IS on and off

[Ron Kruger]

I doubt there is the same effect in IS lens technology, but I am seeing a difference in IS in-camera technology. They may both be trying to accomplish the same thing, but they are doing it from different angles and with different technologies.
Something is happening to saturation and hue between IS on and IS off, and I believe it is due to slightly different (about 1/3-stop or so) auto-exposure readings between IS off and IS on.
I may be confused, as you suggest, or stupid, as Loupey suggests, but something is going on.
I didn't start taking photos yesterday.

[Ron Kruger]

Listen, you guys rule the roost here, so if I'm stepping on your attention toes, I'll bow out.

[Canon_Bob]

Listen, you guys rule the roost here, so if I'm stepping on your attention toes, I'll bow out.

Hey Ron, I don't think anyone's trying to offend here. I know I certainly don't want anyone to "bow out". You may have uncovered a flaw in Pentax's IS system.

From my point of view, I don't want the concept of IS being a bad thing being put out there based on isolated or inaccurate results. There are a lot of folks here who are newbies, and if they get the idea that IS is a bad thing, and turn it off because someone in a forum said it affected hue, saturation, and noise in their camera.....well.....I can't really sit back and ignore that. Perhaps Pentax has a flawed system if you're seeing it, but I guarantee you that Canon and Nikon do not have that issue. To me, IS is one of the best inventions in photography ever. And it's improving more and more. There are isolated instances where it may be counter productive, but the goods outweigh the bads BY FAR.

I suggest you lock your camera in manual mode, and redo your tests with the same manual settings in both IS on and IS off modes. Use the same subject, same light, and post the unmodified pics with Exif intact. THEN I'll buy that hue, saturation, and noise are affected in your camera with IS on or off.

You'll have to excuse me if I seem like I'm coming on a bit strong here, but I'm an Engineer by trade....so I always ask questions when things like this are presented. I am having trouble understanding how a purely mechanical function that has no bearing on exposure settings can have an effect on color and noise level. That's all.

Again, there is absolutely no intention of offending here. You may well be onto something that is either Pentax or "in camera" IS specific.

-CB

[scott-devon]

This is a very interesting thread. I believe you and I both use Pentax K20s Ron. I'd be happy to set it up indoors and run some shots to test this effect. It might be interesting to see if my camera does it as well. Any suggestions on settings?

[Dylan8i]

i agree with what others have said, this is an interesting subject, but as a scientist ( or at least trained to be one) the ONLY way to do a comparative test like this is to keep EVERY setting the exact same EXCEPT the one you are testing. that means manual exposure, manual focus, manual WB, everything except the VR switch and tripod. but think of it this way... with a camera it is REALLY EASY to do this. most experiments its impossible to keep everything but what your looking at the same. and thus you should get really definitive results.

[AgingEyes]

i agree with what others have said, this is an interesting subject, but as a scientist ( or at least trained to be one) the ONLY way to do a comparative test like this is to keep EVERY setting the exact same EXCEPT the one you are testing. that means manual exposure, manual focus, manual WB, everything except the VR switch and tripod.

But if the tests so far seem to suggest that the in-camera IS system of Pentax may be affecting the performance of the in-camera exposure meter, giving different readings depending on if IS is on or off, then leaving the exposure mode to auto seems to be the right way to go for additional testings.

Then again, by using manual exposure mode, and the test results still give us two differently exposed photos, then it's something else that the IS is affecting, I suppose, assuming IS itself does not - I believe - affect exposure at all.

[Canon_Bob]

But if the tests so far seem to suggest that the in-camera IS system of Pentax may be affecting the performance of the in-camera exposure meter, giving different readings depending on if IS is on or off, then leaving the exposure mode to auto seems to be the right way to go for additional testings.

Then again, by using manual exposure mode, and the test results still give us two differently exposed photos, then it's something else that the IS is affecting, I suppose, assuming IS itself does not - I believe - affect exposure at all.

Exactly. the only way to prove this out is to lock out all other variables except IS and then tripod. If the IS on created diffent exposures than IS off under those conditions, then a case can be made for light angle caused exposure variances against the sensor (which is moved on most in-body IS systems). You'd also want to test this in very stable lighting conditions.

