• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

Sigma 18-250mm f/3.5-6.3 DC OS HSM zoom lens - Part 1

Doug Kerr

Well-known member
For two years, we have done 90% of our dSLR work (on a Canon EOS 20D and then an EOS 40D) with a Sigma 18-200mm f/3.5-6.3 DC OS zoom lens. Often reviled as a "vacation lens" (works for me, as I am permanently on vacation), these large zoom ratio lenses can be vary convenient. Of course, there are inevitable performance penalties associated with such large zoom ratios, perhaps especially for lenses in this price range (roughly USD500).

For Christmas this year, Carla got me the successor to this lens, the Sigma 18-250mm f/3.5-6.3 DC OS HSM. Our quest was for:

• Perhaps a "better" focus mechanism (this lens uses the Sigma "HSM" system, somewhat comparable in concept to the Canon "USM" system - more about this later).
• Perhaps a "better" optical image stabilization system (although there had been no real reports of an "improved" design being used in this lens).
• Perhaps improved "reach" as a result of the greater maximum focal length. (More about "reach" in part 2 of this report.)

A secondary hope was that the lens was not any longer physically (from the flange out) when collapsed than the 18-200, so that we could still put our EOS 40D with the lens in place in its compartment in our favorite camera bag.

The lens is now in hand (we are not big on delayed gratification). I thought I would give a brief report on my observations to date.

Collapsed length

This lens is essentially the same length, collapsed, as the 18-200 (the specs show an increase of 1 mm).

Focusing system

The focusing system in the 18-200 uses a geared DC motor. The motor is fairly noisy. Often, focusing takes place in two steps: a period of the motor running, a brief stop, and then a short period of final motion - just a "hiccup". It is hard for me to relate this to what I think I know about the Canon EOS AF scheme. I will probably discuss this further in a separate thread.

I don't think the HSM AF drive in the 18-250 is any faster. I will be able to make some crude measurements of this later, and I'll report on that. But it gives no impression of being noticeably faster.

It does seem more "decisive" (in that I don't perceive the "two-step" mentioned above). And it is almost completely quiet.

As with the Canon USM focus drive system, there are two sub-species of the Sigma HSM (hypersonic motor) system:

• The ring form, which ordinary brings with it full-time manual focus capability (one can manually change the focus after an automatic focus operation without shifting the mechanism to "MF").

The micro-motor form (which does not ordinarily bring with it full-time manual focus capability).

The system used on the Sigma 18-250 is the "micro-motor" HSM, and there is in fact no full-time manual focus. There is evidently a slip clutch so that an earnest effort to move the focus ring with the lens in the AF state allows the ring to turn (presumably before any violence is done to the gear train).

I have not yet made actual measurements of the AF accuracy of the body-lens combination.

My favorite quick check of lens behavior as a component of the overall process (widely reviled by nay-sayers as "ill founded") is to aim the camera at a focus target, half press to do AF, then release the shutter button and half press again. My belief is that if the focus mechanism does not move on then second half press, then:

• the lens has done its part of the job accurately.
• the camera is consistent in its focus error reporting.

The camera of course may have an inaccurate opinion of where correct focus is, a matter which is not revealed by that check.

Based on this, the Sigma 18-250 behaves very well.

Image Stabilization

The image stabilization system on my Sigma 18-200 was worrisome in its behavior. When it was first engaged (by half press after a period of inactivity), the image in the finder would jump, and when moving the camera slowly (finalizing aim, for example) often the image in the finder would "jump". This of course might be the result of a defect in my copy - I had never sent it back to Sigma for investigation.

Although not really problematical, in the 18-200, when the stabilization system was running, there was a fairly perceptible sound from the lens, like a small motor running at high speed in the distance. And when half-press was released for 3 seconds, there was a noticeable "clunk" as the system as "caged".

With the 18-250 there is none of that. The operation is silent, and 2 seconds of release of half-press (which I assume is when the system "parks"), there is a click so tiny I had to go into our anechoic chamber (/walk-in closet/tornado shelter) to hear it .

Thus this is clearly a wholly different system. But nobody seems to know anything about it. I sense it that it is "ready to go" almost instantly after half-press (with the older system, there was a period of perhaps a second before it "settled in", and firing before that can cause noticeable blurring.)

I'm in no position to judge the degree of stabilization performance. However, during some very quick tests, I found that I could consistently get really nice shots of Carla at an indicated focal length of 250mm (230 mm actual) at 1/30 sec.

Construction and fitting

The construction and finish of the lens is quite nice. The full range of the manual focus ring is only about 45°, not really conducive to "delicate" manual focus work. With the slide in the MF position, the focus ring is very smooth and almost supernaturally easy to turn.

