. You can see the effect before you take your shot.

    If I put my camera in Live View and start IncreasIng my ISO, will I hit a point where I start seeing visible noise in my photo before I actually take the picture?

    Steve Thomas

  • Members 2111 posts
    June 3, 2023, 2:12 a.m.

    The only way I know of is in Live View after zooming in and with exposure simulation turned on. But I assume any noise visible in Live View would be after the camera applied any noise reduction according to your camera settings at the time.

  • Members 83 posts
    June 3, 2023, 10:59 a.m.

    While I do not think that you can predict a particular signal to noise number in advance, there are some things about noise you can predict.

    In my Canon cameras, noise increases as the ISO number increases and noise increases as the sensor temperature increases.

    If a larger ISO number is needed to get the shot, then one can predict more noise. If the sensor is warmer because it is hot weather or because it has been continuously active producing live view images, then one can predict more noise.

    You might search for noise in the PDF manual for your camera. For example, in the EOS R5 manual there are several "Note" or "Caution" sections that mention increased noise from setting choices.

    While this does not seem to happen on my EOS 80D, on my EOS R5 when using auto exposure the camera menu settings seem to me to cause a difference in noise. (Others tell me this is impossible). Since I have the EOS R5 save a raw CR3 file, I can do peripheral illumination correction later using DPP instead of when the photo is made. It seems to me that disabling highlight tone priority, peripheral illumination, and auto lighting optimizer in the camera menus will sometimes result in a less noisy image.

    Brighter areas of the image have more noise, but also much larger signal so the signal to noise ratio is better in brighter parts of the image. If the shadow areas are not important, one may reduce visible noise later by darkening the shadows or just tolerate more noise in the shadows. Which visible noise is objectionable and which noise may be ignored and which noise makes the photo look better is a matter of personal taste. I have found this helpful: www.cambridgeincolour.com/tutorials/image-noise.htm

  • Members 468 posts
    June 3, 2023, 11:57 a.m.

    "Live view" usually uses an exposure time that is useful for the live view video feed and is not the exposure time for stills, and often uses a lens wide open even if it is going to stop down for the capture.

    So, if you want something to look proportional, noise-wise, the camera would actually have to do calculations and simulate some kind of noise that it puts in an overlay.

    When shooting video, then the viewing exposure parameters are the same as what is recorded, and is therefore more proportional, but video hides noise very well, and if you want to rip stills from video, they can easily look noisier than they did in rapid sequence. Some cameras have pre-button-press buffered recoding that uses video for stills bursts and those should look proportional. However, the frames are not magnified so noise is going to be harder to see in the EVF than in the stills highly-magnified on a monitor. Again, some kind of overlay with blinking pixels below a certain absolute signal level or other indicator might still be necessary for noise feedback in the EVF.

  • Members 468 posts
    June 3, 2023, 12:18 p.m.

    I, @JohnSheehyRev have written:

    Another option might be to have a raw histogram that is not showing the levels for the given ISO setting, but for the actual number of electron charges expected. IOW, even though you might be set to ISO 12800, you'd see a raw histogram that would be appropriate for base ISO. The histogram could be designed to blink for values below a certain level to get your attention and warn you.

  • Members 468 posts
    June 3, 2023, 12:56 p.m.

    All these simulations are about lightness, though; not absolute exposure. The exposure index of the video driving the EVF is not the exposure index of the capture, unless you're grabbing video frames, and you can't expect the video feed's absolute exposure to be the capture's absolute exposure.

    If you're destined to get a high exposure index because you're using a fast shutter speed like 1/4000 for an active subject, the video feed could be using 1/60 or 1/120.

  • Members 716 posts
    June 3, 2023, 1:05 p.m.

    Certainly is. Understand your camera. Then take a look at the ISO you're using. Mirrorless makes it too easy -just observe the crushed up blacks in the EVF. You're going to get noise. For sure

  • Members 468 posts
    June 3, 2023, 2:28 p.m.

    It's roughly true that you can consider ISO setting combined with an "exposure" simulation (lightness simulation) in the EVF to gauge noise, but different WB needs can hide some very, very weak exposure. Anyone who shoots a lot in green forest shade will know that "ISO 12800 in the shade" is very different from ISO 12800 in the sun or under open overcast skies, and even more different from ISO 12800 with a magenta light source.

  • Members 716 posts
    June 3, 2023, 11:15 p.m.

    For sure. But my very first point, was to understand your camera. That means taking photographs. Lots and lots and lots of photographs, at various settings, in an intelligent, logical manner, to get an understanding of how your particular camera responds. Digital is fabulous for this, because the feedback is instant -although to get a better idea, a computer monitor will give you a more precise indication.
    Which brings us to the next thing. Viewing size, and display medium. These make very large differences to how acceptable, or otherwise, noise is in an image. What or who is your intended audience?
    And also, how well your raw converter deals with noise reduction. I don't think there's any nice, near little package that you can pop all this in, and get a definitive result or idea one way or the other. But that's just me

  • Members 45 posts
    June 5, 2023, 6:32 a.m.

    Yep. The same way we determined exposure latitude with film - test your camera, take notes, learn your cameras behavior at different ISO settings. After a few dozen test shots in different lighting conditions, at different ISO settings, you’ll get a grip on how high you can set your ISO and get acceptable results to you.