All Forum Comments
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 08/01/2004 - 13:32
Hello Guest,
It is the camera doing the metering - in P-TTL flash mode. That is, a preflash is fired from which the camera does the metering, adjusts the exposure appropriately then fires the main flash.
Regarding the metering, it is true that spot metering could affect things, and indeed it does (depending on the subject of course).
Yes, modern lenses are being used. Yes, the guidelines are being followed.
Overall, the *istD produces beautiful results, and the exposure and control are far better than the film based SLRs I already use. I am very pleased with the camera itself and the images it produces. It is been a real help to stick with the equipment I had invested in, and to continue to have digital and film compatibility (maybe my gallery will help in showing that).
However, this is not a simple problem. In the real world, the *istD seems to "randomly" mis-expose some shots. I am of the opinion that it is not random at all. Yet the factors affecting the issue are complex - much more complex than a film camera. So far, these are the findings in trying to work out what is going on (mostly from my experiments, but also from others' comments):
* It appears the metering issue is not just down to the flash. From what I can see, there are differences in exposure (due to factors that I haven't entirely pinned down yet) - if the non-flash exposure is under, then the flash exposure appears to be over exposed (and vice-versa).
* Auto white balance makes testing confusing, since the camera is trying to correct the image post-shutter close.
* Auto sensitivity is a great feature. In some 'real world' situations, I thought it was this feature creating the flash issue. However, after testing, I cannot say conclusively either way. The issue certainly occurs without AS enabled.
* The ISO setting has a bearing
* The zoom setting on the lense has a bearing
* The lens itself has a bearing.
Some theories as to the cause:
* The exposure sensor is sensitive to non-visible wavelengths, or sensitive in selected wavelengths.
* The camera is trying to compensate for something (zoom, lense type etc) which is throwing the exposure off with the kit I have (though I have tested a genuine Pentax lens with similar results)
* There's a little bug in the firmware (indeed, as a firmware developer myself, we have a saying that "the size of the fault is inversely proportional to the size of the bug". ie. make a small mistake in coding and the result will probably look catastrophic. We call this "Golding's law" after the MD who coined the observation.)
To be fair, this is Pentax's first commercial DSLR, and I didn't for one minute expect perfect firmware on the first release. As I mentioned, the issues are not simple ones, and the factors are many, and exposure to the wider market always yields situations that haven't been coded correctly in the firmware.
All that said, I am confident that Pentax will resolve the issue, and maybe add some extra features in the firmware too In the meantime I am taking a little more care (bracketing exposure, double checking results etc) when taking important shots. I wouldn't let this issue put you off considering the camera - it has some great features that stand it apart from the mainstream competition (small body, large clear viewfinder etc) that should not be ignored
Hope that clarifies things a little
Matt
It is the camera doing the metering - in P-TTL flash mode. That is, a preflash is fired from which the camera does the metering, adjusts the exposure appropriately then fires the main flash.
Regarding the metering, it is true that spot metering could affect things, and indeed it does (depending on the subject of course).
Yes, modern lenses are being used. Yes, the guidelines are being followed.
Overall, the *istD produces beautiful results, and the exposure and control are far better than the film based SLRs I already use. I am very pleased with the camera itself and the images it produces. It is been a real help to stick with the equipment I had invested in, and to continue to have digital and film compatibility (maybe my gallery will help in showing that).
However, this is not a simple problem. In the real world, the *istD seems to "randomly" mis-expose some shots. I am of the opinion that it is not random at all. Yet the factors affecting the issue are complex - much more complex than a film camera. So far, these are the findings in trying to work out what is going on (mostly from my experiments, but also from others' comments):
* It appears the metering issue is not just down to the flash. From what I can see, there are differences in exposure (due to factors that I haven't entirely pinned down yet) - if the non-flash exposure is under, then the flash exposure appears to be over exposed (and vice-versa).
* Auto white balance makes testing confusing, since the camera is trying to correct the image post-shutter close.
* Auto sensitivity is a great feature. In some 'real world' situations, I thought it was this feature creating the flash issue. However, after testing, I cannot say conclusively either way. The issue certainly occurs without AS enabled.
