Newbie with underexposure problem on k200d

[pma0]

Hi all,

I've just bought my first DSLR so let me first state that my problem is not for sure coming from the hardware. I really am a newcomer and I'm seeking pieces of advices.

[set up]

I'm experiencing critical mis-exposure problem when using my k200d with or without flash and in every mode (so far I have only used auto, P, M, Tv and Av modes). My k200d being fited (mostly) with a tamron 17-50 f2.8 and a metz 58 flash (I have also tried the pentax 55-300 and the integrated flash). I use AF-S and either multi-segment exposure or center weighted exposure.

[conditions]

Most of my first shots were taken indoor with flash at short range (one to two meters). Few were also taken outdoor and some also without flash.

[result - problem]

I'm always displaying the histogram and without any exposure compensation I have always zero pixels in the right quarter of light intensity. This result in very dark images as you may have guessed .

[unsatifactory solution]

So far the only way I have found to solve this is to play with exposure compensation : I must use a [0.7 to 1.3] eV exposure compensation and for the flash also from [0.7 to 1.3] eV to get pixels all over the luminosity spectrum. Yet this expedient really is a problem in itself, since depending on the situation +0.7 eV can be much too dark or conversely +1.3 eV much too bright etc. The exposure problem I'm encountering is not solved by a constant exposition shift.

[question]

Any way, I would exclude a problem from the flash : its clearly as powerfull as needed to light up any scene in my appartment. Even used indirectly at f8 and 100 ISO. This is easily tested by setting up a strong exposure compensation with the flash : it is able to render any item absolute white. So the problem seems to be coming from either

• me being not able to tell the k200d that I want some correctly exposed pictures ;
• my sample of the k200d having a strong exposure measurment problem.

Since I guess the default purpose of the metering is to render correctly exposed picture I fear very much that my camera has a hardware problem. Any thought or advice anyone (thanks in advance)?

 

[Ishpuini]

My take to indoor flash photography is rather simple: bounce the flash and push the ISO.

Straight flash tends to make the camera underexpose because of the flash light reflected straight back at the camera during the pre-flash (which is used to measure the amount of flash light required for the real exposure). Anyway, that's how I understand it.

Most of the time I take a few test shots before the "real" thing. Just to find out how a room responds to flash. Ceilings come in many types, not all of them as suitable for indirect flash. I generally start with f/8. If I don't get enough exposure I push the ISO instead of touching the aperture. Don't want to loose DOF. My AF540FGZ is mostly set to +1 for extra light.

Solution if bounce doens't work because of too dark or too high ceilings: a diffuser. A Sto-fen omnibounce is very compact and best used with the flash tilted up 45 degrees. If no need for compact, any of Gary Fong's solutions is great (google for his website). I personally use his Lightsphere II.

Wim

--
Belgium, GMT+1

 

[Paul Hunt]

Are you able to post some sample shots?

Do you get the same results with a different lens?

Paul

 

[pma0]

My ceiling are low and white, and I have mostly sticked to indirect lighting (+ small reflector) so far. However it doesn't change anything to my underexposure problem.

 

[pma0]

Both of my lenses result in the same underexposures whatever the aperture, the speed, etc. I'll post a few sample tonight (Paris time).

 

[Dreticus]

Leave out the flash for a while, it's a complicating factor.

Try to take a few "easy" pictures, e.g. inside without windows (no flash -no problem if the pictures are not sharp because of handshake); or outside under a clouded sky. And show us the results with the histograms, that should reveal a lot.

After that it's time for step 2 -whatever that is.

Hi all,

I've just bought my first DSLR so let me first state that my problem
is not for sure coming from the hardware. I really am a newcomer and
I'm seeking pieces of advices.

[set up]
I'm experiencing critical mis-exposure problem when using my k200d
with or without flash and in every mode (so far I have only used
auto, P, M, Tv and Av modes). My k200d being fited (mostly) with a
tamron 17-50 f2.8 and a metz 58 flash (I have also tried the pentax
55-300 and the integrated flash). I use AF-S and either multi-segment
exposure or center weighted exposure.