There is a "zoom lock" slide, but as customary for Sigma lenses, this is really a "transport lock" - it can only be engaged with the lens at minimum focal length.

A nice surprise is that the lens came with the new-style Sigma front cap, whose release tabs are near the center such that one can easily remove the cap when the hood is in working position.

The lens comes with a nice bayonet-attachment petal style hood, although of course a single hood can hardly be optimum for such a wide range of focal lengths. The hood can be stored, reversed, and the lens operated that way (MF would be a little dicey - but doable). The lens takes 72-mm thread filters.

The image circle of the lens will not support any EOS format sizes larger than "1.6x".

At a distance of 500', the actual focal length at the maximum setting ("250 mm") is about 238 mm.

In Part 2, I will speak about "telephoto reach" and show some test images with the new lens (and the older one as well) mainly of interest in assessing that.

Best regards,

Doug

"So, what do the nay-sayers say?"
"Well, 'nay', I suppose."
 

Nigel Allan

Member
Doug, I don't know this lens, but I bought my Nikon D300 with the Nikkor 18-200mm and even though some people were almost rude saying it was a jack of all trades, the results are superb. Are there some compromises with this sort of lens? Of course, speed and no doubt it is not quite up to the standards of a dedicated prime, but so what? ... for 99% of my needs it is just perfect and most people wouldn't be any the wiser when they see the shots. Most of my modern dslr pictures in this forum are with this lens. In fact I was in a photo store today looking at lenses and this lens makes it hard to choose something else as it covers so many bases.
 

Doug Kerr

Well-known member
Sigma 18-250mm f/3.5-6.3 DC OS HSM zoom lens - Part 2

"Reach"

I was originally suspicious of the concept of the "reach" of a telephoto lens, largely owing to vague definitions of what was meant, but one of the regulars here (I must admit I forget who) later proposed a coherent definition essentially as follows: The angular resolution of the lens and camera system.

The significance of that is as follows. For the most part, we look to a longer focal length to do one thing: make a certain subject region fill a larger portion of the sensor frame. It is part of the answer to the question:

"If I want to take a picture of that really distant moose, why, instead of using a heavy and expensive 400 mm lens, do I not just use a nice 100 mm lens and crop a portion only 1/4 as wide/high from the frame to print?

Well, if both lenses are capable of giving the same resolution at the focal plane, the answer is, "the cropped image with the 400 mm lens will have 4 times the resolution of the cropped image with the 100 mm lens.

But suppose that the 400 mm lens offers worse resolution (at the focal plane) than the 100 mm lens? Then the advantage will not be as great as we might think.

It can be shown that the metric that allows comparison under this contest is the angular resolution of the lens+camera system (for the setups of interest in the comparison).

So the only really basic reason we would want a 238 mm lens (the Sigma 18-250) rather than a 188 mm lens (the Sigma 18-200) is to get improved performance on the image of a particular subject region(when we crop it to fill a certain portion of our "print"). And do we get that?

To simplify things for the moment, we'll concentrate just on terh result "a little off center". Out target was the front of a neighboring house, nicely perpendicular to our line of sight, 500 ft from the camera location. Among other things, we shot it with the 18-200 at full focal length and then with the 18-250 at full focal length.

Here we see the full frame captured by the 18-200 at its maximum focal length of 188 mm; the aperture here was f/8.0. The white rectangle shows the area we will crop for examination. This image has been downsized for presentation.

Sigma_18-200_OS_200mm_f08.0_MF_F12775_FF-AR.jpg

Douglas A. Kerr: Sigma 18-200 at 188mm (max), f/8, full frame, downsized and sharpened.

Here we see the cropped area from that image, at original resolution.

Sigma_18-200_OS_200mm_f08.0_MF_F12775_C1.jpg

Douglas A. Kerr: Sigma 18-200 at 188mm (max), f/8, crop, original resolution

Here we see the corresponding area cropped from the frame taken by the 18-250 at its maximum focal length (238 mm). Again here the aperture was f/8.0. The crop has been downsized to match the pixel dimensions of the crop above.

Sigma_18-250_OS_250mm_f08.0_MF_F12766_C1R.jpg

Douglas A. Kerr: Sigma 18-250 at 238mm (max), f/8, crop, downsized to match the pixel dimensions of the crop above.

Subjectively, in this presentation, I do not see any overall significance difference in image quality between the two.

At f/11 (I did not show the f/11 comparisons here), the image from the 18-250 slightly outpaced that from the 18-200.

That would then lead to the question, "was it worthwhile to invest in a USD500 lens for that degree of improvement?" In all honestly, I would have to say "no". I justify it solely on the basis of the improvements - however ethereal - in its AF and OS systems.