Any other testing conditions would quickly cause any conclusions surrounding hue/saturation to be a product of changing light/and or auto exposure setting changes. IS would quickly be dismissed as a possible cause for the variance.

[Ron Kruger]

The noise thing is a whole other issue, related to ISO, not IS. And I'm not suggesting in any way that Pentax has "a flawed system." Quite the contrary: My results with IS on were better than those with a tripod and IS off.
In fact, listening to you Canonites, I was beginning to think I made a mistake switching from Nikon film to Pentax digital, instead of sticking with Nikon or switching to Canon or Sony. I don't tend to think that just because it is the camera I chose, it must be the best.
The thing is, both the IS and ISO experiments I've conducted under actual working environment conditions have reinforced that I've made a good decision switching to Pentax and renewed my confidence in the equipment--not the other way around.
These experiments, though maybe not as scientific as some would like, are very practical for my real world applications, in the field, where I make part of my living. I've sold five magazine covers since getting this Pentax last June. That alone should speak well for Pentax quality.
If I switch everything to manual, and shot indoors under closely controlled lighting conditions, would I see the same (slight) differences in exposure values? Probably not. But what would that prove to me for practical applications? Nothing. That's not the way I shoot in the field, nor is it practical to even think of shooting that way in the field.
So you all can take my experiments and finding for what they're worth to you, and for any beginners out there who also misinterpeted what I was trying to say in an objective way--generally IS is better on than off, and is the greatest part of modern digital technology.
Also, beginners should be aware that what I noticed in exposure values were very slight and that I look at all this with a very critical eye toward professional quality beynd what would matter to most.
Nor do I think that my experiments are conclusive in any way. There are so many variable in photography, and there can be differences in equipment, even between copies of the same model. That's why I've suggested that other conduct their own experiments with their own equipment.
So I'm not scientifically ignorant here, either. I'm just trying to be practical for my own business purposes.
At any rate, I think I'll leave all the important stuff to the important people in the future.
Digital Photography Review has a bunch of working professionals and includes the guy who has written over five books about photography and the business of selling images, plus I got to know Pentax's head of product development over there.
There's a bunch of great people here, too, and I've learned a lot. Maybe I'll hook up with some of them later.

[Canon_Bob]

Sorry you're feeling offended Ron. That was certainly never my intent. I would've been interested in seeing your next wave of tests.

I consider myself completely agnostic anymore when it comes to manufacturer. When I bought my gear, I researched who was the best at the time. It was Canon....Hands down. Now I'm into that system for more than 15k in glass alone....so switching is highly unlikely unless one of the competitors makes a SERIOUSLY significant advancement. In fact, there's a reason I'm still shooting a 1DsMII Canon instead of the Mark III. The reports of the Mark III focusing problems caused me to decide to skip a generation!

That said, if I were buying today...It'd probably be Nikon. They have (IMO) leapfrogged Canon for the time being. I expect Canon will jump ahead again very soon. The dark horse to watch, IMO, is Sony. They have made moves to position themselves right at the top. I expect they will pretty soon. I don't know a thing about Pentax or Olympus. They may be great for all I know. At the end of the day the fact remains that the photographer has a heck of a lot more to do with the resulting image than the camera does.

[Dylan8i]

yep first you would need need to isolate the IS function itself to see any differences. then you can move out to using AE under controlled lighting to simulate the field useage that Ron wants to test.

if there is a difference when only IS/tripod is changed then you know thats whats causing it. if there is no difference when just IS/tripod is changed (think of it as a control) but there is one when AE is on and IS/tripod is changed, then you can say that the AE programing is changing something when the IS switch is flipped. THEN you can worry about having to compensate for it one way or another.

[Singletracklovr]

I don't think the tripod test will show any difference for the nikon users with a VR lens. The Lens manual states it detects panning and activates IS. So on a Tripod VR on or off should result in an identical photo.

VR
Vibration Reduction (VR)
This innovative VR system minimizes image blur caused by camera shake, and offers the equivalent of shooting at a shutter speed three stops (eight times) faster.* It allows handheld shooting at dusk, at night, and even in poorly lit interiors. The lens’ VR system also detects automatically when the photographer pans — no special mode is required.
* As determined by Nikon performance tests.

