[ffking]

Have you seen the Nikon D5's upper extended ISOs? - utterly pointless - I would prefer camera shake or motion blur

[jeffryscott]

As a former newspaper photographer, having the high ISO capability of these modern cameras would have been utterly amazing. High ISO is a tool, like any other, that when you need it, you need it. As a journalist, you often have little control over lighting situations and fast lenses and high ISO are necessities. The quality of digital FAR surpasses film. Compare Tri-X pushed to 1600 to a K3 at 12,800

If you keep your ISO below 800, you likely don't really need high ISO, or are needlessly limiting yourself.

[reh321]

MSNBC online's pictures from the NCAA basketball final were shot at ISO=8000 {using Canon equipment}, which allowed them to use both high shutter speed and a reasonably narrow aperture.

[interested_observer]

I do agree with your thinking in general. I went with the K5IIs over the K3 for several reasons - the ISO 80 setting, and that at ISO 1600 it had less noise than the K3 (night landscape astro - Milky Way images), and the price was fantastic.

Now, having said that - I do use ISO 51,200 every time I go out at night. Out in the desert at O'Dark Hundred (~ 2 am), it is pitch black - as in, you can't see your hand in front of your face. I use the GPS to provide the main bearing to the landscape object. Using a flashlight gives me an idea as to the immediate foreground, but then I dial in 51,200, and take about an 5 second image (its 30 times quicker than ISO 1600) and check the overall composition of the shot (and make adjustments and reshoot to check framing again). Here is an example with the 31Ltd...

It might be level with Liveview, but the only way to check the overall framing is a quick image (as in do I have the main item of interest in the frame the way I want it, proportion of the landscape to the sky, etc.). Also, surprising as it may be - a number of these "throwaway" shots are really not that bad (in terms of overall noise - a bit gritty). Then I dial it down to 400 for the landscape and 1600 for the sky.

ISO 64 or 50 would be nice, but ISO 80 is not native to the sensor, however it provides the maximum dynamic range. I can also bracket and stack to average down the noise.

Then there is ISO 1600 with the Sony sensor. It has been observed that 1600 appears to be the threshold where the sensor maxes out its collection of photons - with higher ISO values just applying various amounts of amplification.

• https://www.pentaxforums.com/forums/groups/135-astrophotography/4340-how-avoi...high-isos.html

Overall, the features are just tools to help with creating / capturing the image that you the photographer want.

[Kath]

In general I agree, but if you like to know your one-in-500 shot is covered then it can be a comfort.

I've found this to be true as well. Seems to me the point in having it there.

[Adam]

This thread might prove to be an interesting read as it touches on may of the points raised here so far:
https://www.pentaxforums.com/forums/76-non-pentax-cameras-canon-nikon-etc/317...ropaganda.html

[osv]

It seems as though Hi ISO specs on cameras is now the latest sales gimmick a -la the previous / ongoing mega-pixel war. Bear in mind this is only my view, I mean whats the point of these hi iso specs? the images look ridiculous. For me I would like the advancements to be made at normal or low iso's. Its obvious that cameras still have noise at base iso levels so why are manufacturers seemingly ignoring this end of the scale?

manufacturers aren't ignoring anything, they are giving people options, that they think people will want: Photographic Dynamic Range versus ISO Setting

notice how the a7rii beats both the d5 and the d810 over most of the dr vs. iso curve; if you want the best overall sensor performance, that's the camera to get... the only place that it could be considered weak is when shooting astro, there have been some complaints.

why are the d5/a7rii curves jagged, not smooth like the d810? because among other things, those cameras use alternative gain strategies at certain iso settings, to improve the performance.

it's going to be very interesting to see what choices pentax has made with the k-1, given the level of performance that the a7rii has brought to the table... the d810 is basically old and obsolete, it's dead tech, even tho it probably is a better camera for astro work.

if you want the cleanest possible iso, period, you need to get a 645z, a modern bigger sensor has huge advantages when it comes to iso noise, because bigger sensors have more surface area, to acquire more total light than smaller sensors.

[creative69]

if you want the cleanest possible iso, period, you need to get a 645z, a modern bigger sensor has huge advantages when it comes to iso noise, because bigger sensors have more surface area, to acquire more total light than smaller sensors.