* The ISO setting has a bearing
* The zoom setting on the lense has a bearing
* The lens itself has a bearing.
Some theories as to the cause:
* The exposure sensor is sensitive to non-visible wavelengths, or sensitive in selected wavelengths.
* The camera is trying to compensate for something (zoom, lense type etc) which is throwing the exposure off with the kit I have (though I have tested a genuine Pentax lens with similar results)
* There's a little bug in the firmware (indeed, as a firmware developer myself, we have a saying that "the size of the fault is inversely proportional to the size of the bug". ie. make a small mistake in coding and the result will probably look catastrophic. We call this "Golding's law" after the MD who coined the observation.)
To be fair, this is Pentax's first commercial DSLR, and I didn't for one minute expect perfect firmware on the first release. As I mentioned, the issues are not simple ones, and the factors are many, and exposure to the wider market always yields situations that haven't been coded correctly in the firmware.
All that said, I am confident that Pentax will resolve the issue, and maybe add some extra features in the firmware too In the meantime I am taking a little more care (bracketing exposure, double checking results etc) when taking important shots. I wouldn't let this issue put you off considering the camera - it has some great features that stand it apart from the mainstream competition (small body, large clear viewfinder etc) that should not be ignored
Hope that clarifies things a little
Matt
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 08/01/2004 - 06:29
Thanks for that Malcolm!
In my tests, when you have AWB off, the results are much more predictable... in fact, I did the Pentax-FA 28-70mm f4 test twice with exactly the same results (once with multi-segment, once with centre weighted).
However, since the problem appears to be zoom related, as well as lense related, as well as ISO related, and the AWB compounds the problem, in normal use it would appear a 'random' event
My efforts were to isolate the Tokina lenses I have from the equation, and to at least pin it down somewhere.
The flash issue is probably partly due to the no-flash exposure issue. If the *istD is not getting the normal exposure right, how can we expect it to get the flash setting right? From what I've seen, if the ambient conditions produce an underexposed image, then the flash will be over-exposed. Conversely, if the ambient conditions produce an overexposed image (in relation to the other images), then the flash will be correspondingly under exposed.... At least that's what I think
(I'm glad I am not the only firmware coder that has his work cut out )
Matt
In my tests, when you have AWB off, the results are much more predictable... in fact, I did the Pentax-FA 28-70mm f4 test twice with exactly the same results (once with multi-segment, once with centre weighted).
However, since the problem appears to be zoom related, as well as lense related, as well as ISO related, and the AWB compounds the problem, in normal use it would appear a 'random' event
My efforts were to isolate the Tokina lenses I have from the equation, and to at least pin it down somewhere.
The flash issue is probably partly due to the no-flash exposure issue. If the *istD is not getting the normal exposure right, how can we expect it to get the flash setting right? From what I've seen, if the ambient conditions produce an underexposed image, then the flash will be over-exposed. Conversely, if the ambient conditions produce an overexposed image (in relation to the other images), then the flash will be correspondingly under exposed.... At least that's what I think
(I'm glad I am not the only firmware coder that has his work cut out )
Matt
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 24/12/2003 - 14:43
Hi All!
Ok, now I am more confused. I decided to go back and try a genuine Pentax-FA 28-70mm F4 and do some more comparisons and tests.
Here's what I did: take the *istD with lense on a tripod looking at an evenly lit magnolia wall. Set focus to manual and infinity. Try different ISO settings at each zoom setting of 28, 35, 50, 70. White balance set at neutral white fluorescent, and pattern metering.
The first set yielded half stop variations at the same ISO setting between each zoom. Ah, you say, but the shutter speed shown in the EXIF is not necessarily accurate. True, the EXIF only shows the nearest "standard" shutter speed. BUT, Photoshop shows histogram shifts - ISO200 not bad, ISO400 has a shift at 70mm, ISO800 shows 28&35 shifted down from 50&70, ISO1600 shows more shifts.
