[rfortson]

That "other site" has a blog posting about the extra headroom to be gained from shooting raw. As an added bonus, they use the K200D as their example.

dpreview.com Editorial blog : Digital Photography Review

The jpeg image they use is perfectly usable, but they do a nice job of highlighting (pun intended) the detail you can recover by shooting in raw.

[Canada_Rockies]

Thanks for the link - it was interesting.

[stevebrot]

Very nice discussion of a complex subject.

I read a few of the comments and agree with the first. They should have picked a subject that did not include large areas that are supposed to be blindingly white. They should also have picked a camera that has a reputation for blowing highlights (say something in the lower end of the N*kon line).

Of course, they may have picked the Pentax because it really shines in dynamic range and preservation of highlights.

Steve

[PentaxPoke]

That was a nice discussion of benefits of raw for recovering from blown highlights. I think it is even more beneficial to eliminate the need to preset white balance (WB). Especially for people who shoot indoors, the auto WB can often be incorrect (especially with fluorescent lights). This can result in a more subtle but worse problem than the one shown. That is, incorrect WB can cause one of the RGB channels to saturate, even though the luminosity might not. I did a test a while ago to demonstrate this. It convinced me to always shoot raw. Then at shooting time, the only thing I need to think about is exposure and focus:
RAW vs jpeg WB test - a set on Flickr

[rfortson]

That was a nice discussion of benefits of raw for recovering from blown highlights. I think it is even more beneficial to eliminate the need to preset white balance (WB). Especially for people who shoot indoors, the auto WB can often be incorrect (especially with fluorescent lights). This can result in a more subtle but worse problem than the one shown. That is, incorrect WB can cause one of the RGB channels to saturate, even though the luminosity might not. I did a test a while ago to demonstrate this. It convinced me to always shoot raw. Then at shooting time, the only thing I need to think about is exposure and focus:
RAW vs jpeg WB test - a set on Flickr

Yes, I remember your test. Well done, also. I suspect (but am too lazy to test) that even with raw, there may be some extreme WB situations (like if you accidently tried the wrong preset) where the image is unrecoverable even in raw, sort of like totally blowing the highlights. However, I've never run into it yet. I think AWB keeps me in the safe zone anyway, and it's just easier/one less thing to worry about.

[PentaxPoke]

The place where I seem to have trouble with auto WB on the k10d and k20d is in certain flourescent lighting conditions like school gyms, and other indoor settings with mixed lighting types. Now that I shoot raw, I don't even care about WB. I set it when I download the pictures to the computer. One problem I have never had with the k10d or k20d is blown highlights. Pentax metering on those two cameras seems to do a good job of preventing blown highlights.

[Canada_Rockies]

The place where I seem to have trouble with auto WB on the k10d and k20d is in certain flourescent lighting conditions like school gyms, and other indoor settings with mixed lighting types. Now that I shoot raw, I don't even care about WB. I set it when I download the pictures to the computer. One problem I have never had with the k10d or k20d is blown highlights. Pentax metering on those two cameras seems to do a good job of preventing blown highlights.

And this results in some strange test comments from those who are used to systems that do not retain the highlights as well as Pentax does. I am perfectly happy with Pentax metering. I don't use [+/-] at all - I prefer manual if I am tinkering with the exposure, or just bracket - it's so easy on my k10d and my MZ-S.

[felix68]

Yes, I remember your test. Well done, also. I suspect (but am too lazy to test) that even with raw, there may be some extreme WB situations (like if you accidently tried the wrong preset) where the image is unrecoverable even in raw, sort of like totally blowing the highlights. However, I've never run into it yet. I think AWB keeps me in the safe zone anyway, and it's just easier/one less thing to worry about.

No, your suspicions are not correct. Raw files are a direct dump of the sensor data without any processing applied. White balance is an adjustment that is applied by the camera during processing so there is no way to screw up a raw file with the wrong WB. The camera saves the setting so that you can refer to it later (Pentax Photo Lab recalls what the setting was as a starting point), and it applies that setting to the preview image but that is it.
Jpegs, on the other hand are really easy to screw up with WB to the point of being unrecoverable. This is the biggest drawback of jpeg, IMHO. The drawback of raw is processing time on the back end.

[PentaxPoke]

Raw files are a direct dump of the sensor data without any processing applied. White balance is an adjustment that is applied by the camera during processing so there is no way to screw up a raw file with the wrong WB. The camera saves the setting so that you can refer to it later (Pentax Photo Lab recalls what the setting was as a starting point), and it applies that setting to the preview image but that is it.
Jpegs, on the other hand are really easy to screw up with WB to the point of being unrecoverable. This is the biggest drawback of jpeg, IMHO. The drawback of raw is processing time on the back end.