Conclusion

The user looking for a lens of wide zoom range that can address a very wide range of situations should certainly look into the Sigma 18-250mm f/3.5-6.3 DC OS HSM.

But the owner of a well-behaved Sigma 18-200mm f/3.5-6.3 DC OS should perhaps stand pat (even more so for the improved version of that, the Sigma 18-200mm f/3.5-6.3 DC OS HSM).

Best regards,

Doug
 
Last edited by a moderator:
Subjectively, in this presentation, I do not see any overall significance difference in image quality between the two.

At f/8.0 you are probably getting hit by diffration a bit (at the pixel level). Perhaps the differences are bigger at f/5.6?

You could also first correct the Lateral Chromatic Aberrations a bit before resampling, it seems there is some resolution to be gained there (especially on the 18-250mm).

Cheers,
Bart
 
Last edited:

Doug Kerr

Well-known member
Hi, Bart,

At f/8.0 you are probably getting hit by diffration a bit (at the pixel level). Perhaps the differences are bigger at f/5.6?
A good point. I did not do a detailed comparison from the f/6.3 shots owing to awareness that the sharpness would probably be compromised there, but I will take a look.

You could also first correct the Lateral Chromatic Aberrations a bit before resampling, it seems there is some resolution to be gained there (especially on the 18-250mm).

Also a good point.

I that regard, I have been looking at the lateral CA behavior of the two lenses, and find that, it is generally visibly worse with the 18-250 than the 18-200 (at around 200 mm for both).

We are very nearly at the point of deciding to return the 18-250 and stick with the 18-200. I really enjoy the quiet AF and less-peculiar seeming stabilization, but there don't seem to be any actual performance improvements in either of those two areas (in one set of very limited tests, the AF time does seem to be about 20% less with the 18-250 than the 18-200).

Thanks so much for your insights.

Best regards,

Doug
 
Last edited by a moderator:

Doug Kerr

Well-known member
Hi, Bart,

Indeed, at f/6.3 there is a noticeable improvement in image sharpness in the 18-250 over the 18-200, although accompanied by more visible lateral CA.

This from the 18-200 at 188 mm, f/6.3 (original resolution):

Sigma_18-200_OS_200mm_f06.3_MF_F12774_C1.jpg

Douglas A. Kerr: Sigma 18-200 mm f/3.5-6.3 DC OS, 188 mm, f/6.3

This from the 18-250 at 238 mm, f/6.3 (downsized to the same pixel dimensions as the image above):

Sigma_18-250_OS_250mm_f06.3_MF_F12765_C1R.jpg

Douglas A. Kerr: Sigma 18-250 mm f/3.5-6.3 DC OS HSM, 238 mm, f/6.3

By the way, regarding focus drive, in tests focusing to a target at 3 feet, with the lens initially at infinity, and ac maximum focal length in each case, the time to shutter trip was about 750 ms for the 18-200 and about 600 ms for the 18-250. I'm looking for ways to make a similar measurement in other situations. (It involves the use of a special clock I made some years ago to test shutter release lag.)

Best regards,

Doug
 
Hi, Bart,

Indeed, at f/6.3 there is a noticeable improvement in image sharpness in the 18-250 over the 18-200, although accompanied by more visible lateral CA.

To me it looks as if the CA of the 250mm is better defined, and thus potentially easier to remove.

By the way, regarding focus drive, in tests focusing to a target at 3 feet, with the lens initially at infinity, and ac maximum focal length in each case, the time to shutter trip was about 750 ms for the 18-200 and about 600 ms for the 18-250.

That can be useful, and 150 ms is quite noticeable.

I'm looking for ways to make a similar measurement in other situations. (It involves the use of a special clock I made some years ago to test shutter release lag.)

Interesting. Some time ago I did some shutter speed timings by recording sound. I wanted to test if there was a delay (which was not there) when using the "Silent" shutter mode of my 1Ds3. Sound allows to get very accurate timings with a good recorder (96 kHz or even 192 kHz). Of course mirrors flapping and shutter curtains starting and stopping is probably much easier to record than a focus motor, I'd have to try that ...

Cheers,
Bart
 

Doug Kerr

Well-known member
Hi, Bart,

To me it looks as if the CA of the 250mm is better defined, and thus potentially easier to remove.
An important observation.

That can be useful, and 150 ms is quite noticeable.
Indeed.

Interesting. Some time ago I did some shutter speed timings by recording sound. I wanted to test if there was a delay (which was not there) when using the "Silent" shutter mode of my 1Ds3. Sound allows to get very accurate timings with a good recorder (96 kHz or even 192 kHz). Of course mirrors flapping and shutter curtains starting and stopping is probably much easier to record than a focus motor, I'd have to try that ...

In my technique, I use a clock with a single continuously running hand, evolving at 1 revolution per second (with the face calibrated in decrimal fractions of a second).