[scott-devon]

As I was shooting some indoor shots with a home-made lightbox I shot these jellybeans as a test for this thread. They were shot with the Pentax K20d in all manual mode, at F22, 1/15 second, 400iso, and 200mm. Using a Tamron 28-200mm zoom. The camera was on a tripod for these of course, tripped with a wireless remote and the only change made was to shoot one with SR on and one with it off. No photoshop was done other than to downsize and convert to jpg.

Here is SR on...



And SR off



Hope this helps, although I may be missing something in the discussion that makes this test irrelevant. I do believe I see a loss in the saturation of colours with the SR on, it's very slight but I don't believe I'm imagining it. Actually with the SR off the entire shot seems just slightly lighter. Actually with the SR off the lighting in the shot seems just a fraction brighter.

[mjs1973]

Interesting. When I first looked at these online, I thought I could see a slight difference. I downloaded both of these images and put them into a single layered file in PS and made them registered exactly the same. When I turn one layer off, I do notice a very slight change, but from what I can tell, it's the noise moving around.

So then I took the color sampler tool, and selected an area using the 11x11 pixel average. When I would check the RGB numbers between the two layers, checking the exact same spot, there was a slight difference. For example, one went from R-171 G-97 B-58 to R-171 G-97 B-56. This could be caused by some different colored noise moving into the selected area but I don't know for sure.

[Dylan8i]

the with out does look lighter to me to. especially in the jelly bean shadows.

[Ron Kruger]

Well, by golllly. That's what I've been talking about.
If you will put the two images so you can see the bottom beans at the same time, you'll notice a slightly darker cast to the whole scene and more color saturation in the beans--all colors. That's what I've noticed during field work. It's not saturaton actually, but slightly different exposure values, and you naturally get the look of more saturation with a darker exposure. I often underexpose my shots by -1/3 to get that, because it's better for publication purposes.
In-camera metering systems are very good these day, but they are designed for the average guy who wants to print nothing larger than 8X10s or post them on the internet, both of which use an sRGB color space. Commercial printing uses a larger and more sophisticated color space of aRGB, but even with my camera set at aRGB, the "perfect" exposure is designed for computer screens and cheaper (relatively speaking) printers and just a tad too bright for printing. So I often underexpose shots to get those richer colors for printing.
I shoot differently than most, because I shoot with the printed page in mind, and printing (especially ink absorbstion of paper) robs images of some color saturation and contrast, so I bump them slightly to get the kind of "magic" I used to get from underexposing Kodachrome 25 back in the day.
We're not talking about a big difference in the shots here or what I got in the field--less than -1/3--but I've been doing this for over 30 years and spend a lot of time looking critically at images. I knew I was onto something.
I still have no idea what's causing it, and it really isn't a big deal, but I'd like to see some similar tests conducted with Canon and IS lenses.

[Canon_Bob]

I too see the subtle exposure differences. I'm willing to do the Canon IS version of the same test.

For consistency though, I suggest shooting 5 to 10 frames of each for comparison. I'll use a 70-200mm IS lens for the Canon test. It would also be interesting to see if the anomaly follows at differeing focal lengths. I won't be able to do it until this weekend though. The jellybeans look like an excellent test subject, so I'll use the same here. We need a Nikonian and an Olympus shooter in this too! If this isn't a photo-geek thread, I don't know what is!

[Dylan8i]

if i can get some time, i might try with my nikon 70-200 (its my only vr) but i might not be able to make anything with enough consistent lighting for it to work.

[Singletracklovr]

If this isn't a photo-geek thread, I don't know what is! -Canon Bob

I'm digging it. :-)
if this isn't some of the best grassroots photography reviewing I don't know what is.

I think Scott's experiment is awesome. I look forward to other inputs.
I will be in the woods shooting all day tomorrow. I'll bring the tripod and try to setup something, probably more along the line of Ron's outdoor test.

