ISO invariance understanding it - just checking something

Started 2 months ago | Discussions thread
Brian Kimball
Brian Kimball Regular Member • Posts: 485
Re: ISO invariance understanding it - just checking something

SigZero wrote:

With every increase of ISO (1 EV) You are loosing around 1 EV in highlights. Overexposing to te limits of the sensor (Zebra at 109% is Your friend here) and then pushing exposure down and shadows up (which means leaving them at real exposure level) gives you the best highlight protection and shadow noise performance.

In case od dual gain sensors the key is to overexpose not more than the limit of dual-gain change. So If the limit is at 640 ISO, then overexpose ISO 100 no more than 2,5EV and overexpose ISO 400 no more than 0,5EV - of course if having best shadow noise performance is the key.

You've got this backwards unfortunately. Dual conversion gain tech does not impact one's strategy when they are able to overexpose. More light at base ISO is always better and will result in higher SNR. The advice to "overexpose ISO 100 no more than 2,5EV" is very misleading. For purposes of ETTR, ISO 100 can be overexposed as much as the scene allows, and more will always be better as long as you're still meeting your needs WRT to clipping, DOF, and motion blur.

We've all been taught that when shooting dimly lit scenes with ISO invariant sensors (with no DCG) there's no practical difference in noise performance between shooting at, for example, ISO 1600, vs shooting at ISO 100 + raising the image lightness by 4 stops in post (assuming using the same aperture & shutter speed as you would have at ISO 1600). Why would one want to do this? To protect highlights, if you're worried about them clipping. Why would one NOT want to do this? Dark- to completely-black JPGs and unusable previews on the camera.

WIth DCG this becomes a little more tricky, but not by much. If your DOF & motion blur requirements don't allow changes in aperture & shutter speed, and your camera meter wants ISO 1600, then on the A7M3 you can drop to ISO 640 and eke out about a half stop more read noise performance than if you dropped further down to an ISO between 100-640.

Whether anyone actually cares about a half stop difference in read noise in the shadows is a separate question that individuals will have to decide for themselves.

In case you need to have better shutter speed and need to overexpose more then the limit, bump ISO to second gain settings (640 in Your case) and overexpose accordingly there - You will loose more highlights but get better shadow noise.

I think this is what you meant:

In case you need to have a [faster] shutter speed [that causes you] to [underexpose] more than [2.5 stops], bump ISO to second gain settings (640 in your case) and [expose] accordingly there - you will lose more highlights but get [around a half stop] better shadow noise.

If the situation is the other way - if in order to keep highlights you need to make negative exposure compensation - if you need to cross dual-gain boundary it is better to underexpose more on lower ISO than underexposing on higher and crossing the dual gain level (is such case You will not saturate sensor to its full capacity).

Yes, if ISO 640 and above clip, and you don't want clipping, it is fine to drop below 640 and raise your image lightness in post. Otherwise, might as well stay at 640 or above (assuming your chosen aperture and shutter speed require a high ISO).

Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow