Sigma 24-70 2.8 how sharp should it be?

M

Mustard shots

Suspended / Banned
Messages
130
Name
Peter
Edit My Images
Yes
I've got a used Sigma EX 24-70 2.8 DG DF on loan from a guy at work who wants to sell it to me. I've been doing some pretty basic trial shots hooked up to my D300 on a tripod wide open and to be honest I'm a bit disappointed with the sharpness just looking through the screen on the back of the camera. Haven't bothered going as far as the computer yet.

So far I've only tried it at 2.8, with a 2 sec timer to try and eliminate camera shake on the tripod. It's been dusk, so mainly tungsten from the room lights - no flash as yet but that's next to see if it helps.

General room shots weren't coming out sharp when I zoomed in to check the detail, so I went for things with more contrast like a door knob. Then tried a cactus on my window ledge which looked awful. Not satisfied, I tried a box with a hard edge, followed by a book with text on. They're rubbish!

In a way, I suppose none of this really matters as I'm more interested in portrait anyway. Just bugs me :suspect:

Am I being unfair do you think, or is this really why the Nikon equivalent is 5 times the price used?

Thanks

Pete
 
well, these are unrealistic tests. at f/2.8, most things will be soft. My 50mm is a superb lens but let's face it at f/1.4 its no stellar performer. Stop that baby down to f/4 or f/5.6 and then retest. Oh, and put them on the computer - the image you see on the screen is only the JPEG preview embedded in the RAW image.
 
Locking the mirror up might help ;)

The D300 has a suprising amount of mirror slap
 
Unfair? Yes, in just about every respect... get it outside or use it with flash and give it a fighting chance....
 
If it was a Nikon he'd be telling himself in every way possible that he was being unfair. It seems that because it's a Sigma, any doubt is the lens's fault :shrug:
 
I owned one of these, the softest lens I've ever owned.

Am i correct in assuming you have the non HSM version? Even the newer HSM version of this lens which is supposed to be 'less bad' gave appalling test results on SLR Gear.

FWIW my advice would be don't buy it.
 
At f2.8 this lens may be soft but at f/9 things get even more interesting. In fact, I'd love to see how a good copy performs if there are any. Skip that lens and look for Tamron or Tokina.
 
Deasy said:
I owned one of these, the softest lens I've ever owned.

Am i correct in assuming you have the non HSM version? Even the newer HSM version of this lens which is supposed to be 'less bad' gave appalling test results on SLR Gear.

FWIW my advice would be don't buy it.
Click to expand...

Really? Do you have a link for said review?
 
daugirdas said:
At f2.8 this lens may be soft but at f/9 things get even more interesting. In fact, I'd love to see how a good copy performs if there are any. Skip that lens and look for Tamron or Tokina.
Click to expand...

Give it a rest dude, you say that about practically every Sigma lens.
 
FlyTVR said:
Really? Do you have a link for said review?
Click to expand...

yup, here it is:



http://www.slrgear.com/reviews/showproduct.php/product/1236/cat/31

To quote from the test, it's a worst in category lens..

On the full-frame 5D, the true character of the lens is revealed. If you are looking for a lens that is corner-to-corner sharp when used wide open at ƒ/2.8, this is not the lens. True to what we noted in our tests with the 20D, central sharpness is still remarkable at 24mm and ƒ/2.8, but outside of this sweet spot sharpness falls off rapidly, reaching (and I think this is a record for us) 17 blur units in the top corners and 13 blur units in the bottom corners.
 
trencheel303 said:
Give it a rest dude, you say that about practically every Sigma lens.
Click to expand...


Some are good, some are bad, but I agree with you on this - give it a fighting chance - it might be one of the good ones...

To the OP: Nothing looks good under tungsten light as the full gamut of wavelengths aren't present.

Do the test again, stopped down to f/5.6-f/11 in daylight or with flash as has been suggested...
 
beg borrow or steal a Nikon 24-70 and do the same tests - be very surprised if you bought the Sigma then ;)
 
trencheel303 said:
well, these are unrealistic tests. at f/2.8, most things will be soft. My 50mm is a superb lens but let's face it at f/1.4 its no stellar performer. Stop that baby down to f/4 or f/5.6 and then retest. Oh, and put them on the computer - the image you see on the screen is only the JPEG preview embedded in the RAW image.
Click to expand...