PR (((ROCKS)))

[mjs1973]

Scott, did you shoot these test shots in RAW or JPEG? The reason I ask is because if shooting JPEG, the camera is processing the images. I wonder if it's possible that the camera could process the images differently between shots? Just a thought.

[scott-devon]

Scott, did you shoot these test shots in RAW or JPEG? The reason I ask is because if shooting JPEG, the camera is processing the images. I wonder if it's possible that the camera could process the images differently between shots? Just a thought.

Mike, these were shot in Raw then simply resized in photoshop and saved as Jpegs.

The difference to my eye seems to be as Ron states a difference in light which is affecting the saturation. It is so slight as to be negligable for my level of photography but for someone shooting for publication it is interesting and worth being aware of although I'm certain it could be corrected for in post processing or with an ev correction while shooting. I'd be interested in seeing comparison tests with other brands simply out of curiosity.

Great eye to see this Ron and thanks for bringing it up. My level of expertise as a photographer isn't very high but that won't stop me from remembering this and shutting off the SR when I use a tripod. I'll take any advantage I can get...

[Ron Kruger]

Differences in exposure are even more dramatic between lenses (depending upon coating), and at different focal lengths (because of total scene light diffusion). That's why we often notice differences in how different lenses handle color saturation. I've even noticed slight differences between copies of the same lens models.
"Color satutration" probably isn't the correct term for this, because I believe it actually stays the same, according to the cameras processor, but what is changing is the tone of the shots, and even the slightest darkening of a scene will give it the apearance of more saturation. What's actually happening here, I believe, is minor light variances within the camera's f-stop perameters, which on my K20D I presume are graduated in 1/3 stops.
I do believe, also, that digital sensors are more sensitive to minor light variances than film was. I played around with Kodachrome for many years and often underexposed it from 1/4 to a full stop, and did a lot of bracketing, to get the maximum Kodack "magic." I find when I bracket shots on my K20D at 1/3-stop, I get almost as much variance in tone and color saturation as when I bracketed Kodachrome by 3/4 or a full stop.
Anyway, I expect you'll get a lot of different exposure/shutter readings with the 70-200s mentioned at different focal lengths, so you might want to make note of those readings as well, and if you have a focal lock, be sure to lock it down for each series of shots at each focal length.
For you "controll" freaks, you're talking about less control than with fixed focal lengths, but I like the idea, because it will expand the experiment greatly (although it might get a little confusing keeping track of everything). Thankfully, computers give us all those paramaters without taking extensive notes.

[Singletracklovr]

I was at my local state park today hunting anything that would jump in front of my lens when I remembered to make a few tests of VR on and off using the tripod and RF shutter release.

VR ON
>>
>>
EXIF Information File name: VR_on.JPG
File size: 581462 bytes
File date: 2009:02:20 15:02:44
Camera make: NIKON CORPORATION
Camera model: NIKON D80
Date/Time: 2009:02:20 12:43:27
Resolution: 1024 x 685
Flash used: No
Focal length: 70.0mm (35mm equivalent: 105mm)
Exposure time: 0.0020 s (1/500)
Aperture: f/22.0
ISO equiv.: 800
Exposure bias: -1
Whitebalance: Auto
Metering Mode: matrix
Exposure: aperture priority (semi-auto)

VR OFF
>>
>>
EXIF Information File name: VR_off.JPG
File size: 572009 bytes
File date: 2009:02:20 15:02:48
Camera make: NIKON CORPORATION
Camera model: NIKON D80
Date/Time: 2009:02:20 12:43:34
Resolution: 1024 x 685
Flash used: No
Focal length: 70.0mm (35mm equivalent: 105mm)
Exposure time: 0.0020 s (1/500) Aperture: f/22.0
ISO equiv.: 800
Exposure bias: -1
Whitebalance: Auto
Metering Mode: matrix
Exposure: aperture priority (semi-auto)

I did several tests moving from spot to spot with the same results. Shot 1 was always more saturated(darker) than shot 2. I even stitched up and took shots with VR off first. Maybe the cameras sensor is storing a little light from the first shot.

[scott-devon]

That is interesting Bob, it shows that this isn't just Pentax Dslrs that have this variance. It would be interesting to hear a manufacturer's explanation of the cause.