Nice to know

[Rondec]

manufacturers aren't ignoring anything, they are giving people options, that they think people will want: Photographic Dynamic Range versus ISO Setting

notice how the a7rii beats both the d5 and the d810 over most of the dr vs. iso curve; if you want the best overall sensor performance, that's the camera to get... the only place that it could be considered weak is when shooting astro, there have been some complaints.

why are the d5/a7rii curves jagged, not smooth like the d810? because among other things, those cameras use alternative gain strategies at certain iso settings, to improve the performance.

it's going to be very interesting to see what choices pentax has made with the k-1, given the level of performance that the a7rii has brought to the table... the d810 is basically old and obsolete, it's dead tech, even tho it probably is a better camera for astro work.

if you want the cleanest possible iso, period, you need to get a 645z, a modern bigger sensor has huge advantages when it comes to iso noise, because bigger sensors have more surface area, to acquire more total light than smaller sensors.

I don't really buy that the D810 sensor is obsolete. These are slightly different sensors that perform differently at different points on the iso curve. The D810 is maximized for low iso performance, allowing for iso 47 (DXO measured) as lowest iso, which allows it to have significantly better dynamic range at base iso -- 1 EV more. The two sensors have almost exactly the same SNR throughout the iso range, the difference is that beyond iso 800, the A7r II sensor maintains a 1 EV improvement in dynamic range. That is a real improvement, but it is scarcely such a big improvement as to render the D810 obsolete.

The biggest thing is that the A7r II has really good video performance and if that is really important, then yes, it is certainly the way to go.

To the OP's question, I would say that I agree that most of the iso ranges are wasted, as (a) they are grossly overstated. The Sony A7r II at iso 102K is measured by DXO Mark at only 75K. In fact, both the D810 and A7r II iso values (I looked up both their curves on DXO) are overstated by half a stop. (b) they are scarcely usable without a large amount of work. Yes, if you shoot RAW and buy a high quality noise reduction software, you can probably make some of these work, but above 3200 on APS-C and above 12K on most full frame cameras is going to take a lot of work to be usable.

The solution a lot of folks suggest is just to turn these photos into black and white images, but even there, you need a certain amount of dynamic range to have a good black and white image and for the most part, that is lacking in these ranges. Yes, cameras like the A7s and A7r II are a little better, but they still look pretty poor once you go to the extreme values.

[zen3d]

As has already been pointed out, high ISO serves as an option. One that I rarely use. But I have two shots that I wouldn't have been able to take otherwise.

Both examples below are taken with a K-5.

The first one is ISO 51,200. It was taken hand-held in a dark interior and I didn't have a tripod with me. It is pretty grainy, but you can clearly see the subject and the grain arguably adds to the atmosphere:
https://flic.kr/p/rUe9tn

The one is ISO 20,000 (I originally wrote 40,000, but that was wrong). Time was of the essence, since I didn't want to miss the shot if I got my tripod. Yes, you can see some noise if you know where to look. But the subject and the background do a good job camouflaging that noise :
https://flic.kr/p/mQoM6c

[I tried inserting thumbnails into my post, but I couldn't figure out how to do it; you'll just have to click on the Flickr links to actually see the images.]

[Fauxton]

In general I agree, but if you like to know your one-in-500 shot is covered then it can be a comfort. Here's an ePM2 crop (~60% crop?) at iso 4k and 200k, straight jpeg, nr set to Low. I think this is one click down from max ISO, exposure = EV3. Not a poster candidate but quite acceptable to me for jpeg.



The Sony 4:3 sensor is yet another sweet-16 winner!

To jimr-pdx - pretty ironic that the OP's title was "The HI ISO Scam" and you reference an Oly camera. See Sensorgen.info data for Olympus PEN-E-PM2 . If any camera maker plays a bit fast and loose with ISO its Olympus. But none the less you did make a valid point - the technology in the Sony sensor since that generation has been class leading.
However I would think that the ISO increases are rather more a secondary outcome of the vendors pursuing greater Full Well capacity and lower noise at base ISO (for greater DR and tonal fidelity). Once that is accomplished well the camera essentially becomes ISO less. Shoot what you want and then apply corrections in post. In other words you don't need to shoot at a high ISO camera setting to enjoy the benefits that allow that to be offered.

[freddyisaac]

Here is one interesting thing I noticed about so called high ISO

If you look at DxoMark web site and say take a K5 along with a comparable Nikon on Canon the ISO graphs show that the Pentax is consistently closer to the actual ISO than either of the other two.