Stopping to 6.3 or 8 makes a huge difference to sharpness... 2.8 is rubbish in comparison.

Maybe I'm missing the point, but what's the point in spending cash on a 2.8 lens when it can't perform at that aperture? :bonk:
 
Maybe, just maybe, try honing your skills as a photographer rather than pixel peeping at 200% or whatever. The majority of paying customers don`t look as much as we do.

Saying that, maybe you have a sh@t lens...........:shrug:
 
Mustard shots said:
Stopping to 6.3 or 8 makes a huge difference to sharpness... 2.8 is rubbish in comparison.

Maybe I'm missing the point, but what's the point in spending cash on a 2.8 lens when it can't perform at that aperture? :bonk:
Click to expand...

It's only rubbish at f/2.8 on the cheap stuff though.. try a top drawer Nikon/ Canon/ Sony Carl Zeiss and you'll see there's a huge difference.
 
Richard Peters said:
You need to re-test at at least f5.6. The depth of field isn't going to be big enough to get everything sharp at f2.8 :)
Click to expand...

Understood.

The thing is I'm not getting anything sharp at 2.8 - nada. 2.8 and 70mm that is. Will try 2.8 and 24 and 50ish as well... Just interested in knowing whether that is the lens or my bad technique or I'm simply just expecting too much from the kit. I'm slowly working my way through a few trials to see....
 
I've tried the current versions of both the Nikon 24-70mm f/2.8 and the Sigma 24-70mm f/2.8 outdoors in natural sunlight, with speedlights, and under studio stobes. I could've picked up the Sigma quite inexpensively recently, but I passed for 2 reasons.

1) While the Sigma *can* be sharp, to me it wasn't as sharp as the Nikon at any aperture and the increase in sharpness as I went from wide open down each stop was much more dramatic on the Nikon. If I got it, I knew I'd only end up replacing it with the Nikon one anyway.

2) I wasn't impressed by the 82mm filters. This meant not only duplicating all the pro filters I've already bought at 77mm (and quite happily switch between my Sigma 10-20mm f/4-5.6, 70-200mm f/2.8 VR and 300mm f/4 AF-S), but paying an average of about 20-30% more per new filter I'd have to buy (B+W 77mm Pro UV filter is around £69, the 82mm is £89. B+W 77mm MRC Circular Polariser is aruond £155, the 82mm version is £179 - Warehouse Express prices). The Nikon one uses 77mm filters.

So, I've passed on the Sigma, and the Nikon one is currently close to the top of my shopping list.
 
I had one of these in Canon fit, it was soft as hell at F2.8 but was absolutely razor sharp at F8, it was a nice lens to use though, it feels nice... the worst thing about it was having to do two operations to switch from auto to manual focus, flick a switch and pull the ring... not sure if the Nikon one works like that.
 
No, the Nikon one is AF-S, so assuming you don't use your shutter half-press to auto-focus (or you do and you're in single servo AF mode and it's already locked onto something), you can just turn the ring as normal - even if the lens & body are set to AF modes.
 
Thanks for your thoughts John.

The results I'm getting are a bit semi-hit and total-miss. The addition of flash only seemed to make things worse (not entirely convinced I know why, but off camera flash gave sort of a ghosting effect similar to a small amount of motion blur) so I've abandoned tests and will have another crack tomorrow.

There's a few Nikons that have been advertised recently - just not convinced I can justify the 900+ for a pro piece of glass when I'm not making a living out of it. Someone mentioned looking at the Tokina and I may well do just that, but really liked your justification thoughts on the filters - that's my kind of logic!

Thanks all

Pete
 
Ahh that would be better, the Canon one was irritating because of if but you couldn't really use it wide open. I broke it too actually, dropped it and my 30d and broke the bayonet off it, it was repaired and recalibrated by Fixation and was no better. I assumed it's just the way they are. Decent value for money though.
 
