Perhaps this thread should be laid to rest but I'm uneasy because I don't see that the case has been made yet for including RAW mode in the G7.
I don't think anyone here is arguing that RAW mode does not provide an opportunity for
some degree of improvement. At least I'm not.
To recap, is there any good reason to suppose that RAW support in a P&S camera can provide a greater percentage improvement than can be achieved in a DSLR with its significantly greater dynamic range?
Both Jonathan Wienke and John Sheehy initially implied that it can, claiming a whole stop of dynamic range improvement.
I questioned this. Jonathan did some comparisons, then claimed a 2/3rds stop DR improvement for his Olympus SP-350. Unfortunately, he made the comparison with in-camera contrast settings at zero on a scale of -5 to +5. If those contrast setting on the Olympus are not just for show, it's almost certain that the DR improvement will be
less than 2/3rds of a stop, perhaps much less.
If it's not much less, then the question still arises. Why the greater percentage improvement, than with a DSLR? If we make an assumption that the DR of a 5D is 7 stops, then the DR improvement using RAW converted in ACR, according to my own experiments, is around 10% (ie. around 0.7 stops. Maybe 0.8 stops, but less than 1 stop).
If we make an assumption that the DR of the G7 is 3 stops (
by the same standards we have used to describe the DR of the 5D as being 7 stops) then we could expect a DR improvement in the G7, using a RAW image, of 3/7ths of 0.8 stops = 0.34 stops, close enough to 1/3rd of a stop.
If this is true, then Jonathan has overstated his case, as indeed Michael might also have in this quote from his latest article - "It's not the Camera, it's the Photographer - Right?"
In this case, adequate for the job. But, not ideal, The Canon G7 doesn't shoot raw, and therefore only 8 bit JPGs. The JPG part wasn't that big an issue in this case, but the 8 bit file was found to be a real problem when I tried to enhance the separation between the rays of light and the shadows. It kept posterizing. (If anyone tells you that in-camera jpgs are sufficient, tell them to go to the back of the class. That's where they'll find the Canon marketing and engineering people who decided that this fine camera didn't need raw).
Now I'm not sure what bit depth is used in the G7, but it's probably less than the 12 bits used in Canon DSLRs and I can't help wondering if Michael is making the same mistake of attributing the 'magic bullet' cure to all RAW processing.
In other words, because a certain degree or level of posterization, visible on a high quality monitor, can be completely fixed by shooting in RAW mode with a DSLR and converting from 12 bit to 16 bits, it does not follow that the same degree of posterization from a tiny P&S camera could be completley fixed if such a camera also offered a RAW mode.
It's reasonable to suppose there would be
some improvement but perhaps not enough to fix the problem. One could also argue that any improvement, however small, is better than no improvement. That's true. But how much time, effort and expense is that small improvement worth, both from the perspective of the end user as well as the manufacturer? That's the real question.
Canon have said essentially, it's not worth bothering with. Jonathan has said it is, but seems to have been under a misapprehension all along that he's been getting a whole stop of DR improvement using RAW mode with his SP-350 (which has a 10 second delay between shots.)
He's also tried to demonstrate the superior WB results from RAW, but again it seems the RAW image is only very marginally superior when one works on the jpeg image.
As I said, the issue does not seem resolved yet.