Pls show examples where increasing brightness in post better than increasing ISO.

Started Oct 1, 2013 | Discussions thread
gollywop Veteran Member • Posts: 8,284
Re: The meaning of ISO, NEX-7 vs. EOS 7D

ProfHankD wrote:

wchutt wrote:

ProfHankD wrote:

ISO film speed is a bit of a judgement call for sensors used in digital cameras. Basically, the notion of ISO change is implemented by two things:

  • Change of the exposure settings (aperture and shutter speed)
  • Change of the sensor gain

It is physically impossible to increase or decrease the sensor gain.

All sensors operate at a single gain value. Any change in gain (electronic or digital) occurs after the shutter closes.

Yes, gain is applied after the integration period ends (with or without a shutter); I wasn't referring to something like adjusting the quantum efficiency, which is impractical if not impossible. The sensor gain is the gain (amplification) of the analog sensor outputs as they go into the ADCs.

In principle, there should be no difference between increasing brightness (not exposure) by increasing (ISO) using electronic amplification or digital multiplication during post-processing. However the actual results will be highly dependent on the specific camera being used. Besides the convenience of viewing results immediately, the primary reason to use in-camera gain is to get the most out the analog-to-digital converter. Some cameras have ADCs that require no electronic gain to accommodate the ADC technology. The Nikon D7000 and Phase One digital backs are just two examples.

That's what I believe this thread is about. The "ISO-less" sensors would then be the ones where the analog amplification going into the ADCs is essentially a constant. Then again, the raw files do have different values for different ISOs given the same exposure settings, so something is causing that. In addition, Sony has said some things in the past about dynamically adjusting analog values on sensor to reduce noise -- perhaps compensating for differences between ADCs by incorporating some type of reference level check?

That brings up the question of why postprocessing applying the same digital gain (tone mapping) that is done in-camera would do better than the presumably well-tuned implementation in the camera. It could just be it isn't so well tuned in some cameras.

Because that well-tuned camera processor (1) may or may not be that well-tuned, (2) results in an 8-bit jpeg with either sRGB or Adobe RGB gamut rather than a 16-bit psd (or otherwise) with a choice of a wider gamut, and (3) and has an in-camera ISO applied to it that may have been too high, resulting in undesired clipped highlights that you were unaware of at the time of shooting.

-- hide signature --


Post (hide subjects) Posted by
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow