iirc the only difference in browser is that it can identify more lenses and bodies, and open PEFs from later cameras.

I seem to recall that Lab got a significant improvement between versions 2 and 3, but I can't remember exactly what the differences were.

sorry I can't be more help.

Curses! bang goes another Nobel prize! back to the drawing board I suppose......

lol, well if you win I expect a cut

6 13 26 31 35 and 42

Thanks George, that's very kind. Now, if only I could convert my understanding of the technology into photographic ability!

Simon, it is my experience that whatever Pentax pattern metering actually does works extremely well in almost all situations. My K10D tends to stay on pattern almost all the time.

They say a picture is worth 1000 words, so here is a 4000 word essay These are terrible photographs which I just snapped to demonstrate the difference between pattern and spot metering. All taken with the K10D and kit lens.




This is the pattern metering shot. You could argue that it's a touch underexposed, but it's broadly speaking correct. The important thing to note is that the meter reading remains constant if I change the composition while keeping similar proportions of wood, phone and wall.




Now I have flicked over to spot metering and changed the composition very slightly. The spot is on the table, which is slightly darker than a mid tone. The camera pulls up the exposure to brighten the table and the scene ends up slightly over exposed.




Moving again so that the spot covers the phone, we get a more drastic overexposure to try and bring the black phone up to mid grey.




finally putting the spot on the cream wall causes some underexposure to bring the brightness down to mid grey.

If I leave the camera on pattern mode, then none of these three is enough of a change in the composition to change the meter output at all, while in spot mode going from the phone to the wall gives a change of 3 full stops.

I hope this little demo helps to show what difference it's supposed to make. I certainly agree that in most situations pattern metering is your best option.

well that deppends very much upon what you expect.

At the end of the day, all three are designed to achieve the same end - to aid you in selecting the correct exposure for your photograph.

Metering is based upon the assumption that the average over the metering area will be a mid tone. For many images, this a fairly valid assumption, and the first metering systems simply applied this assumption directly. The whole scene was considered equal and an average brightness reading was used to determine the correct exposure.

Over time, various improvements have been applied to this system, to give a better "hit rate" of correct exposures. There are three modes present in the K10D, as you have observed, and they work as follows:

1. Pattern metering (is that the correct Pentax term? I get lost) - This is Pentax's most advanced metering mode. It splits the scene into 16 segments and uses cunning cleverness of which I have no real knowledge to decide how to weight those segments in its readout.

2. Centre Weighted Average - This was the forefunner of pattern metering. It assumes that the object of interest is somewhere near to the centre of the scene, so the reading is biased towards the reading from the centre. Note that the outside is still used, it is simply given a lower weighting.

3. Spot Metering - This mode is reserved for really hard lighting situations and will probably usually be used in conjunction with manual exposure (indeed on some older Pentax bodies the only way to engage spot metering was to first engage M mode). Metering is performed considering only a small spot in the centre of the scene, and the rest is ignored. On the K10D, the spot metering area corresponds roughly to the size of the central circle in the viewfinder.



If you point all three modes at a low contrast scene which really is roughly even brightness throughout, you'll get the same result from all three.


If, on the other hand, you select a complex scene with wide tonal variation then you will get different results. Spot metering particularly should give you a wide variation of meter reading as the spot moves from bright to dark areas

fantastic capture Chirpy!

Malards can be evil. Sometimes I suspect they intentionally engage in excessive antics to make up for the fact that they are so common.

you're going to have me down the building society taking out a mortgage for one of those lenses at this rate!

thanks Golf, I hope it makes sense!

the ones which really make me laugh are the phone cameras. 5 megapixels in a phone? give me a break!

The really annoying thing is that the consumer drive to push up the pixel counts is hurting image quality. For a while the major camera brands seemed to resist the megapixel madness, but it seems they have now given in. You can't blame them, they have to make what will sell, but I dread the day when my Lumix FX-7 breaks. I just don't know of a current compact camera which can equal its performance.

what you are seeing is essentially a limitation of compact camera design, and one of the reasons most "serious" photographers end up with a DSLR.

Modern compact digital cameras cram an awful lot of pixels onto a very small sensor, and that is a recipe for high ISO noise.

You can understand why if you think of light using the particle model.

In this model light is considered to be made up of particles called photons. These are just little packets of energy which wizz around at the speed of light.

Think of the sensor in your camera as a series of buckets aranged in a rectangle to catch the light particles. If you have to get 10 million buckets into a sensor 1/4 inch across, those buckets are very very small. Smaller buckets hold fewer photons.

Now, it turns out that all photons are the same size (physicists bear with me here, it's close enough to the truth for this discussion), and you can't cut them up. Further, it turns out that you can't count them particularly well.

Lets imagine for a minute that a single pixel in your compact camera holds 10 photons of light, plus or minus 1 (due to difficulties in counting).

A larger camera has a sensor 10 times the size with the same number of pixels. This cameras pixels will each hold 100photons, plus or minus 1.

The error in the measured value from the small sensor is +-10%, which is what causes noise. The larger sensor has an error of only +-1% giving far less noise.


How does this relate to ISO? simple. To get a higher ISO out of a given sensor you multiply up the results from it. In so doing you also multiply the errors, so that a noisey signal becomes even worse. Add to this the fact that you have to multiply the smaller signals from smaller sensors by larger factors to get the same ISO and you have a one way ticket to noise central.


This is all very simplified, and other factors are also important, but it's more or less correct.

From the point of view of image quality, Pentax and the other camera makers made a grave error in pushing compact cameras beyond 6MP.

I don't have any images for you, but Dave Coffin has written a very well regarded open source RAW conversion engine called DCRAW. The latest version supports the K20D, so he must have figured out the ins and outs of reading its PEF files.

http://cybercom.net/~dcoffin/dcraw/