After resetting the camera (using the menu function), the results were identical, although the Av/Tv combinations used were slightly different, and seemed more consistent. Having laid out a composite image of zoom vs ISO with two layers (one before reset, one after) showed no significant differences.
To summarise:
It appears that genuine lenses have issues too - so it points to the body.
The differences vary greatly between different lenses and different zooms (different lenses have very different colour responses - some are more sensitive to low level light etc). Here are the results: http://www.pbase.com/image/24411186/large
I am positive that this can be resolved... I would just like to pin it down. It would be interesting to know more about how the exposure is read - presumably in a similar way to the MZ-S with a separate sensor - and what the interplay is between the zoom, f-stop, lense ID, ISO and other settings.
Any comments or suggestions would be greatly appreciated
Matt
Ok, now I am more confused. I decided to go back and try a genuine Pentax-FA 28-70mm F4 and do some more comparisons and tests.
Here's what I did: take the *istD with lense on a tripod looking at an evenly lit magnolia wall. Set focus to manual and infinity. Try different ISO settings at each zoom setting of 28, 35, 50, 70. White balance set at neutral white fluorescent, and pattern metering.
The first set yielded half stop variations at the same ISO setting between each zoom. Ah, you say, but the shutter speed shown in the EXIF is not necessarily accurate. True, the EXIF only shows the nearest "standard" shutter speed. BUT, Photoshop shows histogram shifts - ISO200 not bad, ISO400 has a shift at 70mm, ISO800 shows 28&35 shifted down from 50&70, ISO1600 shows more shifts.
After resetting the camera (using the menu function), the results were identical, although the Av/Tv combinations used were slightly different, and seemed more consistent. Having laid out a composite image of zoom vs ISO with two layers (one before reset, one after) showed no significant differences.
To summarise:
It appears that genuine lenses have issues too - so it points to the body.
The differences vary greatly between different lenses and different zooms (different lenses have very different colour responses - some are more sensitive to low level light etc). Here are the results: http://www.pbase.com/image/24411186/large
I am positive that this can be resolved... I would just like to pin it down. It would be interesting to know more about how the exposure is read - presumably in a similar way to the MZ-S with a separate sensor - and what the interplay is between the zoom, f-stop, lense ID, ISO and other settings.
Any comments or suggestions would be greatly appreciated
Matt
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 24/12/2003 - 12:55
Don & Malcolm,
I've managed to rattle off a few test shots using my three Tokina zooms, the AF360FGZ, and comparing in PS-CS (live histogram has been useful here)...
First interesting thing: Pentax Photo Browser reports the three lenses as follows:
A) Tokina 20-35mm = "smc PENTAX-F 35-70mmF3.5-4.5"
B) Tokina 28-70/f2.8 = "smc PENTAX-F 28-30mmF3.5-4.5"
C) Tokina 24-200/f3.5 = "-- -- --"
Now, the lense ID (I suspect) is probably just an ID number, and I don't think has too much to do with the issue.
Second big thing when comparing images is to force the White Balance. When you have AWB, the different colour characteristics of the lenses changes the way the *istD compensates for white balance. When using AWB I was getting lense C as the brightest, A as the next, and B as the darkest. However, after testing with Flash WB lenses A and C produced equivalent images (with slight colour shifts), while B was a good stop darker.
Now, I tested the following flash combinations:
(1) Straight flash, (2) Manual zoom flash, (3) Diffused, (4) Auto sensitivity and auto flash zoom. Note, 1-3 were on P-mode, ISO200, non auto sensitivity.
Why, oh why, the AF360FGZ shows the zoom size 1.5x shorter that the real zoom, I don't know (it shows 13mm when the 35mm format is 20mm - it should have shown 30mm, or just 20mm, on the AF360)! Since the EXIF information and the zoom markings are all in 35mm format, why didn't they stick to that on the AF360?? However, judging by the noises the zoom makes, it appears that this is only an incorrect display, rather than an incorrect zoom.
All sets (barring the AWB issue mentioned) produced consistent results. Lenses A and C produced good balanced images, and B was underexposed.
