High ISO

[rpcrowe]

I am a fan of actual real-life testing vs. charts etc. This is especially true if you have the gear in hand to test. As an example, if a camera gives me quality decent enough for my uses (whatever those uses are) then that camera is decent enough for me to continue to shoot with.

The decision whether the camera is good enough for my use is predicated on the final product which I can hold in my hands and view or view on a monitor at the size that I would normally need it.

If that quality is sufficient, then I have no problems with the camera, despite whatever charts or extreme enlargements show me.

OTOH... If I don't have the camera in hand, then various tests that are published "just might" give me a basis for choosing one camera over another.

Here is a shot that I did in Hong Kong using a Canon 30D. At that time, I was shooting with a 30D and a 40D. I used the 30D at ISO 1600 with some noise reduction (this was years ago and I am not sure if I was working with NIK at that time.) Provided enough quality for my needs. I could not use a tripod because I was shooting over the heads of a crowd of spectators. The shot was hand held at f/2.8 using 1/45 second at a focal length of 44mm.



I am now shooting primarily with a 7D and a 7D Mark-2. This was an experiment using auto ISO at ISO 1250. The quality after some global Dfine noise reduction is plenty decent enough for my use..



I didn't need to go that high in ISO. I shot with a 1/1250 second shutter speed at f/8. The IS capability on this lens could easily have allowed me to shoot at a much slower shutter speed and the lens performance wide open is quite good. My dog was stationary so shutter speed did not impact her sharpness. I could have easily captured this scene using ISO 400 and very likely could have done O.K. at ISO 200 of even slower... ISO of 150 or so could have given me an exposure of around 1/300 second or so at f/5.6... I know that I can easily hand hold the camera at 1/300 of a second using 400mm. And I know that image quality of the lens is still very decent side open.

[billtils]

. I guess what I am looking for is to get a feel for when ISO 800 or 1600 might produce an acceptable image;

Rufus, there are several responses to your question as originally posed, but as was perhaps inevitable, the thread has in part morphed into how to deal with ISO noise - fair enough given what you say in the above.

Depending on the editing software that you have, there is a good and free (even better!) way to deal with ISO noise, namely stacking. The procedure is to shoot a few (the number will depend on how pushed you are for light, but the results below used 4) then edit as a stack. The example below was set up to give an idea of the results - a bowl of fruit on a black background, available light, ISO6400, 2.5 sec and f6.3. (C & C absolutely not requested on the image itself !!! ).

One of the original burst of 4:




The image after stacking:



EDIT 1: Looking at these as posted the difference is not as clear as on my screen, but does show some improvement

EDIT 2: The first image was uploaded without any problems using Tinypic; gave up on the second one after 10 failed attempts and uploaded from flickr

[Rufus]

You say, Richard:

"I am a fan of actual real-life testing vs. charts etc. This is especially true if you have the gear in hand to test. As an example, if a camera gives me quality decent enough for my uses (whatever those uses are) then that camera is decent enough for me to continue to shoot with.

"The decision whether the camera is good enough for my use is predicated on the final product which I can hold in my hands and view or view on a monitor at the size that I would normally need it.

"If that quality is sufficient, then I have no problems with the camera, despite whatever charts or extreme enlargements show me.

"OTOH... If I don't have the camera in hand, then various tests that are published "just might" give me a basis for choosing one camera over another."

I think this is a very sensible pragmatic approach.

[Rufus]

I have difficulty seeing the difference between the 2 images, Bill. But then I probably don't know what to look for!

Aha...! Having enlarged the images on screen I now see there is some noise (not much) in the first image and none apparent in the stacked one.

I think I have seen photo stacking done in Photoshop (I use CS6) but that was with a view to using different focal lengths, not noise reduction. I may ask my friendly PS person about that.

[billtils]

Correct Rufus. I did it in a bit of a rush this morning, and was not sure how far to push it - my camera handles high ISO noise well but I decided to keep it down to 6400 to stay within your range. Since you have S/W that permits stacking, give it a try and see how it works for you.

[William W]

My 10 year old Canon 450D camera has ISO settings from 100 to 1600. . . I guess what I am looking for is to get a feel for when ISO 800 or 1600 might produce an acceptable image . . .

This is an important part of the question, but I think quantitative testing becomes somewhat irrelevant when you come across a scene which you really want to capture. All the testing that you do (and I encourage you to do testing) will not account for every image that you might want to make.

***

I refer to John’s Post #2

. . . Your decision to use higher ISOs will be dependent upon the subject, any movement within composition, the dynamic range of the subject/scene. . .

And Manfred's Post # 10

. . . I'd rather get the capture and try to do something about the noise in PP than not get the image at all. . .

***

Below is an image made with an EOS 400D and the EF 35 F/2 at ISO1600.

