Joseph Marney wrote:
I'm attaching a crop that has been significantly downsized. The left hand side is pulled, the right is original exposure.
OK, now I'm even more confused by your example. It looks like you took a very underexposed area of the right image and pushed the shadows by probably more than a couple of stops. Originally I thought you were taking an out-of-camera exposure at the very high ISO and simply looking at it at a very high magnification factor.
If you are pushing the nearly black deep shadow area of the right half image shot at 6400 or higher in order to get the exposure we see in the left half, I think that there are a few take-aways to consider:
1. It is no surprise that the left side looks awful.
2. With such a super-bright area on the right, why would you use such a high ISO?
3. Aren't we really looking more at how a poorly lit and exposed exposure fails, and not so much an indication of the defects of a camera? (In the real world you would almost certainly not make this photograph this way — you might fill the dark area with a reflector or electronic flash... but you'd really probably move the lens to a different location without that extremely bright background on the right side.
This reminds me to the "tests" we used to see in forums and elsewhere that "proved" that some sensor has terrible banding, poor gradients in the shadows, awful noise, and so forth... which were based on making a "photograph" with the lens cap on and then pushing the resulting 0 luminosity value image by 8-10 stops.
This sort of thing can be kind of useful for identifying the edge case failure mode for various cameras — and they all will fail this test in some way. But it isn't predictive of real-world performance.
If you are in the habit of lighting and exposing this way and then doing what you did here in post...
... you probably need something other than a APS-C camera from Fujifilm.
The general rule is that if you are really going to. push the outer boundaries of current camera/sensor technology with extremely high ISOs as a major portion of your work, you are probably better served by a full frame or larger sensor camera with less-than-the-maximum sensor resolution — e.g. a larger sensor where photo site size is larger at the expense of resolution.
Again, in my experience (a lot of which is doing handheld night photography in the urban street environment using Fujifilm cameras), qualty I have no concerns getting quality at ISO 200, 400, and 800. I can get quite excellent results after normal post processing at 1600 ISO, and quite decent results at 3200. I prefer not to go above 3200 for my photography, but I can and do get workable images at ISO 6400, but I virtually never go higher. (My standard for image quality is that at 20" x 30" print should look very good.)
Joseph Marney wrote:
I'm attaching a crop that has been significantly downsized. The left hand side is pulled, the right is original exposure.
OK, now I'm even more confused by your example. It looks like you took a very underexposed area of the right image and pushed the shadows by probably more than a couple of stops. Originally I thought you were taking an out-of-camera exposure at the very high ISO and simply looking at it at a very high magnification factor.
If you are pushing the nearly black deep shadow area of the right half image shot at 6400 or higher in order to get the exposure we see in the left half, I think that there are a few take-aways to consider:
1. It is no surprise that the left side looks awful.
2. With such a super-bright area on the right, why would you use such a high ISO?
3. Aren't we really looking more at how a poorly lit and exposed exposure fails, and not so much an indication of the defects of a camera? (In the real world you would almost certainly not make this photograph this way — you might fill the dark area with a reflector or electronic flash... but you'd really probably move the lens to a different location without that extremely bright background on the right side.
This reminds me to the "tests" we used to see in forums and elsewhere that "proved" that some sensor has terrible banding, poor gradients in the shadows, awful noise, and so forth... which were based on making a "photograph" with the lens cap on and then pushing the resulting 0 luminosity value image by 8-10 stops.
This sort of thing can be kind of useful for identifying the edge case failure mode for various cameras — and they all will fail this test in some way. But it isn't predictive of real-world performance.
If you are in the habit of lighting and exposing this way and then doing what you did here in post...
... you probably need something other than a APS-C camera from Fujifilm.
The general rule is that if you are really going to. push the outer boundaries of current camera/sensor technology with extremely high ISOs as a major portion of your work, you are probably better served by a full frame or larger sensor camera with less-than-the-maximum sensor resolution — e.g. a larger sensor where photo site size is larger at the expense of resolution.
Again, in my experience (a lot of which is doing handheld night photography in the urban street environment using Fujifilm cameras), you can get quite excellent results after normal post processing at 1600 ISO, and quite decent results at 3200. I can and do get workable images at ISO 6400, but I prefer to stay below that... and virtually never go higher. (My standard for image quality is that at 20" x 30" print should look good.)
Good luck.
Nov 29, 2022 at 04:20 PM
Previous versions of gdanmitchell's message #16106834 « XT5 white specs? »