This is true. The only camera settings that affect the raw file are the exposure settings (ISO, shutter speed, f#) and focus. That is the beauty of raw. You don't have to think about anything else when you are taking a shot. Everything else(white balance, sharpness, saturation, etc. etc. etc) doesn't matter when you take the picture. You can set those at your leisure when you download the images to the computer. raw processing time really is minimal. I much prefer it to taking time thinking about what WB, and the miriad of other processing settings are best at the time the photo is taken. Why waste time setting those with the chance of being wrong when you are trying to take the picture? With raw, if your exposure and focus are correct, then nothing else matters, and you can get any "look" you want when you do the jpeg conversion.

[Marc Sabatella]

No, your suspicions are not correct. Raw files are a direct dump of the sensor data without any processing applied.

This is true in general, but people who have carefully examined the actual contents of Pentax RAW files have discovered a few exceptions:

- On the *ist series (and that series only, if I recall correctly), WB *is* partially applied to the RAW data. Not in the sense of totally correcting the color, but more some sort of shift to avoid clipping. I forget the specifics, but it seemed to have been proved pretty conclusively. The gist of it was that using AWB or the "wrong" setting wouldn't screw anything up - you'd be getting the results straight from the ADC - but actually choosing the tungsten present would result in, hmm, maybe less gain on the amplifiers on the red channel to avoid clipping? Something like that, anyhow. The effect was small but measurable, and only on the *ist series (or perhaps K100D/K110D too). Nothing like the huge difference WB makes for JPEG.

- On the K20D, *some* NR is applied at ISO 1600 even if you turn off the high ISO NR option.

- On all models, dark frame subtraction (slow shutter speed NR) is performed even in RAW

There is also the possibility that on some camera models, things like D-range may be implemented such that the results are actually reflected in the RAW file.

[felix68]

This is true in general, but people who have carefully examined the actual contents of Pentax RAW files have discovered a few exceptions:

- On the *ist series (and that series only, if I recall correctly), WB *is* partially applied to the RAW data. Not in the sense of totally correcting the color, but more some sort of shift to avoid clipping. I forget the specifics, but it seemed to have been proved pretty conclusively. The gist of it was that using AWB or the "wrong" setting wouldn't screw anything up - you'd be getting the results straight from the ADC - but actually choosing the tungsten present would result in, hmm, maybe less gain on the amplifiers on the red channel to avoid clipping? Something like that, anyhow. The effect was small but measurable, and only on the *ist series (or perhaps K100D/K110D too). Nothing like the huge difference WB makes for JPEG.

- On the K20D, *some* NR is applied at ISO 1600 even if you turn off the high ISO NR option.

- On all models, dark frame subtraction (slow shutter speed NR) is performed even in RAW

There is also the possibility that on some camera models, things like D-range may be implemented such that the results are actually reflected in the RAW file.

I didn't know about WB changes in the *istD. The rest is not WB related and that was what I was responding to. Should have made that more clear.

[jbinpg]

<snip>

- On all models, dark frame subtraction (slow shutter speed NR) is performed even in RAW

I think you meant to say:

"On all models, dark frame subtraction (slow shutter speed NR) is performed even in RAW...and in the K20D, you don't have a choice whether to apply DFS anyway."

Sorry, just had to get that off my chest.

Jack

[rfortson]

This is true in general, but people who have carefully examined the actual contents of Pentax RAW files have discovered a few exceptions:

- On the *ist series (and that series only, if I recall correctly), WB *is* partially applied to the RAW data. Not in the sense of totally correcting the color, but more some sort of shift to avoid clipping..

This is what I was thinking about. I've noticed that white balance shifts the various R, G, and B channels (as it should), but if one of the channels is shifted too far right, I was thinking you potentially could reach an unrecoverable situation, like in exposure.

As I also said, I've never run across this and this was more of a mind exercise. Since I've never really seen anyone else have this problem, then I guess I'm just worrying for no reason....not that I really worry about it anyway.

[Marc Sabatella]

As I also said, I've never run across this and this was more of a mind exercise. Since I've never really seen anyone else have this problem, then I guess I'm just worrying for no reason....not that I really worry about it anyway.

When I had my DS, I *did* change WB to tungsten when shooting indoors, both because of concerns over the possibility of red channel clipping (which realistically happens anyhow in tungsten lighting unless you underexpose), but perhaps more because the view on the LCD looked *so* bad when using AWB. But I wasn't always consistent about it. Whatever effect the in-camera WB might have had in theory, I can't say that it made a noticeable difference.

The K200D RAW data is *not* affected by WB in the same way the DS supposedly was, and the AWB works enough "better" in tungsten light that I no longer mess with WB in camera, ever.

[Ron Boggs]

rfortson, nice reference to the "what me worry" kid...