K10D - underexposed images

I've recently bought used K10D with the kit lens, shutter count 1950 according to PhotoMe. The first images I obtained with the camera and Tokina 28-70/2.8 lens were quite too dark. I tried +0.5EV and +1.0EV: +0.5EV resulted in correct exposure, +1.0EV was slightly overexposed. I shoot all photos with multisegment metering and ISO 100, no corrections of contrast/saturation/sharpness, only brighter tone instead of normal.
I still need to take the same photos with the kit lens to compare.

Does anybody have encountered similar experience? I hope there is no problem with the camera - though it is still under warranty.

I am new into digital SLR but I've been using digital point and shot and Pentax MZ-5n for several years shooting slides quite successfully.

PS one more question: any possibility to find out the true number of shutter actuations, other than EXIF info from PhotoMe?

Good quality digital SLRs such as the K10D will tend to expose less highly than lesser cameras. This is in order not to blow highlights. You can always boost the brightness with software afterwards, but you can't recover detail from a highlight if it was never there in the first place.

Digital has less dynamic range than film, so exposure is even more critical.

The camera can behave differently with diferent lenses, though, and it may be that it's not behaving quite normally with this particlar lens. I don't know. Anyway it should not be too much bother to use EV compensation - and it's perfectly normal to use it sometimes to get the exposure YOU want rather than what the camera wants.

Use the flashing highlights display on the LCD screen to tell for yourself when the highlights are blown. When not blown, you can safely increase the exposure if you like.

PhotoMe is probably the most reliable way to find out the actual number of shutter activations.

Your camera sounds just fine, the K10D's default is to underexpose slightly - this is to ensure that highlights are not blown out (as they are impossible to recapture in post-processing).

If you're shooting JPEGs then it's probably best to leave your camera set at +0.5EV so that your results are usable straight away; if you shoot RAW then the extra headroom provided by the underexposure means you can obtain superior results when processing on your computer - many K10D users (including myself) think of the underexposure issue as one of the camera's positive selling points!

As for shutter actuations, I think you'll find PhotoMe is generally accepted as being an accurate indicator. I don't know of any other way of checking to be honest.

I didn't type quick enough!

Thanks for your help and explanations. I'll try the kit lens though I don't think there will be substantial difference in exposure under the same conditions.

One secondary question about over/underexposure with digital SLRs: overexposed photo is then not containing any extractable info in white (saturated) areas. But what about the underexposed areas that are all black? Can some information be revealed from them rather than from the white ones?

I have my camera set to -0.3EV to prevent blowing highlights and a simple boost on Shadows and mid tones usually does the job. I just check the histogram and if necessary, adjust EV accordingly.

It is possible to regain some highlight detail but usually only from RAW files where as shadows contain better details, even in JPGs, though it can result in noise.

I also have set my K10D to 1/3 stops for EV (the default is 1/2 stops) as I believe this gave me better control.

The exposure will also depend on the metering mode used.

I'm using multisegment metering, EV steps set to 1/3 too and PEF+JPEGs... Lots of new things to learn with the digital!

This could quite possibly be down to the subject matter, can you post some examples?

When you process RAW files on the computer you can usually extract more detail from shadow (black) areas, however, the highlight (white) areas are lost for good unfortunately.

As was said above, if you're shooting JPEGs for use "straight-from-the-camera" then you may wish to lift the exposure levels and sacrifice the extreme highlights in favour of a better balanced overall picture. This is the way most point-and-shoot digital cameras work as they (mostly) don't have a RAW facility and expect the user to be less critical of their pictures. I'm sure with practice you'll find a way of working that suits you, but I think it's safe to say that your camera is working as it is intended to.

Not if they really are completely black. But that rarely happens with a moderate amount of under-exposure (especially at the RAW level), whereas blown highlights happen often with only a slight amount of over-exposure.

Also the appearance of black or featureless shadows can look less offensive than blown highlights, though of course this depends on the subject and your taste.

total blackness is much less likely to occur than total burnout.

You can prove this for yourself by setting the camera to -3ev compensation. Shoot raw and load the shot into your raw converter of choice. By thumping the exposure right up you'll get most if not all of the detail back, though it will be noisy.

do the same thing the other way around and the result wont look half as good.