Now, it's interesting to note that Malcolm had the problem with a Pentax 28-80, and my lense is ID'd as a Pentax 28-80
Just to throw a real spanner in the works, when taking non-flash images (with white balance on sunny), lense B produced consistently brighter images! Lenses A and C had seriously long "tails" on their histogram - stretching about 1/3 across the histogram. (The order, from bright to dark was B, C, A, with C about 1/3 stop brighter than A, and B about 1 stop brighter than B)
Exposure information:
Available light: A=1"/f4.5 (EV4.34), B=0.7"/f2.8 (EV3.49), C=1"/f4.0 (EV4.00)
Flash: A=45/f4.5, B=45/f3.5, C=45/f4.0
Why the same scene can produce such a wide range of available-light exposures is a mystery :
So, it looks like there's something a little strange going on. Whether it is down to colour shifts in the lense, or (more likely I think) a firmware issue, let's hope we and Pentax can get this sorted out soon
Matt
I've managed to rattle off a few test shots using my three Tokina zooms, the AF360FGZ, and comparing in PS-CS (live histogram has been useful here)...
First interesting thing: Pentax Photo Browser reports the three lenses as follows:
A) Tokina 20-35mm = "smc PENTAX-F 35-70mmF3.5-4.5"
B) Tokina 28-70/f2.8 = "smc PENTAX-F 28-30mmF3.5-4.5"
C) Tokina 24-200/f3.5 = "-- -- --"
Now, the lense ID (I suspect) is probably just an ID number, and I don't think has too much to do with the issue.
Second big thing when comparing images is to force the White Balance. When you have AWB, the different colour characteristics of the lenses changes the way the *istD compensates for white balance. When using AWB I was getting lense C as the brightest, A as the next, and B as the darkest. However, after testing with Flash WB lenses A and C produced equivalent images (with slight colour shifts), while B was a good stop darker.
Now, I tested the following flash combinations:
(1) Straight flash, (2) Manual zoom flash, (3) Diffused, (4) Auto sensitivity and auto flash zoom. Note, 1-3 were on P-mode, ISO200, non auto sensitivity.
Why, oh why, the AF360FGZ shows the zoom size 1.5x shorter that the real zoom, I don't know (it shows 13mm when the 35mm format is 20mm - it should have shown 30mm, or just 20mm, on the AF360)! Since the EXIF information and the zoom markings are all in 35mm format, why didn't they stick to that on the AF360?? However, judging by the noises the zoom makes, it appears that this is only an incorrect display, rather than an incorrect zoom.
All sets (barring the AWB issue mentioned) produced consistent results. Lenses A and C produced good balanced images, and B was underexposed.
Now, it's interesting to note that Malcolm had the problem with a Pentax 28-80, and my lense is ID'd as a Pentax 28-80
Just to throw a real spanner in the works, when taking non-flash images (with white balance on sunny), lense B produced consistently brighter images! Lenses A and C had seriously long "tails" on their histogram - stretching about 1/3 across the histogram. (The order, from bright to dark was B, C, A, with C about 1/3 stop brighter than A, and B about 1 stop brighter than B)
Exposure information:
Available light: A=1"/f4.5 (EV4.34), B=0.7"/f2.8 (EV3.49), C=1"/f4.0 (EV4.00)
Flash: A=45/f4.5, B=45/f3.5, C=45/f4.0
Why the same scene can produce such a wide range of available-light exposures is a mystery :
So, it looks like there's something a little strange going on. Whether it is down to colour shifts in the lense, or (more likely I think) a firmware issue, let's hope we and Pentax can get this sorted out soon
Matt
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 22/12/2003 - 16:05
(In the meantime...)
Here is an interesting link regarding the AF360 and *istD combination:
http://www.mail-archive.com/pentax-discuss@pdml.net">target="blank">http://www.mail-archive.com/pentax-discuss@pdml.net/msg149194.html
(focal length appeared wrong on flash, manually set zoom and exposure okay)
Something else to check...