--
Andreas

 

[steephill]

Not all subjects should have pixels in the top quarter of the histogram. It all depends on the subject. For example, if your subject consisted entirely of mid-tones then there shouldn't be any pixels in the highlight or shadow regions. One example of this would be a shot of a neutral coloured wall which would only show a central peak with no shadow or highlight detail.

Useful articles on histogram here http://www.ronbigelow.com/articles/histograms-1/histograms-1.htm

Your aim should be to obtain correct exposure so that the approriate tones are placed correctly on the histogram. This may involve applying some exposure compensation to the suggested exposure on canera and/or flash.

One thing to watch out for with P-TTL flash in particular is that if there are any specular highlights such as reflections from glass or metal then the expsoure will try to protect these highlights and the result will be underexposure. Look for a spike at the extreme right hand side of the histogram to check for this.

--
Steve

http://www.pbase.com/steephill

 

[pma0]

Not all subjects should have pixels in the top quarter of the
histogram. It all depends on the subject.

Yes. Yet I would await for the peak to be at the center of the histogram not in the middle of the first half. However this is a bright idea for a test. I'll do that and post the result.

One thing to watch out for with P-TTL flash in particular is that if
there are any specular highlights such as reflections from glass or
metal then the expsoure will try to protect these highlights and the
result will be underexposure. Look for a spike at the extreme right
hand side of the histogram to check for this.

So far, without exposure compensation, I get the highlight peak (in case of such reflection) almost exactly on the line that separate the third and the fourth quarter of the histogram.

 

[GaryDeM]

my suggestion on one of yourn pics is to use auto levels in photoshop elements. this is designed to put then overall brightness at the correct 18% amount. so sim[ply open the image in pe and hit auto levels, ignore the histogram, just watch the image to see the before and after brightness amount. if there is a large change in brightness then the pic was shot with the wrong exposure. if there is little or no brightness change then the image was correctly exposed. this is a real simple test.

the following may be of interest to you being a new dslr user-

as far as i know no dslr under or over exposes.

the user must know what the meter is doing how it is doing it why it makes certain decisions how it is processing the info and how it is presenting the info to the user.
unless the above is known any talk of under or overexposing is meaningless.
--------------------------

matrix metering is the same as averaging metering with an addition. it also includes the use of EC to prevent highlight blow outs. the reason that matrix metering has many segments is as follows- in matrix the meter determines the correct exposure, then checks each segment to see if at that exposure if any have a highlight that would blow, if yes then the meter backs down the exposure so that the segment will not blow any highlights.

so if you have the meter in center weighted and then you are using EC to prevent blowout then the matrix system is doing the same thing. if you are spot metering and seeing which areas are the brightest and setting the exposure that way; well, the matrix system is also doing that for you.

this is exactly why i use matrix almost exclusively on my dslr. the object of metering is to get the correct exposure and not blow highlights, and that is what matrix(evaluative) metering is doing.

once the pics are downloaded to the pc, all that is necessary is to use auto levels in pe6 or csx, or similar in another pp program and the light level of the pic with be restored to normal. and no blowouts. matrix metering is protecting you from making highlight blowouts. let it do it, that is what it is there for.
as far as i know no dslr under or over exposes.

the user must know what the meter is doing how it is doing it why it makes certain decisions how it is processing the info and how it is presenting the info to the user.
unless the above is known any talk of under or overexposing is meaningless.

 

[GaryDeM]

i wrote the following posts some time ago, they may be of interest.

no matter which dslr you buy.

heavily consider the following. there are NEW DSLR owners' writing in all over these forums on this subject.
when changeing from a p&s to a dslr, there is a huge difference.

when you take p&s out of the box add a memory card and a fully charged battery you can now shoot and take very good pics.

