[pathdoc]

This is by no means a divine miracle - 800K is unusable - but 51K is pretty good for Facebook fodder, and there have been times in a dark auditorium where the ability to turn out a good result at that ISO would have been really worthwhile to me.

It will be interesting to see what the Raw-->JPG conversion software they ship with this camera can do in terms of improving the result over the in-camera conversion.

[photoptimist]

ISO 819,200 seems like a very welcome advance in performance for three reasons:

Changing definition of "poorly lit": With the earliest digital cameras, anything dimmer than bright shade was poorly lit. Thankfully, the cameras have gotten better and better so that the definition of poorly lit is getting darker and darker. I look forward to hand-held photographs on full moon hikes and around camp fires.

Embrace the grain: I'm amused by the people that think you can't have an "acceptable" photo at higher than ISO X. Sure, I love a sharp, buttery-smooth, noise-free images as much as the next person. But then I go over to some of the film threads and see people doing push-process, lomo, and caffinol images that look like they were printed on gravel but also look excellent. I'm sure we'll see some very high ISO images where the grain helps the image rather than ruins it.

Digital polaroid-back: Some 4x5 users have a polaroid back for taking test shots to check exposure and composition. That's what ultrahigh ISO can be used for. Before doing a long exposure, time-lapse, or laborious bracketing shot at low ISO, these new high ISO settings let one do a quick exposure to carefully review the image for details or issues that simply aren't visible in the viewfinder or live view.

[awscreo]

Will be interesting to see raws. Jpegs are fairly impressive, no doubt.

---------- Post added 01-30-17 at 11:14 AM ----------

I respectively disagree with your opening comment, "why we want to take memories of that". Are you saying all night time pictures shouldn't even be attempted since they are all going to be bad? Maybe you need to learn how to take pictures in dark lighting? Or in situations where shadowed areas are also part of the composition (tunnels, museums, etc.)? There are PLENTY of interesting subjects, night city skylines, fireworks, astronomical, just to name a few, that have a purpose for a camera that can accurately capture the contrast between light and dark. As for flare, I don't have photos that suffer from flare. Maybe its because I use quality coated lens that reduce it, along with knowing how to compose a shot to not induce flare. If you think these example pictures are the sum total of "interesting subjects", or the epitome of nighttime photography, instead of just some samples to demonstrate the resolution, then you maybe missing the point of their use? It illustrates how the gradual degradation of noise starts to occur at a much higher ISO than earlier models, where it begins become unusable at 6400. Instead of saying that night photography has no value, it may be because the results you are getting aren't good? Maybe having a tool like this might add capabilities that would allow satisfactory and creative possibilities. That's all I was saying.

The major benefit of the technology n this camera is that what used to be unusable, noisy shots can now be properly exposed. And that also extends to the opposite side of the lighting spectrum. With an electronic shutter as fast a 1/24,000 sec, you can reduce overexposed areas of shot. If you look at the Pentax 645Z, a MF, it's ability to take high resolution images at higher ISO is it's strength.

Interesting subjects are the photographer's function, not the camera. Having a tool that handles a wider range of situations where less than ideal lighting conditions exist, seems to me to increase the number of memories that can be captured.

Interesting, I thought mf was lagging slightly behind of ff in terms of high iso performance? At least from tests I've seen (granted I haven't looked into it in detail).

[Quicksand]

But then I go over to some of the film threads and see people doing push-process, lomo, and caffinol images that look like they were printed on gravel but also look excellent. I'm sure we'll see some very high ISO images where the grain helps the image rather than ruins it.

Oh, I don't mind grain. But those high-ISO large chroma blotches are really off-putting, even in web-size versions.

[eddie1960]

well to me 819200 looks about the same as 3200 did on my *ist ds , maybe even a little better. certainly what i see here has very usable for a lot of scenarios iso 51200. Very usable as in with the right pp you could make decent prints ate a reasonably large size never mind perfect for web use even without doing any serious pp. I'm always gobsmnacked how far it's come. and using the ultra high for compositional purposees and determining long exposures before commiting the time and battery to the final shot is a huge benefit , particularly for guys shooting mid winter at night....