In this situation, I preset the lens focus to infinity, engage the AF, aim at the clock, and as the hand crosses the 0 point fully mash the shutter release.

I of course get an image showing the hand where it was at the time the shutter opened.

Of course, there is some uncertainty as to my manual tripping of the shutter (an electrical release tripped by a cam on the clock would be would be nice, but I haven't made the necessary provisions for that yet).

I suspect that the experimental error due to that is on the order of 150 ms.

Thanks again for all your observations. A decision I have to make soon is whether to keep the 18-250 (and likely sell the 18-200) or return it and revert to the use of the 18-200. (We are of course not speaking of a gigantic amount of money here - the 18-250 was USD 500 landed.)

More than the cost issue of course is the tradeoffs, largely concern over the increased CA on the part of the 18-250 (which of course can be corrected in settings where it is intrusive, and I appreciate your observations about it perhaps being better-defined in the 18-250 and thus more susceptible of mitigation).

And I sure appreciate the better-behaved AF and OS machinery!

Best regards,

Doug
 
More than the cost issue of course is the tradeoffs, largely concern over the increased CA on the part of the 18-250 (which of course can be corrected in settings where it is intrusive, and I appreciate your observations about it perhaps being better-defined in the 18-250 and thus more susceptible of mitigation).

And I sure appreciate the better-behaved AF and OS machinery!

Hi Doug,

Also don't forget to check it out at the short end of the focal length zoom range. A newer lens design may have some improvemnets there, perhaps it has a newer coating to improve contrast or resistance to flare. Due to the enormous range of focal lengths, there will be optical quality compromises, no surprise there, but one of the main features of such a lens is its versatility, even on cropped fields of view due to smaller sensors. So it should perform good enough under a wide range of focal length settings.

Cheers,
Bart
 

Doug Kerr

Well-known member
Last night, while shooting the infamous "Beans on stove 2009", I discovered a truly unique property of the Sigma 18-250 f/3.5-6.3 DC OS HSM lens' image stabilization system.

I had seen this referred to in a post on some other forum, but I wasn't exactly able to grasp what the writer was saying, and missed seeing this effect until last night.

In the typical optical stabilization lens, when the lens is de-energized, the natural position of the aim-changing element is either "centered" of "floating. However, when the lens is inactive (and usually even when the stabilization is disabled), a "caging" mechanism forces the element to its centered position and holds it there. This has many purposes, one of which is to prevent the element, when the lens is de-energized, from swinging around and possibly causing damage. It of course also eliminates the need to operate the servo system when image stabilization is turned off.

It seems that in the Sigma 18-250, with the lens de-energized, the movable element is spring biased against a stop at one limit of its travel ("parked"). Thus, no separate "caging" mechanism is needed.

When we first half-press, or (in the case of my 40D) when Live View is enabled, the aiming servo system smoothly shifts the aim of the lens to "centered" (taking perhaps a second to do so).

Then, after 60 seconds of inactivity (no half-press, no Live View), the servos seem to be shut off, and the aiming element unceremoniously drops to its "parked" position with a modest but perceptible "clunk".

If the image stabilization is shut off, the servo system still operates (once we half press or turn on Live View) so as to move the aiming element to the "centered" position and actively hold it there.

Of course, there could be some bad side effects of this. If I was setting up a shot through the reflex finder, and had yet had no reason to poke the shutter release, my aim would have been inappropriate.

So I'm not sure how to feel about this.

In any case, I thought it was worth reporting.

Best regards,

Doug
 

Doug Kerr

Well-known member
It's all over

This is to close out this saga.

We decided to return the Sigma 18-250mm f/3.5-6.3 DC OS HSM to the vendor. This was basically motivated by:

• The fact that the additional focal length over the 18-200 did not bring any more "reach" (that is, making a subject fill a certain portion of the frame could be done by cropping with just as good resolution in the delivered image).

• The fact that lateral CA was in fact greater in some regimes of interest.

• My discomfort with the odd behavior of the optical stabilization system (in that the aiming axis of the camera with the lens "lit" is different than with everything "idle").

• The fact that I got no bids on the 18-200 on eBay at a starting price of $300.00!

Essentially, we found that the overall benefit was not worth the cost.

We have had good results for two years with the 18-200 and plan to do more of the same.

When better photographic performance is needed, we can certainly use our Canon EF 24-105 f/4L USM or the EF 70-200 F/2.8L USM.

Sic transit gloria mundi.

By the way, the vendor (Sigma4Less, a.k.a. Photo4Less) was very accommodating about this, even though we decided to return the lens well after their normal limit for no-penalty returns (7 days after receipt). I've had good results with that firm on several occasions.

Best regards,

Doug
 
Top