BUT, you cannot do this with dslr. the camera HAS TO BE SETUP first. you have to adjust the contrast/saturation/sharpness/shooting modes(color style or whatever it is called) to your likes. if you don't it is quite likely you will disappointed with results. your p&s will likely outshoot the dslr.

to setup-you have shoot a test shot make ONE adjustment reshoot check pc screen readjust, until you are satisfied. and you do this with each of the adjustment types. then you have all the custom adjustments in the menu to check and if wanted change.

when done you can put the camera into AUTO or PROGRAM and get reasonably nice shots. i would advise at first staying with jpeg. as you learn about the camera and photography you can then go to the other shooting modes and try RAW if you wish.

dslrs are made to see the shot through the optical viewfinder not through the lcd. this is true of almost all dslrs including the k10d.
dslrs and color.

if you mean heavy saturated colors then no dslr is going to do that. they are not made to give strongly saturated colors. they are made to give ACCURATE COLORS. not heavy saturated colors.

this is not the same thing at all. too many people who come from a p&s are very disappointed in the dslr colors, because they are not bright and saturated. this is because they are and have been using a p&s which has been giving them saturated and incorrect colors for so long that they think it is the right look. nothing could be further from the truth. the p&s colors are wrong, wrong. the camera manufactures know that the public buys high megapixel and heavy saturated colors and is what they make and sell to the public.

but the slr/dslr is a whole different world. for the dslr accuracy of the scene in terms of view and color is a religion rpt religion. you want accurate color that is what you are going to get with dslr. but they will not be the bright saturated colors of a p&s. ytou can with adjustments in the menus up the color is dslr, but it will not look the INACCURATE CARTOON COLOR of the p&s.

if you are wishing to buy a dslr for more and brighter color, save you money the p&s is what you want.

not too long ago a new owner of a dslr was on these forums talking about the poor color of his new dslr. it seems as if he was shooting on an overcast day. many many people replying to him told him that cloudy day shots give the most accurate color, which they do. he couldn't believe and get over that idea. he also owned a p&S previously.
you might be interested in this; which i posted a while back.

http://forums.dpreview.com/forums/read.asp?forum=1036&message=23677257
----------------------------

to setup for jpeg with new camera-

there are 4 functions that may be adjusted. the color mode(or whatever it is called) saturation contrast and sharpening. i assume you are using a calibrated monitor. simply select a scene immediately outside your house. hopefully it has lights darks and colors. all settings in the camera are at zero or default. adjust color mode first then check the shot on the monitor, decide if ok, if not adjust reshoot and recheck. go on to each of the other adjustment settings. the object is to get the monitor scene as close a possible to the real scene outside. do not be concerned if the finished monitor scene has enough color for your tastes; the amount of color can be adjusted in pp. you are going for accuracy between the 2 scenes. the real and the one on your monitor; when done the 2 scenes should look identical or as close as possible. do not hurry. the adjustment process could take several hours. but once done leave the settings alone. at this point you know that the camera will accurately make the best most accurate pics possible of the scene. after i set my dslr up 3+ yrs ago about, i have not ever moved the settings. It took me 2-3 hours to setup my dslr.

if i needed/wanted more color or whatever that is what pp is for. i also try very hard to do my composing in the camera and not crop heavily in the pc. my thinking is why buy a 10mp camera and crop away 40%. you are then no better that a 6mp that is not cropped. besides which the cropped 10mp is noisier.

i would not adjust the contrast to get more DR. to me you just have to get used to the idea that digital has DR limitations. i shoot slides for 32yrs; the DR in digital and slides is about equal. i never had a problem. While DR limits exposure, lighting should/can be adjusted to compensate. if you want more headroom in your camera for taking jpegs, use adobeRGB color gamut. it gives slightly more headroom.