---------- Post added 30th Jan 2017 at 12:49 ----------

That's one of the situations of the sunny f16 rule. In short, the f16 rule tells you that for the most common still shooting situations, you don't need more than ISO1600, even with a Pentax long lens (f5.6 / f8). All current apsc cameras can cover this range. You need at least one full stop ISO improvement when shooting moving subjects with >200mm long lenses, that where full frame helps a lot.

sunny 16 is fine on days where it is sunny, or just in daylight in general. if all your shooting is in beutiful sunny most of the year near the equator areas and during daylight hours that is great. anyone shooting events will appreciaqte an extra stop or better performance at the iso they consider emergency only level. astro photography has also been mentioned in detail. ....sunny sixteen worked when i shot with a spotmatic mostly in daylight , digital has changed my expectation

[majortom67]

Just saw damples on the German site. At least at 6400 my Fuji X-Pro 2 is far more better, both on noise and details.
Marketing.

---------- Post added 01-30-17 at 01:14 PM ----------

As already posted elsewhere:
---
Just compared samples images of the German site with my Fuji X-Pro 2. At least at 6400 I have far better noise and details. I just envy the 5 axis stab.
Therefore that's Marketing

Sorry for the duplicate answer.

[eddie1960]

Just saw damples on the German site. At least at 6400 my Fuji X-Pro 2 is far more better, both on noise and details.
Marketing.

---------- Post added 01-30-17 at 01:14 PM ----------

As already posted elsewhere:
---
Just compared samples images of the German site with my Fuji X-Pro 2. At least at 6400 I have far better noise and details. I just envy the 5 axis stab.
Therefore that's Marketing

Sorry for the duplicate answer.

I shoot Fuji and they are noted for exaggeration in their ISO rating of about 1 stop, so compare the 6400 to 12800 on the xt2 -can the xt2 shoot 12800 raw, the XT1 cannot- (I've not done a measured analysis because i lack the tools but based on experience with several brands i'd say that yes Fuji claims about a stop more than real. Even on the Fuji forums this is pretty much accepted

[Simen1]

Since much of ISO improvements lately are software improvements (in camera jpeg converter), I would like to see side by side photos from KP and K-5 processed from raw in the same modern software, in stead of in camera jpegs.

[shardulm]

This is ridiculously controlled noise.. No skittles? That is no fun.

[biz-engineer]

sunny 16 is fine on days where it is sunny

What's the f4 in the Sunny f16 rule table? If I understood correctly f16 is , yes , for sunny day, and it says, if cloudy shadow, open up the lens to f4.

[MarkJerling]

Eh, it's easy to hide noise at web sizes.

You can look at the original file size then: http://www.ricoh-imaging.de/media/fffbbfbdc4c531488177eedfbd4b02b9/IMGP0638_ISO6400.jpg

[stevebrot]

Embrace the grain: I'm amused by the people that think you can't have an "acceptable" photo at higher than ISO X. Sure, I love a sharp, buttery-smooth, noise-free images as much as the next person. But then I go over to some of the film threads and see people doing push-process, lomo, and caffinol images that look like they were printed on gravel but also look excellent. I'm sure we'll see some very high ISO images where the grain helps the image rather than ruins it.

You make an interesting point. There is a place for "rough" interpretation of a subject.

There is also a very important difference between digital noise and film grain. The former is composed of pixels that have no data. As noise increases, detail, tonality, and color fidelity disappear and you can't mitigate loss due to noise. Film grain, OTOH, is intrinsic to image formation for most emulsions and is not distinct from "data". Paradoxically, minimization of grain is not a sure recipe for better fidelity. Fine grain tends to defeat definition despite improving resolution.* That may explain the continued popularity of the Rodinal developer formula.**

Digital polaroid-back:

Indeed! Live view in the dark is nasty work. I have done as you state on several occasions where I needed fairly precise framing but had very little light.


Steve

* The nature of grain as it relates to image formation and fidelity is a complex subject and what I wrote above is overly simplistic. It is usually enough to simply state that grain is "supposed to be there", despite the fact that it may be obtrusive.

** Rodinal is renowned for promoting excellent acutance (definition and edge sharpness) with slow and medium speed film, though at the expense of more prominent grain.