Kaouthia said:
No, the Nikon one is AF-S, so assuming you don't use your shutter half-press to auto-focus (or you do and you're in single servo AF mode and it's already locked onto something), you can just turn the ring as normal - even if the lens & body are set to AF modes.
Click to expand...

This one you push the ring forward or back to snap between manual and autofocus. Manual focus feels like you are turning the whole gear mechanism and to be honest I've found it hard to get accurate.
 
Mustard shots said:
Thanks for your thoughts John.
Click to expand...
No problem. :)

Mustard shots said:
The addition of flash only seemed to make things worse (not entirely convinced I know why, but off camera flash gave sort of a ghosting effect similar to a small amount of motion blur) so I've abandoned tests and will have another crack tomorrow.
Click to expand...
Do you have an example image you could post with EXIF info?

I'm thinking what might be the case is that you've just got too long a shutter speed and too wide an aperture to block out enough ambient light. So it's partly exposing from the ambient light (and because you're handheld it's going to be a bit soft), and then the flash is just finishing off the exposure to lighting up your subject.

Sometimes this can be a useful effect/feature. Sometimes, not so much, although it's often a way I work with flash outside of a studio. Exposure settings for 2-3 stops below ambient, then bump up my subject with flash.
 
Mustard shots said:
This one you push the ring forward or back to snap between manual and autofocus. Manual focus feels like you are turning the whole gear mechanism and to be honest I've found it hard to get accurate.
Click to expand...


The Canon one is even worse as you have to flick a switch as well though I found it easy to use once you'd changed it, the focus ring is nice and big and the viewfinder is nice and bright too.
 
Oooh, if you're getting it on a tripod, then that does seem odd (unless it's possibly mirror-slap causing movement during the ambient exposure and screwing it up that way). Interested in seeing the image now.

Have you got the delay mode enabled or disabled? (it's menu option D10 on my D300s, so I'm assuming it's the same on the D300).
 
Trying to asses the sharpness of a lens when tested in isolation is extremely difficult. Of course you can usually spot a lemon but anything more critical is impossible without doing a full lab test under very controlled and repeatable conditions, so that you can measure and compare resuts.

However, if you just want to see if a lens is any good or not, it's done easily enough by comparing it with identical images shot side by side with another lens of known quality. That way, you might not have measureable or empirical results, but you know that the results are absolutely comparable and you can assess the relative performance pretty well.

This particular Sigma is never going to compete with similar lenses from Nikon/Canon. It fits Sigma's hallmark of adventurous spec and low price, so something has to give. But almost any lens should be sharp enough in the centre, even at full aperture. It will be at the edges and corners where the design economies will be most evident.

To test for a 'good copy' the usual problem is de-centering of the elements, either through poor assembly or because it's taken a knock. That is easy enough to test by comparing the corners and if one or more of them is substantially worse than the others then that's a sure sign.

The way to check that is to focus on a distant subject, to reduce potential problems of focus error or field curvature. I use a road sign or car number plate. Shoot at lowest f/number and take four pictures with the target in each corner of the frame. Don't expect high sharpness with this test as you won't get it - it's a tough test and will probably be quite poor, that's not the issue and in normal picture taking you might never notice it. What you're looking for is an even level of sharpness in all four corners. If you have to look twice, then it's probably okay on that score - significant errors are quite obvious.

D300 has a good LCD - you can see all you need on that. You can do it all hand-held too if the light is good and you can keep the shutter speed well up. Raise the ISO to get it.
 
After reading that it almost encouraged me to go and 'test' my lenses.... I managed to stop myself, you can spend your whole life chasing sharpness when there's nothing really wrong and I'm actually happy with my lenses. My old Sigma was great at F8.... to use F2.8 would have needed an emergency but you pay your money and get your choice. I might have some test images somewhere I did with it but they were on a Canon 30d so not really a fair comparison.
 
I would disagree with you a little on your assessment of Sigma as a producer of adventurous spec at low price.

I think that they make some lenses that do more than stand comparison with the body manufacturers. 12-24mm, 50mm f1.4, 30mm f1.4 and more...these are very good lenses and would be no matter what makers name was on the lens.
 
woof woof said:
I would disagree with you a little on your assessment of Sigma as a producer of adventurous spec at low price.