Hello Everybody,

I've read all your posts with interest, and it seems that some of what you are saying is not particularly true - or accurate.
I constantly read and learn as much as I can on digital photography, and I recently read this article on a magazine. It actually states the exact opposite of what Hefty was saying about the amount of detail in Shadows / highlights. So I'm taking some time, and writing for you the entire article... Please remember that it only regards RAW files, and not Jpegs, but then again, if you are serious about your K10, you surely wouldn’t want to loose the possibility of shooting in RAW..

Hope you enjoy this!

Nathan


Digital SLR Photography - June 2008
"Exposing to the Right"

First thing “exposing to the right” is fast becoming a widely accepted approach to help maximise image quality – although it applies if you shoot in RAW. With this technique you effectively push exposure settings as close to the overexposure as possible without actually clipping the highlights. The result is a histogram with the majority of pixels grouped to the right of midpoint – hence the name “expose to the right”.
The logic of this approach can only be understood once you appreciate that CCD and CMOS sensors count light photons in a linear fashion. Linear capture has important implications on exposure. Most digital SLR’s record a 12 bit image, capable of recording 4096 total values over 6 stops. However, whilst you might automatically think that each f-stop of the 6 stop range would record an equal amount of the tonal value total, this is not so. The level corresponds exactly to the number of photons captured, so in reality, each stop records half the light of the previous one. For example, half of the levels are devoted to the brightest stop (204, half of the remainder (1024 levels) are devoted to the next stop and so on. As a result, the last, and darkest of the 6 stops, only boasts 46 levels. At first this might seem very confusing and the relevancy not obvious. However, in simple terms, what it means is if you do not properly use the right side of the histogram, - which represents the majority of tonal values, you are wasting up to half of the available encoding levels. So if you deliberately underexpose to ensure detail is retained in the highlights – a common practice among digital photographers – you are potentially loosing a large percentage of the data that can be captured. Noise can be significantly increased as a result – particularly to mid-tones and shadows. This is because, when you open up the shadow areas during RAW conversion, in order to lighten an image, you have to spread the 64 levels in the darkest stop over a wider tonal range. So, whilst it remains important not to overexpose images to the degree where the value for pure white is blown, “exposing to the right” is a logical approach to exposure. Whilst the method needs applying with care, and relies heavily on using the histogram to avoid “clipping”, image quality will be maximised. When using this approach you may notice that images appear light and washed out on the LCD monitor. Don’t worry; colour and contrast can be quickly restored during RAW conversion. “Exposing to the right” is an advanced technique that many experienced photographers are still getting used to, so don’t worry if you find it confusing.

ETTR is correct - if you want to retain the whole contrast range.

Normally you use ETTR when using fully manual mode. And in that circumstance you can choose to decide which highlights to sacrifice for shadow detail.

However, when the camera makes the judgment, it tries to keep the majority of the highlights in. I, for one, appreciate this - especially when using fill-in flash with the AF360/540. Wedding dresses are a particularly troublesome area for photographers... and with the K10D every shot had the dress detail fully there (which couldn't be said of my mate's 30D )

Blown highlights are completely unrecoverable (unless you are an absolute master at Photoshop and can generate detail from nothing).

Dark shadows, as has been pointed out, can be recovered.


If you want the maximum amount of colour detail, then you should switch to fully manual mode and ETTR

Matt

Well I'll need to have a closer look at the ETTR theory calculations...

About the 'underexposed' photos: they're not underexposed, they just look underexposed when looking at less than 50% of its size. Though there were some 'very dark' areas on the image (with no exposure correction) the sky showed shadows and was not completely white nor grey (rainy day). I'll try to post the photos to my picasaweb in PEF format.

Definitely there is a lot to learn.

Hold on. The actual advice given in the article -- to expose to the right, without clipping -- does NOT contradict what any of us (including Hefty) have said here.

The article is somewhat self-contradictory, and confusing to you, because it seems to criticise those who expose to preserve the highlights by "under-exposing". But often (depending on the scene) if you don't "under expose", you WILL clip the highlights. They advise that you DON'T clip the highlights. So that's a contradiction.

The article was not clear enough. What I think they mean to criticise is the practise of going too far - under-exposing by an amount that is more than necessary to preserve the highlights. In that case, yes, you can end up with more noise than you would otherwise.

Not everyone will want or need to "expose to the right" anyway, particularly those who won't be doing PP and don't want unnaturally bright pictures. And there is less need for this method when the ISO setting is low.

It has also been said on this thread that you can choose to clip the highlights deliberately, sacrificially, when it wouldn't be possible to get detail in the lowlights otherwise.