[stevebrot]

What's the f4 in the Sunny f16 rule table? If I understood correctly f16 is , yes , for sunny day, and it says, if cloudy shadow, open up the lens to f4.

Aha!

The sunny 16 rule is somewhat distinct from the tables (roughly based on EV100 estimates for common lighting situations) that usually accompany articles on "the rule". Unfortunately, there is a dedicated group of users on this site and elsewhere that treat the tables as being the same as the rule and simply proclaim "sunny 16" whenever a noob has a sticky exposure question.

The rule:
"In full mid-day sun at middle latitudes, exposure should be 1/(value of ISO speed) at f/16."

The rule tends to fail above latitude 45°N, below 45°S, in the tropics, in mid-summer/mid-winter, and wherever "full" sun is seldom the norm. In my region, it is closer to "sunny 11" in mid-summer and "sunny 8" at the winter solstice. A person keeping the rule in mind may do a reasonable job of estimating daylight exposure, particularly when shooting to a medium with high dynamic range.

The tables, OTOH, are based upon long-standing conventions (over 100 years?) of EV* based on common subject luminance (full sun, open shade, full moon, etc.) for available emulsions. Most have been applied to ISO 100 to allow easy calculation of reasonable exposure for common lighting situations without relying on a meter. I used to carry a card in my bag with a fairly comprehensive list. They are only tangentially related to the "sunny 16" rule.

Rant over...


Steve

* Exposure Value (EV) is simply a number that represents equivalent combinations exposure time and f-number. When used as an expression of scene luminance, convention is to assign ISO 100 unless explicitly stated otherwise. A full step on the EV scale is equivalent to one "stop" of exposure.

[photoptimist]

You make an interesting point. There is a place for "rough" interpretation of a subject.

There is also a very important difference between digital noise and film grain. The former is composed of pixels that have no data. As noise increases, detail, tonality, and color fidelity disappear and you can't mitigate loss due to noise. Film grain, OTOH, is intrinsic to image formation for most emulsions and is not distinct from "data". Paradoxically, minimization of grain is not a sure recipe for better fidelity. Fine grain tends to defeat definition despite improving resolution.* That may explain the continued popularity of the Rodinal developer formula.**



Indeed! Live view in the dark is nasty work. I have done as you state on several occasions where I needed fairly precise framing but had very little light.


Steve

* The nature of grain as it relates to image formation and fidelity is a complex subject and what I wrote above is overly simplistic. It is usually enough to simply state that grain is "supposed to be there", despite the fact that it may be obtrusive.

** Rodinal is renowned for promoting excellent acutance (definition and edge sharpness) with slow and medium speed film, though at the expense of more prominent grain.

Indeed! I totally agree that the statistical properties of film images are different from those of digital images. Although both media are quantum photon counters at heart, digital's rectilinear array of pixel buckets is very different from film's analog emulsion of silver halide grains. Moreover, as you allude to, the developer isn't just a simple linear convertor of the latent image in to a developed one. There's a lot of chemical magic going how the developer catalyzes the conversion of an excited grain into metallic silver and also modulates the conversion of adjacent grains. It's a chemical version of photoshop. And on the digital side, there's some mathematical non-linearities in demosaicing Bayer-filter RAW data that creates some strange chromatic artifacts in high-ISO images.

And yet at a deeper level, signal-to-noise is an invariant concept across the media. The human ability to perceive structure and tone in the image is a function of noise levels whether it is digital or film. One of the most welcome advances in digital sensor design and manufacturing has been the near-elimination of most forms of pattern noise in digital images. Digital noise is a lot more "film-like" than it has ever been.

What fascinates me at the creative level is that no matter what the deficiencies of a given medium, there's someone who finds a way to turn deficiency to advantage. Moreover, I'd not be surprised if in 20 years there will be "retro" filters that give the super-clean 64-bit images of that era the look of old 8-bit, ISO 800 from the dawn of digital.

[gm4life]

The really good news is how good ISO 6400 and 12800 look because you seldom need more.

Amen to that 12,800 looked like 1,600 on my K3II. Even 25,600 was decent, heck 51,200 might be usable. Anymore than 3,200 on my K3II is pretty iffy. Maybe 6,400 but beyond that then aren't really any good.

So glad my pre-order is in!