I wrote the Hand Held Helper Script (H3S) for a specific use case: that is a time limited bracketing situation, further constrained through hand holding needs.
One dimension to the constraint field is the shutter speed, that is the lowest hand held you can get away with. As this will vary with FL and the individual’s skills, it has a wide scope, however, to illustrate things, let’s assume we are hand holding with a WA lens, and our lowest shutter is sensibly placed at, say, 1/30s.
For most scenes this means that if we wish to cover the highlights and the shadows, so we will likely be time bracketing between 1/30s and, say, 1/1000s, ie 5 stops.
If we were shooting with a fully ISO invariant camera, ie invariance down to ISO 100, then this time bracketing would likely be sufficient to cover us, ie we (hopefully) could ‘recover’ shadow exposure in post, and then post process the brackets, eg either through tone mapping, fusion blending or masking.
However, as we’re shooting with a Canon system, the ISO invariance advantage doesn’t really kick in until we are at, say, ISO 1600-3200. Thus we need to be a bit more respectful at managing our brackets, ie trying to maximise the signal (number of photons) and minimise the noise (made up of photon statistical noise and ‘electronic’ noise).
Therefore, the second bracketing dimension that we need to utilise is the ISO, say, 100 to (say) 3200 or 6400 (on a full frame camera with ‘large’ photo sensors, eg the 5D3).
Thus the full bracketing scope that we have when hand holding looks like this:
That is some 10 stops of coverage.
Of course, much of the time we don’t need access to the above Ev space, as we are intelligent enough as photographers to know when to limit our bracketing needs, eg through metering or simply our (visual) skill/experience.
Such Scene-Informed-Bracketing (SIB) allows us to then choose the bracketing scheme we wish to adopt, eg ML-Advanced-Bracketing, Dual-ISO or ‘simple’ Canon.
But what if you don’t have the luxury, or ability, to carry out a SIB capture? For example, you are time limited and/or hand holding: that is you are Scene-Agnostic-Bracketing (SAB). This is why I wrote the H3 Script; which will capture a full range of brackets, irrespective of the scene’s DR; thus allowing you to choose, in post, what brackets you take forward to final post processing, ie one through to six!
The following illustrates the coverage that H3S provides, shown relative to an illustrative Dual-ISO bracket at 100/800:
The above, however, is an apples and pears comparison, as the Dual-ISO requires some SIB awareness. That is the Dual-ISO needs to be positioned to maximise the chance of covering the scene of interest. The H3S backets are SAB based.
Note the above only illustrates one potential Dual-ISO capture, the two red squares can be positioned ‘anywhere’ in the exposure range. Of course, as we know, using a single Dual-ISO not only limits your Ev coverage, but it also slightly compromises your resolution. But Dual-ISO remains a very valid choice for a SIB capture.
To fully maximise the H3S value, I also included the user option of capturing ‘high ISO (noise) brackets’ if you can’t (sic) capture time brackets, eg you are limited from a shutter perspective. The following shows such an H3S bracket example – once again reference to the Dual-ISO example:
So what’s the bottom line?
With ML we always have choices. If you have the luxury of assessing the scene, ie its dynamic range, that is able to carry out a SIB capture, and able to fully use the shutter, eg be on a tripod, then ML Advanced Bracketing is likely to be the way to go.
If you are handholding and/or need to manage scene motion, then Dual-ISO will likely be your friend. But you will still need to approach the capture from a SIB perspective.
If, however, you don’t know what the scene’s DR is, or haven’t the time to spend assessing the scene and/or you are hand holding, then using the SAB-based H3 Script may be the best way to ensure you have ‘all’ the images you need for post processing, ie to choose from.
It's your choice.
No comments:
Post a Comment