I specifically dug out this image because I recall (hopefully correctly) that you like Architectural Photography and also the EF 35/2 is one of your two fastest Lenses (the other is a 50 F/1.8 or F/1.4 from memory).

*

The EOS 400D is very similar to your camera. There were improvements, and those included:

> 12.4 megapixel sensor and 14bit A/D converter capable of relaying 14 bit information;
> 'Highlight Tone Priority', which will retain more information in the bright areas of the image but, at the possibility of increased noise in the shadow areas;
> High ISO noise reduction

However it is relevant to note that dpreview sums up when comparing the 400D and 450D, thus -

“There's little visible (or measurable) difference between the EOS 450D and its predecessor [EOS 400D] at higher ISO settings, and both show Canon still has a slight edge over its rivals at this end of the market (thanks to superior processing and a more light-handed approach to noise reduction)”.


*

That preamble stated, the point I want to make is:

Imagine yourself is a situation like this - you have limited time; you are not allowed a tripod; stacking is not a possibility; . . . and you really, really want the shot.

You decide that you need as wide a lens as possible to get a feeling of the expanse and grandeur of the Mosque, but you also need a fast aperture because you need a fast shutter speed to ensure that there is little or no perceived SUBJECT MOVEMENT (yes he is praying, but he is still moving), remembering that Image Stabilization (for example on your Standard Zoom Kit Lens) will not assist that.

So you choose your 35/2 – you know if used at F/2, there will still be good IQ at the edges on your APS-C Format Camera, because of the cropping affect of the sensor.

You note the scene has a very large dynamic range and probably if you want to keep good detail in the interior, you will have to blow some of the highlights streaming in from the exterior strong sunlight.

You make a few meter readings from the interior walls and you note, that to even get close to a shutter speed effective enough to arrest any Subject Movement and Camera Shake you absolutely and definitively must pull the shot at ISO1600.



And then when you get home you use your Post Production Skills to make the best image that you can.

But the point is “a feel for when ISO 800 or 1600 might produce an acceptable image” is so much dependent upon how valuable any particular image is to you.

And the question is:

Do you make the shot or do walk away, simply because your quantitative testing has shown that using ISO 1600 will be “unacceptable”?


Here is a crop of the same image:



WW

REF:
https://www.dpreview.com/reviews/canoneos450d
https://www.dpreview.com/reviews/canoneos450d/32

All Images © AJ Group Pty Ltd Aust 1996~2017 WMW 1965~1996

[rpcrowe]

BTW: as an added comment... I will prefer an image with a bit of perceptible noise, any day, over an image that is blurry from using a shutter speed too slow...

Of course, I come from an old school which had black and white Kodak Super-XX at ASA (equivalent to ISO) 200 as the fastest film in my arsenal

[Rufus]

First, Bill, you have a very rigorous system that allows you to retrieve a shot from your archive that is so closely matched to my equipment!

Second, though, I love your detailed description of the scenario and thought process, which we seldom pay much attention to.

At the end of the day, my conclusion from all the helpful advice from you and others is that I must follow my "shoot at a low ISO" mantra but stop being afraid to take it up as necessary.

[William W]

First, Bill, you have a very rigorous system that allows you to retrieve a shot from your archive that is so closely matched to my equipment!

Thanks. A condition of the profession - meticulous Archiving and Indexing - personally I loathe the process but it has been useful and rewarding to have the process.

Second, though, I love your detailed description of the scenario and thought process, which we seldom pay much attention to.

Much appreciated comment. Quality Thinking for any situation is very important, IMO.

At the end of the day, my conclusion from all the helpful advice from you and others is that I must follow my "shoot at a low ISO" mantra but stop being afraid to take it up as necessary.

Sage. Keep things in balance.

WW

[paintingwithlight]

Do you make the shot or do walk away, simply because your quantitative testing has shown that using ISO 1600 will be “unacceptable”?

+1. I have met a few photographers who say, for example, "I NEVER shoot above 6400 ISO." I don't understand such sentiments. I am not a professional photographer, but personally I'd rather capture the scene and worry about noise later. Your question could be alternatively phrased "What ISO is too noisy to use today, and what ISO will become acceptable in a few years from now when I get some better noise reduction software?"

I believe that a comparable new camera will have a sensor with significantly less noise than your 10 year old camera but I can't substantiate this.

[DanK]

Of course, I come from an old school which had black and white Kodak Super-XX at ASA (equivalent to ISO) 200 as the fastest film in my arsena

That's old school. I don't think I ever used it. When I started in the 1960s, my fastest film was Tri-X, rated at ASA 400 with standard developing. You could push it about one stop faster in developing, if I recall. It was introduced in 35mm format in 1954, according to Wikipedia.