 

[GaryDeM]

to use the stofen omnibounce on a 360 or 540-

mount the omnibounce on the flash. it will be very tight.

put the bounce angle of the flash at 45degrees. if you have the 540 which swivels, then you can use the the omnibounce in portrait mode as well. simply put the bounce angle at zero, use portrait mode and use the swivel only to put the 45degree angle back on.

note the indicated distance on the back of the flash for possible shooting distance is now invalid. the actual distance will be less; the omnibounce uses up some of the possible range.

also, to remove the omnibounce from the flash is a struggle because it is so tight. just keep working your fingers around the omnibounce and pulling and it will slowly come off.

the omnibounce effect is the same as the bounce effect, with one addition-the light that is coming from the omniboune is always white. the light bounced from the ceiling or walls is dependent on the color of the walls.

 

[GaryDeM]

it would be very helpfull if you could post some of the pics you are talking about.
we could look at and analize them for you.

 

[jeffkrol]

• my sample of the k200d having a strong exposure measurment problem.

Since I guess the default purpose of the metering is to render
correctly exposed picture I fear very much that my camera has a
hardware problem. Any thought or advice anyone (thanks in advance)?

Incorrect premise no. 1 "meter is used to get the correct exposure") That is a function of Matrix metering but not CW or spot.
See my favorite quote in my sig.

Incorrect premise no. 2 "center spiked histograms ". Generally speaking the average of the exposure should be in the neighborhood of 110. Left of dead center.

And as stated below, watch out for "hot" spot reflections. It will definitely pull your metr down and give you a very dark exposure.........

--
360 minutes from the prime meridian. (-5375min, 3.55sec) 1093' above sea level.

'The exposure meter is calibrated to some clearly defined standards and the user needs to adjust his working method and his subject matter to these values. It does not help to suppose all kinds of assumptions that do not exist.'
Erwin Puts

 

[improbable]

This is an aside from the original poster's question...

matrix metering is the same as averaging metering with an addition.
it also includes the use of EC to prevent highlight blow outs. the
reason that matrix metering has many segments is as follows- in
matrix the meter determines the correct exposure, then checks each
segment to see if at that exposure if any have a highlight that would
blow, if yes then the meter backs down the exposure so that the
segment will not blow any highlights.

I don't think that's a terribly good description of what matrix metering does. The classic problem it was initially designed to solve was almost the opposite: shooting one person against an overcast sky, an averaging meter will underexpose badly, putting the sky at mid-grey, while you want the person there and the sky kept bright.

If you find this sort of thing interesting, I highly recommend reading Nikon's original description of how matrix metering works:

http://www.kenrockwell.com/nikon/matrix01.htm

(Can't read this without getting all nostalgic for the days when companies didn't assume their customers were complete idiots to be advertised at, not taught. Although I realise this is probably one more pastoral myth...)

 

[Marc Sabatella]

Yes. Yet I would await for the peak to be at the center of the
histogram not in the middle of the first half.

Not necessarily. What should really be happening is that the "average" (not necessarily peak) value should be a little to the left of center - corresponding to around 12-13% reflectance, as opposed to 18%, which is more or less the center.

And of course, if you're too far from your subject, or trying to bounce flash off a reflector where the total distance is too far, or using too small an aperture and/or too low an ISO, there is nothing the camera can do to make the picture any brighter. But it sounds like that isn't the problem here.

--
Marc Sabatella
http://www.marcsabatella.com/
Blog: http://marcsabatella.blogspot.com/
Flickr: http://www.flickr.com/photos/marcsabatella/

 

[GaryDeM]

you should check the date of the article that you suggest people read. the date was 11/25/2000. that is 8 yrs ago. nikon and everyone else have changed their metering since then. nikon's 3d matrix is a fairly recent developement and did not even exist in the yr 2000. and it works the way i described but adds a check of 10000 to 40000 stored scenes against which it checks the metering to also help determine the correct exposure.

 

[pma0]

'The exposure meter is calibrated to some clearly defined standards and the user needs to adjust his working method and his subject matter to these values. It does not help to suppose all kinds of assumptions that do not exist.'

Sensible quote. The point is there is no clues in the manual about what precisely the metering is supposed to do. At least nothing like the "improbable" link at Ken Rockwell's site.