Matt
Here is an interesting link regarding the AF360 and *istD combination:
http://www.mail-archive.com/pentax-discuss@pdml.net">target="blank">http://www.mail-archive.com/pentax-discuss@pdml.net/msg149194.html
(focal length appeared wrong on flash, manually set zoom and exposure okay)
Something else to check...
Matt
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 21/12/2003 - 22:09
Don,
I've spent a little time this afternoon trying to test out the AF360... I have three Tokina zooms: AT-X 24-200, AT-X 28-70/f2.8, and 20-35mm (not ATX)
I tried to take the same scene each time, using P-TTL, Auto sensitivity, ISO800 (starting at), and 35mm on each lense. I found the 28-70/f2.8 produced underexposed shots (nearly 1 stop), while the 20-35mm produced the brightest. I tried this in Depth, MTF, and Normal program lines - each the same. However, the 20-35mm did only use f5.6 for depth (which is weird) while the other two used f10 and f11 (I haven't double checked the ISO choice for each shot, since it was on auto sensitivity - it's difficult with my lively kids running around )
Going back to non-flash, the 28-70/f2.8 produced the brighter image!
Soooo confusing! And I am sure it is not down to the Tokina lenses per-se. Maybe the *istD is reading something from the lense and making other judgments I am unaware of.
In addition, I did find out that putting the diffuser on and trying 20mm often produced badly underexposed images. I need to try this with and without auto-ISO, and with different subjects.
This is a really knotty problem, and I am going to have to set up a controlled experiment with a static subject.
I am not sure that it is down to any particular lense problem, but more of a lense-to-body-communication and algorithm issue. When I had the problem with the 20mm, AF360 wireless I tried to repeat the experiment the next day and all was fine! It may even involve the subject - perhaps the contrast ratio or maybe even lense flare of some kind.
Still, it's a good challenge
Matt
I've spent a little time this afternoon trying to test out the AF360... I have three Tokina zooms: AT-X 24-200, AT-X 28-70/f2.8, and 20-35mm (not ATX)
I tried to take the same scene each time, using P-TTL, Auto sensitivity, ISO800 (starting at), and 35mm on each lense. I found the 28-70/f2.8 produced underexposed shots (nearly 1 stop), while the 20-35mm produced the brightest. I tried this in Depth, MTF, and Normal program lines - each the same. However, the 20-35mm did only use f5.6 for depth (which is weird) while the other two used f10 and f11 (I haven't double checked the ISO choice for each shot, since it was on auto sensitivity - it's difficult with my lively kids running around )
Going back to non-flash, the 28-70/f2.8 produced the brighter image!
Soooo confusing! And I am sure it is not down to the Tokina lenses per-se. Maybe the *istD is reading something from the lense and making other judgments I am unaware of.
In addition, I did find out that putting the diffuser on and trying 20mm often produced badly underexposed images. I need to try this with and without auto-ISO, and with different subjects.
This is a really knotty problem, and I am going to have to set up a controlled experiment with a static subject.
I am not sure that it is down to any particular lense problem, but more of a lense-to-body-communication and algorithm issue. When I had the problem with the 20mm, AF360 wireless I tried to repeat the experiment the next day and all was fine! It may even involve the subject - perhaps the contrast ratio or maybe even lense flare of some kind.
Still, it's a good challenge
Matt
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 20/12/2003 - 16:19
Don,
Thanks for the information. I have been trying to work out what goes on with the flash too, and I do remember having wireless probs with a Tokina 20-35mm - I took a few shots all underexposed (but corrected them in PS).
Armed with that knowledge, I'll go back and retry all lense/flash combinations to try and sort this thing out once and for all
You never know, it may not be the lense, but firmware. Hang on in there...
Matt
PS. If you haven't tried Photoshop CS - it's superb There are some reallly, really nice features for photo editing. Let me know if anyone wants a quick review - we have the full Creative Suite CS Premium here, such a killer package
Thanks for the information. I have been trying to work out what goes on with the flash too, and I do remember having wireless probs with a Tokina 20-35mm - I took a few shots all underexposed (but corrected them in PS).
