Fuji x10 6MP RAW DR400 EXR mode vs old 6MP DSLR

Started Apr 14, 2014 | Discussions thread
(unknown member) Contributing Member • Posts: 500
Re: EXR versus X-Trans, or X10 vs X20

Trevor G wrote:

cantanima wrote:

Trevor G wrote:

The above is RAW. Silkypix predominantly uses the second, shorter exposure. In fact, except at very low shutter speeds, I cannot see when the first frame is used.

Hence the noise.

The JPEG quality would be even worse, with much more detail smeared away..

Perhaps I didn't express myself well enough.

The shorter exposure (in RAW, not JPG) will inherently be noisier in the shadows, because it has less time to collect information. The longer exposure (in RAW, not JPG) will not be as noisy in the shadows, because it has had more time to collect information.

If Silkypix & Fuji predominantly use the second, shorter exposure, they will produce a noisier picture. I'm saying that, since it's possible to combine the two exposures in a way that relies much more on the longer exposure for the shadows, it's possible to achieve a better result than what you are seeing here.

I don't doubt that the X20 has a better sensor for low light than the X10; I've read quite a lot about recent improvements in sensor technology in that regard. However, taking one low-light shot & improving the S/N ratio using software alchemy, or combining several short, low-light exposures using software alchemy, and combining one short exposure in good light with one longer exposure in good light using software alchemy, are all very, very different things, with lots of options for how to combine, etc.

But that is not what we are looking at.

You would have to apply the same sort of effects to both images. That's why I don't like supplying RAW because unless you handle both the same (noise reduction turned off, and so on) the test becomes meaningless.

Here's what I'd do: extract them as 16-bit TIFFs using dcraw, and adjust the exposures in digikam, primarily to see when noise appears in the darker image. Then I'd export the X10's exposures and combine them results using 2-3 different techniques. I would not perform any noise reduction, would use only free, open-source software, and I'd document it all in detail (including options supplied to command-line tools) so any fool could reproduce it.

If you don't want to participate, here's something else I can try: I'll take two shots of a high-contrast scene in the X10, one with M size ISO 100 DR 400 and one with L size ISO 400 DR 400. I'll then use the technique described above to see which has more information, and/or can be manipulated into a more pleasing image. This would demonstrate more convincingly whether EXR hardware techniques are a gimmick, especially since the same sensor would be used. (As I've acknowledged elsewhere, improvements in sensor technology can mean the X20's sensor may well have a wider dynamic range, which would suggest why its Raw files are 14 bits, to the X10's 12.)

I'll probably do the second either way, assuming I find time. If you have any suggestions on how to improve the test, I'm open to them.

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