I noticed this when looking at the image tool from dpreview and noticed that Pentaxes seem to be more noisy when comparing ISO 1600,3200,6400 on others but if you look at the dxomark it shows that ISO for many Nikons are way off. Not talking about the D2,D3,D4 which are usually pretty close but the D7000, D5000,D3000 are usually an ISO closer to the lower. So if you do look at dpreview image tool it is better to compare Pentax ISO3200 with Nikon ISO6400 as they are often times closer in actual measured ISO and so the image noise is a better comparison and will often show the Pentax as better.

So Not only is high ISO a marketing gimmick but also manufacturers do cheat on this scale.

Again a sales driven metric and not a real life one.

Of course when I can take a picture of a black cat in a black room and see rainbows emanating from its eyeballs I'll know the whole ISO argument has gone from the sublime to the ridiculous.

cheers

P.S. of course it is subjective as to whether or not you agree with dxomark and dpreview.

[rawr]

I mean whats the point of these hi iso specs?

The 'high-ISO' scam seems a bit like the 'high-shutter speed' or 'small aperture' scams.
I mean, who ever needs or uses 1/8000+ sec shutter speed or f32 aperture or above? ...

Real pros use cameras with fixed shutter, fixed focal length, and fixed aperture. Everything else is a scam.

[Rondec]

The 'high-ISO' scam seems a bit like the 'high-shutter speed' or 'small aperture' scams.
I mean, who ever needs or uses 1/8000+ sec shutter speed or f32 aperture or above? ...

Real pros use cameras with fixed shutter, fixed focal length, and fixed aperture. Everything else is a scam.

Well, it is a little different. I do shoot 1/8000 second occasionally and it really does work (mainly shooting multiple exposures). I don't have many lenses that go to f32 (most stop at f22), so not sure. Seems like it would be pretty prone to diffraction. But no one buys a lens because it goes to f32, or do they?

The thing is that if a camera lists an iso of 1,000,000 as the high iso capability, they know that it isn't going to be usable. They are just sticking it there to sell cameras. Nothing more or less. It looks good on spec sheets and unsuspecting customers look and see that the D5200 can shoot to 25K iso while the D7200 can shoot to 102K. Clearly the D7200 must be better (well, it is, but not because of that).

[osv]

The 'high-ISO' scam seems a bit like the 'high-shutter speed' or 'small aperture' scams.
I mean, who ever needs or uses 1/8000+ sec shutter speed or f32 aperture or above? ...

Real pros use cameras with fixed shutter, fixed focal length, and fixed aperture. Everything else is a scam.

---------- Post added 04-10-16 at 05:55 PM ----------

I don't really buy that the D810 sensor is obsolete. These are slightly different sensors that perform differently at different points on the iso curve. The D810 is maximized for low iso performance, allowing for iso 47 (DXO measured) as lowest iso,

it looks attractive, even in the raw editor, but it's a fake iso that gains nothing, same problem with the a7rii: Nikon D810 fake ISOs | The Last Word

which i guess proves that the o.p. was right all along about iso scams, he was just at the wrong end of the iso spectrum

the d810 is still an xlnt camera, but it's technology is obsolete, the d5 is a more modern camera, with newer technology, including similar dual-stage sensor gain to what the a7rii has... i think that we'll be seeing more dual-stage gain in the future, it quite obviously pays off, what would be really nice is if you could turn it off and on or even various levels of sensor gain off and on.

the a7rii has the very latest tech, like fully usable efcs up to 1/1000th for instance, in practically all shooting circumstances, can't do that with any dslr, and no mf camera that i'm aware of... that's huge, no shutter shake and no mirror vibration up to 1/1000th.

bill claff uses dxo measurements whenever possible, but different criteria can be applied to that data, notice here how the data doesn't show any dual-stage gain curve on the a7rii, and there isn't any d5 data, because dxo hasn't posted it: DxOMark Derived Photographic Dynamic Range versus ISO Setting it looks like the curves on dxo because it's the same, afaik, but since it doesn't show the dual-stage gain, it's probably not as accurate a methodology.

here is bill himself, with links and discussion about his data: Sensor Read Noise comparison chart - DSLR & Digital Camera Astro Imaging & Processing - Cloudy Nights