Armed with that knowledge, I'll go back and retry all lense/flash combinations to try and sort this thing out once and for all
You never know, it may not be the lense, but firmware. Hang on in there...
Matt
PS. If you haven't tried Photoshop CS - it's superb There are some reallly, really nice features for photo editing. Let me know if anyone wants a quick review - we have the full Creative Suite CS Premium here, such a killer package
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 19/12/2003 - 09:24
I've been trying to get the AF360 to be underexposed...
If the power switch is on the middle position, and the wireless mode (the small lower switch) is set to "C", then it will always be underexposed.
If the contacts between the flash and the *istD are not clean enough, there is a possibility of under/over exposure.
If the mode on the AF360 is not at P-TTL then is could be improperly exposed. Conversely, when in P-TTL, the flash can set the proportional power, so it should be at 0.0 to be properly exposed. (This is very useful indeed to control the flash amount and balance with ambient light.)
The metering mode also affects the flash - in spot metering, the flash will meter only for the centre, and could adversely affect the rest of the image (unless that's what you want ).
In my case, I have had a few situations where things didn't work out as expected. When I tried to work out the cause, it disappeared. If I can pin it down to anything in particular, I will, of course, let everyone (including Pentax) know
Matt
If the power switch is on the middle position, and the wireless mode (the small lower switch) is set to "C", then it will always be underexposed.
If the contacts between the flash and the *istD are not clean enough, there is a possibility of under/over exposure.
If the mode on the AF360 is not at P-TTL then is could be improperly exposed. Conversely, when in P-TTL, the flash can set the proportional power, so it should be at 0.0 to be properly exposed. (This is very useful indeed to control the flash amount and balance with ambient light.)
The metering mode also affects the flash - in spot metering, the flash will meter only for the centre, and could adversely affect the rest of the image (unless that's what you want ).
In my case, I have had a few situations where things didn't work out as expected. When I tried to work out the cause, it disappeared. If I can pin it down to anything in particular, I will, of course, let everyone (including Pentax) know
Matt
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 04/12/2003 - 11:16
Hi Don Schoolmaster,
I too have the *istD and AF360 and general get beautiful results.
However, occasionally, I have had underexposed shots. I have been trying to pin it down, and I think I may have had the flash mode (on the *istD) set to wireless, rather than automatic/redeye/always (ie the little remote control icon appears on the *istD top LCD).
I would suggest setting up a little test - a still-life about 6ft away - and go through the flash and camera modes. If that still doesn't work, try the same setup with the built in flash alone - double check the internal flash is correctly exposed.
In the meantime, I will see if I can get the same results as you in some way to add anything to this.
Matt
I too have the *istD and AF360 and general get beautiful results.
However, occasionally, I have had underexposed shots. I have been trying to pin it down, and I think I may have had the flash mode (on the *istD) set to wireless, rather than automatic/redeye/always (ie the little remote control icon appears on the *istD top LCD).
I would suggest setting up a little test - a still-life about 6ft away - and go through the flash and camera modes. If that still doesn't work, try the same setup with the built in flash alone - double check the internal flash is correctly exposed.
In the meantime, I will see if I can get the same results as you in some way to add anything to this.
Matt
Comment by MattMatic posted on Pentax *istD underexposed with AF-360FGZ at 01/12/2003 - 07:39
No, not insulted
When I first decided to go digital, I considered the Dimage 7. Specification wise it looked great, but the handling - eurgh. Having a proper DSLR is just fantastic. Don't think I could go back now. Even film seems daunting having got used to the advantages of pure digital
Yes, some of the flashes have zoom (the ones ending FTZ - AF360FTZ etc). The zoom on the flash can be overridden, but this seems to make little, if any, difference to me.
The firmware is always upgradeable these days (the code is stored in a flash memory chip inside the camera). In developing the code for such devices you always need to throw new code at a device quickly to bugfix. Such benefits also extend to the end user too Normally a little Windows application will fire the data via USB, or you load up a CF card with the appropriate files on and shove it in the camera which upgrades itself. (That all assumes it is a firmware bug, and not something more fundamental.)
Matt