[GoldenWreckedAngle]

1) 1/500 Synch but 1/250 with ISO down to 50 would be a livable alternative

2) Zero noise at ISO 1600

3) Faster auto focus

Revision to #3 - I would rather have gradient mapping on the sensor where the top half of the sensor could be set to a different ISO than the bottom half of the sensor. I could tame a bright sky and bring up the foreground exposure without having to carry graduated filters for all of my lenses.

I can think of a lot of uses for such technology. If you could do it radially where the center of the sensor could be set for a lower sensitivity than the corners you could compensate for lens vignetting. It would be a handy way to radically increase the highlight and shadow details captured in any frame where they surpass the dynamic range of the sensor.

[falconeye]

the top half of the sensor could be set to a different ISO than the bottom half of the sensor

Sounds cool. But isn't.

Because ISO doesn't do much else than amplifiying the available signal, you can achieve an almost identical effect in RAW postprocessing w/o lack of quality or DR compared to your proposal. I know, it was a joke. But anyway ...

[Erik]

Sounds cool. But isn't.

Because ISO doesn't do much else than amplifiying the available signal, you can achieve an almost identical effect in RAW postprocessing w/o lack of quality or DR compared to your proposal. I know, it was a joke. But anyway ...

That's not true. Setting the ISO on the sensor amplifies the analog signal before the ADC stage. If you do the same thing in RAW PP you are amplifying the digital information including all the quantization errors. At the analog stage you have an infinite amount of bits, in a RAW file you have 12, or 15, or whatever. If you take a picture at ISO 200 and amplify it to 1600 in PP you are using only a couple of bits per channel. It will look like junk compared to native 1600.

This is why ISO 3200 is so much worse than 1600 on the 6 MP Pentax bodies -- it is just 1600 digitally pushed a stop in firmware.

[GoldenWreckedAngle]

... I know, it was a joke. But anyway ...

No, actually I'm quite serious but I'm just the idea man.

[Cambo]

Revision to #3 - I would rather have gradient mapping on the sensor where the top half of the sensor could be set to a different ISO than the bottom half of the sensor. I could tame a bright sky and bring up the foreground exposure without having to carry graduated filters for all of my lenses.

I can think of a lot of uses for such technology. If you could do it radially where the center of the sensor could be set for a lower sensitivity than the corners you could compensate for lens vignetting. It would be a handy way to radically increase the highlight and shadow details captured in any frame where they surpass the dynamic range of the sensor.

that'll probably happen AFTER my suggestion of a digital hyper-focal scale in the viewfinder, and a built in digital level - in other words, NEVER!

Too bad - a digital neutral density filter would be just amazing, especially if ytou could chooe the shape or even have it pick an outline for you based on contrast...we'll have to keep doing it manually in Photoshop or some time to come.

Great idea, though.

Cheers,
Cameron

[GoldenWreckedAngle]

... you can achieve an almost identical effect in RAW postprocessing ...

Sure, with expensive software, a couple of bracketed exposures and a bit of time and know how. I can also achieve a similar effect with gradient filters with a lot less effort and expense. It would sure be nice to side step all of that though wouldn't it?

[Kunzite]

that'll probably happen AFTER my suggestion of a digital hyper-focal scale in the viewfinder, and a built in digital level - in other words, NEVER!

Never say never

[JFMichaud]

In reply to that thread

Oups i have to add the words....

Pentax K-7

there you go

[GoldenWreckedAngle]

... if ytou could chooe the shape or even have it pick an outline for you based on contrast...

Now you are thinking! It will happen, mark my words. I don't care how impossible it is today, someone will figure it out and implement it and then every camera made will have it. It will be called something like "Automatic Dynamic Range Compression." I don't know why I give these flashes of brilliance away for free.

[GoldenWreckedAngle]

Hey, this rounds on me. Take it out of my patent royalties.

[jeffkrol]

That's not true. Setting the ISO on the sensor amplifies the analog signal before the ADC stage. If you do the same thing in RAW PP you are amplifying the digital information including all the quantization errors. At the analog stage you have an infinite amount of bits, in a RAW file you have 12, or 15, or whatever. If you take a picture at ISO 200 and amplify it to 1600 in PP you are using only a couple of bits per channel. It will look like junk compared to native 1600.

This is why ISO 3200 is so much worse than 1600 on the 6 MP Pentax bodies -- it is just 1600 digitally pushed a stop in firmware.

Actually if you convert at 16 bit it is the same. ADC is at most 10-12bit anyways....
Underexposing and "push processing" is just as effective as in camera boost for a good part of the image.
Bottom line: High exposure zones and/or high ISO, where photon noise and pre-amplification read noise dominate the noise, are rather insensitive to what ISO is chosen once a choice of exposure is selected and care is taken not to clip highlights. Underexposing by a stop, and doubling the raw values in post-processing (that is, applying exposure compensation), yields the same image quality as 'proper' exposure under these conditions. On the other hand, in lower exposure zones at low ISO, where post-amplification read noise becomes important, the read noise goes down by a bit less than a factor of two (in electrons) when the ISO doubles. In this situation, underexposing by a stop and doubling the raw values in post-processing, yields more noise than proper exposure, particularly in shadows.
http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html#ETTR

[GoldenWreckedAngle]

Actually if you convert at 16 bit it is the same. ADC is at most 10-12bit anyways....
Underexposing and "push processing" is just as effective as in camera boost.
Will get you proof later.

I've been experimenting a bit with this in LAB mode and I'm inclined to agree with you, but again, in camera would be the bomb. ACR and Photoshop are pricey and time consuming accessories.

Come to think of it, there was a sudden scramble to redesign something about the K-7 around the time frame of post #208 in the dear Pentax thread... I wonder...

There, bought my round and added my rumor for hitting a tenth post - Who's next?

[nanok]

i think i have seen this implemented on some makers camera. now to remember where i saw it.. (might be panasonic g1h, but not sure). i think it's a great idea, but i doubt of how much use it is with todays technology (it sounds more like a matter of convenience -- less post processing needed -- than a matter of better image quality). for a sensor capable of significant gain with no loss (a very very smart amp, to get it up to iso800 or so with extremely high quality), and it would be a cmos (so asto enable separate amp adjustment for each photosite), it would make some sense. oh, wait, i am almost talking about the k20d sensor here.

/me walks away scratching head..

[nanok]

about price of software: for landscape shooters, who shoot form a tripod and can use braketing (no moving subjects across the frame), i recommend having a look at enfuse. it is _exactly_ what you (we) need. it's not hdr, it's designed by somebody who thinks like a photographer (like an ND grad filter user photographer).

here is just one example (there are muchmore impressive ones on their wiki)

web-adamov-202503-505

[GoldenWreckedAngle]

Thanks for the tip. I'm checking into it now. Looks promising.