[sTi]

It's in the section labeled Manufacturers Notes.

Since it's below 0 outside and I had no desire to get out and do anything, I spent a good while looking through my photos and reading the exif in PhotoME. My K5 regularly listed the SR Focal length as 45mm when I shot at 17mm. I began checking some other focal lengths and found similar results. There were also slightly off SR results at longer lengths but only by a couple of mm's and not enough to likely cause a problem. I also found some differences in other lenses too, but only by a very little. Then another problem surfaced. I went back and looked at some of the other data in those 17mm shots I checked earlier and they are now listing the SR length as 52mm!

Since these are all pretty sharp photos, I'm beginning to think this is a problem with PhotoME. Why would the data change? I also found instances where my DA 10-17 was listed as an "M" or "K" lens even though in other places in the data it is correctly listed as a DA. I think PhotoME isn't correctly reading the data. My AV software also has tried to block PhotoME as malware.

Rather than send a lens back, the real answer to all this is to follow the rules of photography and use a tripod if you want to shoot at slow shutter speeds. It's always a crap shoot under 1/30 and for longer lenses, the minimum is usually the focal length; don't try to shoot with a 200mm under 1/200, etc and even that is no guarantee.

I'm pretty confident this is no PhotoMe problem, and it's always possible to cross-check with other software. Of course all this will be a non-issue if you shoot at 1/100 or 1/50 seconds (or with flash), as these speeds are fast enough even with screwed-up SR. But try getting decent shots at 1/10 seconds with the Sigma @17 and e.g. a DA 18-55 or 18-135 @18 if you have one. With the Pentax lenses (which show the correct 18 or 19mm as SR focal length) this is no big deal, with the Sigma it's nearly impossible. As written above, it's actually easier with SR off.

---------- Post added 07-01-15 at 19:28 ----------

What is the difference between this an my older 17-70? I took a few test shots at 17mm 1/15 in my dim office and they looked plenty good....my K5 original is not the best at low light shooting at slow speeds, but they appeared decent using the Sigma SR.

Regards!

The 17-70 Contemporary is a completely new construction and lacks the in-lens stabilisation. Your lens is a previous version and most likely may not be affected at all, especially when you use the in-lens stabilizer, so no worries

---

Last edited by sTi; 01-07-2015 at 11:32 AM.

[Spodeworld]

I must be dense (no need to agree). I don't see a Manufacturer Notes section in Photome. I see a Manufacturer notes row in the Camera section, but it does not have this data in that row.

---------- Post added 01-07-15 at 01:52 PM ----------

BTW, I just took a number of shots of a mug at 17mm @ ISO 400 with SR on and off. The shutter speed was around 1/15". All the ones with SR on came out a little blurry. But when I turned it off, most of them were very sharp.

[sTi]

I must be dense (no need to agree). I don't see a Manufacturer Notes section in Photome. I see a Manufacturer notes row in the Camera section, but it does not have this data in that row.

Your density seems within normal range - Sorry, I tried with a sample K-3 JPG from the web and it seems this camera no longer writes this value in the metadata, at least I couldn't locate it. There are tons of (undocumented) values in the manufacturer notes, but it seems impossible to know which is which. So there seems no way to check it for this camera. But you could try a real-world test @17 or 21mm and 1/10 seconds. and see if the results look OK.

By the way, if you want to check your pics without installing PhotoMe, there's also http://regex.info/exif.cgi that lets you check the metadata online. It shows the same "SR focal length" value like PhotoMe, for my images so both seem to give trustworthy results.

[Spodeworld]

Just gave regex a try, and it doesn't seem to be burried in there.

I mentioned above that I did a few snaps at 17mm at about 1/15". The ones without SR were sharper....actually, very sharp.

Your density seems within normal range - Sorry, I tried with a sample K-3 JPG from the web and it seems this camera no longer writes this value in the metadata, at least I couldn't locate it. There are tons of (undocumented) values in the manufacturer notes, but it seems impossible to know which is which. So there seems no way to check it for this camera. But you could try a real-world test @17 or 21mm and 1/10 seconds. and see if the results look OK.

By the way, if you want to check your pics without installing PhotoMe, there's also http://regex.info/exif.cgi that lets you check the metadata online. It shows the same "SR focal length" value like PhotoMe, for my images so both seem to give trustworthy results.

[sTi]

[/COLOR]BTW, I just took a number of shots of a mug at 17mm @ ISO 400 with SR on and off. The shutter speed was around 1/15". All the ones with SR on came out a little blurry. But when I turned it off, most of them were very sharp.

