get_live_histo() strangeness

pigeonhill

1164

« on: 06 / April / 2023, 11:43:43 »

Advertisements

@reyalp

I hope you can help me understand something about get_live_histo() use.

I'm close to publishing an update to my ETTR meter, but am seeing some strangeness.

I've decided to use the top three bins of the histogram, after calling histo,total=get_live_histo()

That is histo[253], histo[254] and histo[255]

What I see, as I move the exposure to being overexposed, is what looks like real data in histo[253], but histo[254 or 255] never get any meaningful data, ie just low level counts. The top cell thus appears to be histo[253].

Do you have any insight ;-)

Logged

pigeonhill

1164

Re: get_live_histo() strangeness

« Reply #1 on: 06 / April / 2023, 13:15:46 »

Forgot to say the above behaviour is on an M3.

On the M10, I observe histo[254] is the peak bin, by far, with a little bit more in histo[253] but only 'noise' in histo[255].

Logged

reyalp

14118

Re: get_live_histo() strangeness

« Reply #2 on: 06 / April / 2023, 14:30:04 »

Quote from: pigeonhill on 06 / April / 2023, 11:43:43

What I see, as I move the exposure to being overexposed, is what looks like real data in histo[253], but histo[254 or 255] never get any meaningful data, ie just low level counts. The top cell thus appears to be histo[253].

Do you have any insight ;-)

How the viewport values are derived is unspecified (outside of Canon, at least). It is not unusual for the extremes (near 0 and 255) to have few or no counts regardless of exposure. As discussed previously (https://chdk.setepontos.com/index.php?topic=14342.msg146177#msg146177 for example) some cameras have a much larger dead zone. It could also vary depending on settings or output device, although I don't think that's been confirmed.

Logged

Don't forget what the H stands for.

pigeonhill

1164

Re: get_live_histo() strangeness

« Reply #3 on: 06 / April / 2023, 15:03:18 »

Thanks for the reminder.

I’ll adjust my algorithm ;-)

Logged