In my case using the default metering choice gives the following sample shots (very representative of the whole production):
http://anglade.googlepages.com/IMGP0738.JPG/IMGP0738-full ;init:.JPG
http://anglade.googlepages.com/IMGP0739.JPG/IMGP0739-full ;init:.JPG

Viewing the histogram on my k200d I get most of the histogram in the two first quarter of the range.

My (maybe incorrect assumption ; yet no actual information doesn't help a lot) would be that, with default exposure parameters, I should get something closer to the following image:
http://anglade.googlepages.com/IMGP0733.JPG/IMGP0733-full ;init:.JPG

Here I have set +1 exposure compensation and the histogram spreads over almost the complete spectrum.

I have posted and described with more details some more images here : http://anglade.googlepages.com/k200d-metering_problems . Nothing beautiful I was just trying to provide a strong ground for further discussion based on the comments of steephil, dreticus and Paul Hunt.

Do you, experienced shooters, feels I am doing something wrong or expecting something stupid from the camera ?

 

[Jonathan Demarais]

Pentax's (K10D, K20D anyway) shoot dark, by 0.5 to 0.7 f-stop. Bracketing is good, and in most cases, possible.
--



'I cried because I had no E-3. Then I met a man with no E-510'

Olympus E-410, Nikon D100 (IR) & Pentax K20D.
57 lenses of various types from most brands.

 

[pma0]

I have posted some samples at the (current) end of the thread... This will be a safer ground for discussion.

my suggestion on one of yourn pics is to use auto levels in photoshop
elements. this is designed to put then overall brightness at the
correct 18% amount. so sim[ply open the image in pe and hit auto
levels, ignore the histogram, just watch the image to see the before
and after brightness amount. if there is a large change in brightness
then the pic was shot with the wrong exposure. if there is little or
no brightness change then the image was correctly exposed. this is a
real simple test.

Although my personal taste is usualy for slightly overexposed pictures, I hope I'm not at the level I need a computer to tell me if a picture is over or underexposed ! Anyway, using a much less expensive (and much more open) software than photoshop (TM) auto exposure tool I get some sensibly lighter pictures.

as far as i know no dslr under or over exposes.
the user must know what the meter is doing how it is doing it why it
makes certain decisions how it is processing the info and how it is
presenting the info to the user.
unless the above is known any talk of under or overexposing is
meaningless.

Unfortunately this is absolutely not described in the small manual coming with the k200d. What a pity. That's the very reason I'm asking.

this is exactly why i use matrix almost exclusively on my dslr. the
object of metering is to get the correct exposure and not blow
highlights, and that is what matrix(evaluative) metering is doing.
once the pics are downloaded to the pc, all that is necessary is to
use auto levels in pe6 or csx, or similar in another pp program and
the light level of the pic with be restored to normal. and no
blowouts.

Doing so you loose part of the accuracy of your pictures if you end up extending its dynamic range. If you don't and your picture is widely underxposed (as mine are) you should end up (unless you really are a magician) with a dull pictures laking contrasts. My own approach is to tune camera settings in order to get something as close as possible to the final result I expect directly in the camera. I guess this is what people call an "inexperienced user attitude" huh ?

as far as i know no dslr under or over exposes.

Having spend some time (8 years) reading reviews here and there before making a buying decision, I can tell you that some model or at least some samples do have this kind of problem. This is not my words but those of experienced reviewers. Allthough they can mistake, their judgement is likely better than mine.

the user must know what the meter is doing how it is doing it why it
makes certain decisions how it is processing the info and how it is
presenting the info to the user.
unless the above is known any talk of under or overexposing is
meaningless.

Very sensible statement. I hope that with the pictures posted below people will have some interesting comments on what I feel like being a real problem.

 

[pma0]

PS: thanks for taking the time to explain so much things "a priori". Although in this case it was not exactly helpful its a very appreciable attitude and I am grateful for that. I hope the pictures below will be able to help you give me a more accurate diagnostic about my, so called problem.