Now you're talking! Thanks for the test, so similar results like for my copy on the K-5 II.
So far we have several copies of the lens on K-5, K-5II(s) and now K-3 that show this behaviour. That's enough for me now to contact Sigma and point them to this issue. Hopefully we'll see a firmware upgrade that solves this problem.

[Spodeworld]

Assuming there is a firmware upgrade, I assume that we would have to buy the dock to apply it.

[sTi]

Assuming there is a firmware upgrade, I assume that we would have to buy the dock to apply it.

It should also be possible to send it to a Sigma repair center, which will apply the new firmware free of charge during the warranty period.

[Spodeworld]

Not for those of us who bought it from a PF member! Anyway, probably not a bad idea to invest in the dock, assuming it will have additional use in the future.

It should also be possible to send it to a Sigma repair center, which will apply the new firmware free of charge during the warranty period.

BTW, post the message you send to Sigma. Maybe we can send the same message to local Sigma centers in our respective countries. Thanks!

Here are my results, and they were typical with multiple tries. I focused on the cup using AF.S:

17mm 1/15" no SR



17mm 1/15" SR on

[MadMathMind]

Some of my 17-70 shots are reported correctly. Many are not. About 33% of the ones I tried were correct. All the others were wrong. I only looked at 17-21mm for the most part, but quick checks on longer ones looked right. Then I decided to try out some other lenses:

Pentax DA35 f/2.4: Had SR focal lengths of 35, 36, 37

Pentax F50 f/1.7: 52mm across the board (most had SR turned off, whoops, and I no longer have this lens)

Pentax F 70-210 f/4-5.6: Consistently higher than actual. Had a 188 for actual 180, 121 for 118, etc. Doesn't seem to be too big a deal.

Pentax FA77 Limited: All looked correct at 77mm

Pentax DA*300: SR focal length of 292 across the board

Sigma 18-35 f/1.8 Art: All correct

Sigma 70-200 HSM II: Mostly all correct; had a 105 for an FL of 106, and a 188 when actual was 200, but otherwise all matched


In short, this is very bizarre. How would it be 1-2mm too high? Zoom lenses aren't always at the FL they claim, but if the lens were just giving the wrong numbers, then how would the camera get these random numbers? Seems like the camera is measuring something to cook up these numbers, but what?

[vonBaloney]

Could it be something like the shutter was half-pressed and the SR stabilized, and then the lens was zoomed and fired before it could be stabilized at new value? Or something to do with the actual level of movement detected?

[reeftool]

Could it be something like the shutter was half-pressed and the SR stabilized, and then the lens was zoomed and fired before it could be stabilized at new value? Or something to do with the actual level of movement detected?

The data also records the time of the shutter half press. I have half press times of 1 second to nearly 5 seconds and the SR still records 45-52 mm instead of 17. I also noticed small differences on many of my other lenses.

[vonBaloney]

The data also records the time of the shutter half press. I have half press times of 1 second to nearly 5 seconds and the SR still records 45-52 mm instead of 17. I also noticed small differences on many of my other lenses.

Yes, well that's what I'm getting at, the shutter could remain half-pressed the whole time while zooming and composing but if the focal length is changing the SR system still may need a second to catch up. It's a long-shot, just a thought. It does seem weird, those numbers must come from somewhere...

[IchabodCrane]

Yes, well that's what I'm getting at, the shutter could remain half-pressed the whole time while zooming and composing but if the focal length is changing the SR system still may need a second to catch up. It's a long-shot, just a thought. It does seem weird, those numbers must come from somewhere...

The actual focal length of an internal focusing lens changes with focusing distance. That alone might explain why there are two numbers reported which aren't alsways identical.

[vonBaloney]

The actual focal length of an internal focusing lens changes with focusing distance. That alone might explain why there are two numbers reported which aren't alsways identical.

Mmmmm...that sounds like a possibility also, but some of those numbers were WAY off...

[sTi]

Yes, well that's what I'm getting at, the shutter could remain half-pressed the whole time while zooming and composing but if the focal length is changing the SR system still may need a second to catch up. It's a long-shot, just a thought. It does seem weird, those numbers must come from somewhere...

I can assure you that this is not the issue here, I never zoomed after half-pressing the shutter. The SR focal length numbers seem indeed pretty random for the Sigma in the 17-21 range, even shooting two images one after the other with no changes could lead to wildly different values.
As said above, slight differences between focal length and SR focal length are normal and most likely explained by small deviations of true focal length from the nominal value because of internal focusing desgin or simply because the exact focal length is a bit different than stated, even for prime lenses. This "SR focal length" value was probably introduced by Pentax to ensure optimum SR performance, but in this case it badly backfires because for some reason the communication between camera and lens fails and the resulting wrong SR focal length values actually increase camera shake.