Spotmeter inconsistent % across builds?

Started by timbytheriver, December 11, 2019, 06:18:47 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

timbytheriver

This is a weird one.  :o

Across two different builds and modes I get inconsistent readings from the spotmeter in crop/non-crop modes:

1) Vanilla install of 4K raw video recording; lossless compression build from the experiments page, my spotmeter maxes out (on the brightest highlight) at 99% in crop, and 100% in non-crop mode.

2) A recent Danne build I get 99% Max in both crop and non-crop mode.

Is this expected behaviour?
5D3 1.1.3
5D2 2.1.2

a1ex

The percentage spotmeter is JPEG-based, so the output depends on how the rendering pipeline is configured (whether the saturation level is rendered as pure white, or not). Picture style settings will affect this. Changing ISO gains will also affect this.

The EV spotmeter is RAW-based, and it depends on how well the white level is detected or hardcoded. This varies with ISO, sometimes also with shutter speeds, and sometimes also with aperture (actually digital gain related to aperture). Also, once you drop the gains below the "harsh clipping" level, this becomes more on of an art than a science. In the iso-research branch, raw_diag uses a different method that attempts to address all of these things, but it's pretty much guesswork and I didn't test it extensively. It's not integrated in the ML core yet.

timbytheriver

5D3 1.1.3
5D2 2.1.2

timbytheriver

The only time I manage to get 100% max spotmeter readout on my 5D3 (obviously pointing at a 'hot' light source) – is in non-crop modes when half-pressing the shutter. Then 100% appears for a moment – and is gone again!

Changing pic styles seems to make no difference on mine. I have a 99% camera...  :(

:P

Anyone have any thoughts on how to retrieve my 1% (no pun intended).
5D3 1.1.3
5D2 2.1.2

Dmytro_ua

Quote from: timbytheriver on December 12, 2019, 03:38:06 PM
Anyone have any thoughts on how to retrieve my 1% (no pun intended).

Thanks to Alex for an explanation because I didn't know this.
I'm still trying to find the best workflow. For now I'm using a percentage spotmeter with a grey card close to a face to get an equal lighting in any situation. I'm not sure if this is correct. Still experimenting.
5d3 1.2.3 | Canon 16-35 4.0L | Canon 50 1.4 | Canon 100mm 2.8 macro
Ronin-S | Feelworld F6 PLUS