[falconeye]

(And on the subject of fine glass I have a Leica Noctilux 50mm f/0.95 on the way)

wow

[extravagrant]

That wasn't my point.
My point was that you shift ISO down to 1/10th if you discuss things with 1600 wide images. So, your ISO51200 tutorial becomes a ISO6400 tutorial in fact. Go for it if you like but you should know about.

I just want to make sure I understand this.

Are you saying that when the file size is reduced, each pixel in the final image has (4928/1600)^2 ~ 9.5 pixels worth of information from the original image? This makes sense to me, I just want to confirm.

[falconeye]

I just want to make sure I understand this.

Are you saying that when the file size is reduced, each pixel in the final image has (4928/1600)^2 ~ 9.5 pixels worth of information from the original image? This makes sense to me, I just want to confirm.

Yes. Easiest to see this way: if the original pixel received 10 photons, than the resampled pixel received 95 photons. The same what would happen if you reduced ISO from 9500 to 1000.

[Jonson PL]

At first, they looked pretty identical. But looking closer, the edge definition is better in the first one. And so is the colors.

[pscl57]

There is a tiny difference in IQ.
But a big difference of sensibility (exif data).

It is extraordinary

[mikemike]

Drink the contents of the bottle and the picture won't matter.

[JohnBee]

Samples updated,
Full size images added.

[jpzk]

Agreed. The K5 produces excellent straight images out of the camera that don't require any NR manipulation. Images up to 6400, if shot and exposed well, don't require much of anything. Even images at 12800 and some 25600 images often just need a quick pass of light NR, which only take a minute or two. And yes, at 51200, you probably need to put some time into the PP to get a decent image noise-wise (John, please chime in and correct me where/if I am wrong). However, my understanding from speaking with John (who owns a D700) is that from a noise perspective, this is no different than the FF (albeit from FF sensors that are 1-2 years old at this point).

This is all very comforting for someone who does a lot of wildlife/bird photography in often not-so-good lighting conditions.
Even with underexposed images, done inadvertently, we can now assume that it is possible to get decent results at much higher ISO's than with the K20D/K7.

On a cloud covered day, even with some mist and/or rain, especially late during the day, the available light isn't very good. Unfortunately, these are times when one can actually encounter lots of wildlife, especially water birds near/on shores of rivers and lakes.
Same goes for early mornings with a misty atmosphere and low available light.
Contrasts are just not there and I usually have to crank up the ISO way beyond what the K20D/K7 can properly handle, keeping in mind that a reasonably fast shutter speed is near always needed.

(I can imagine what this would do to the K5 AF as well, but that's another story).

This is like a dream come true, I just need to get my own K5 .... sooner than later.

Thanks for all of the images shared.

JP

[dgaies]

This is all very comforting for someone who does a lot of wildlife/bird photography in often not-so-good lighting conditions.
Even with underexposed images, done inadvertently, we can now assume that it is possible to get decent results at much higher ISO's than with the K20D/K7.

On a cloud covered day, even with some mist and/or rain, especially late during the day, the available light isn't very good. Unfortunately, these are times when one can actually encounter lots of wildlife, especially water birds near/on shores of rivers and lakes.
Same goes for early mornings with a misty atmosphere and low available light.
Contrasts are just not there and I usually have to crank up the ISO way beyond what the K20D/K7 can properly handle, keeping in mind that a reasonably fast shutter speed is near always needed.

(I can imagine what this would do to the K5 AF as well, but that's another story).

This is like a dream come true, I just need to get my own K5 .... sooner than later.

Thanks for all of the images shared.

JP

Yeah, it is nice to be able to push about 2-3 stops higher than the K7 and get similar (or better) results.

Also, keep in mind that while 6400-12800 is quite usuable with a minimal amount of effort, it's still not going to match the IQ at lower ISO levels. I suppose that goes without saying, but I just want to make sure I was clear that while I am really impressed with the IQ at 6400, it's not to say it's as good as the IQ at 100-400.

[jpzk]

Yeah, it is nice to be able to push about 2-3 stops higher than the K7 and get similar (or better) results.

Also, keep in mind that while 6400-12800 is quite usuable with a minimal amount of effort, it's still not going to match the IQ at lower ISO levels. I suppose that goes without saying, but I just want to make sure I was clear that while I am really impressed with the IQ at 6400, it's not to say it's as good as the IQ at 100-400.

Of course, I always want to use the lowest ISO possible.
There is no way the IQ at ISO 3200 can be a good at the IQ at 100-400 as you mention.
But it is not always possible and the gain in 2 to 3 stops can make the difference between a souvenir shot and a keeper shot.

Thanks for the reply.

JP

[vancmann]

All I can saw is WOW! Whatever the differences you can find, it is expected and expectations far exceeded (atleast for me). The fact remains that this High ISO sample retains enough info/detail to be usable for many situations. Great team you make with the K5 John!

[JohnBee]

There is no way the IQ at ISO 3200 can be a good at the IQ at 100-400 as you mention.

Have you seen these?

SCENE_1: ISO200 JPG EXTRAC

SCENE_1: ISO3200 JPG EXTRAC

Personally, the ISO3200 is slightly undereposed in contrast to the ISO200 sample. However, I'm willing to bet that if they were both shot exactly the same, we would be hard pressed to differentiate them.

PS. both samples are with no NR/PP etc.

Have you seen these?

SCENE_1: ISO200 JPG EXTRAC

SCENE_1: ISO3200 JPG EXTRAC

Personally, the ISO3200 is slightly undereposed in contrast to the ISO200 sample. However, I'm willing to bet that if they were both shot exactly the same, we would be hard pressed to differentiate them.

PS. both samples are with no NR/PP etc.

Looking at the dark areas of the photos, it is completely obvious which is which. Also I notice that in almost every photo posted that the color of the subject really hides the noise. Just an observation.

[secateurs]

I'm still unsure why you are using a photo with mainly red's and black's for a tuturial.... Why not have a shot with a balance of colours?

[JohnBee]

Looking at the dark areas of the photos, it is completely obvious which is which. Also I notice that in almost every photo posted that the color of the subject really hides the noise. Just an observation.

I would think so, given the ISO3200 doesn't have any NR or PP
But how about that detail?
A little NR and the high ISO version will be looking mighty nice.

Also(off the top), has anyone happened to see a D7K ISO3200 NEF yet?
I think we have something to be proud of