I think that they make some lenses that do more than stand comparison with the body manufacturers. 12-24mm, 50mm f1.4, 30mm f1.4 and more...these are very good lenses and would be no matter what makers name was on the lens.
Click to expand...

Fair comment - what I meant to say was adventurous spec or low price. And yes, they do make some excellent lenses of high spec, but not always low price!

Sigma 50 1.4 is generally regarded as the best lens out there of that specification, bar none. I used to have the 30 1.4 and it was excellent - fab on a crop camera.

I would also add that I've tested several copies of several different Sigmas side by side, and every one of them was a 'good copy'. Sigma have a reputation for poor quality control but whatever the truth of that or not, the chances of you getting a duffer are still pretty low.
 
I have a Sigma 28-70 and a 24-60

The 28-70 is sharp at 2.8 until 55mm and very soft after that
The 24-60mm is sharp from 24 to 60mm

So it depends on both the lens model, the actual copy you have and the length you are using it at.
 
i was thinking about buying this lens for my EOS body since the canon 24-70 is pretty expensive and i probably wouldn't use it enough to justify the cost.

the review on this site http://www.the-digital-picture.com/Reviews/Sigma-24-70mm-f-2.8-EX-DG-Lens-Review.aspx is quite balanced, but it does say "To me, 70mm f/2.8 is not usable on this lens. " this put me off a bit since 70mm at f2.8 is a nice length and aperture for portrait work.
 
I think that this was the reason Sigma sold the 24-60mm which was sharp at 60mm but maybe pressure to match the length of the Canon 24-70mm led to the 70mm model which is not sharp at all between 60 and 70mm!!! This was the case with the 28-70 I had and also a 24-70 I tried.

So either get the 24-60mm or get the 24-70 and use it up to 60mm. Portraits on a crop sensor are best taken at 60mm or below anyway IMHO
 
What's the point of having a lens you can't use at f/2.8? I've recently changed from a Sigma 18-50mm f/2.8 to a Tamron 17-50mm f/2.8 and the difference wide open is night and day. The Tamron isn't sharp right to the corners, but center sharpness is awesome. The Sigma was just plain soft everywhere.
 
Well I've been playing around again with the lens and generally different results again to last night with a bit more ambient light.

One portrait handheld of my flatmate with some pretty harsh off-camera flash came out just fine at f3.5. Some other shots weren't as crisp with the same settings.


I'm mulling over a few thoughts:

I took a few test shots using a tripod and a handsfree earpiece with sharp edges - pretty boxy object and on white, so hopefully simple for the camera to focus on. Shot slightly from above - no flash. One thing that became really obvious was what an incredibly small depth of field 2.8 has. You could see the roll off over the body of this thing (shot at a slight angle rolling away from the lens).

This also made me realize how sensitive or picky this lens seems to be - some shots were ok or passable and some were pretty awful with same settings but re-visiting. I've tried to minimise my input into bad results - but accept that I haven't quite bolted the camera to the floor to keep it rock steady.

So it made me doubt it's ability to actually lock-on with AF every time.

I tried the numberplate trick mentioned before hand, but the light had faded so much by the time I got around to that and with the ISO cranked up that far the results were pointless really.

I imagine the Sigma is fine, just a little more sensitive to focus IMO and doesn't always get it right :suspect: a different story to my 2.8 70-200 AF-S VR and even 24-120 which always give me great results unless its pilot error.

So far the conclusion is that I'll never be 100% happy with the Sigma - although it certainly has merit for £200 in my opinion. The hit and miss focussing close up would probably wind me up.

Looks like I'll have to save my cash and flog a few bits and wait my turn for a Nikon piece of glass :shrug:

Cheers all

Pete
 
Pete, your findings don't surprise me. I own the same lens (Nikon mount) and when it focusses properly, the results are decent enough (given I don't expect any lens in my budget range to hit F2.8 and be sharp corner to corner). The AF is dog slow and a bit hit and miss though.
 
I've got a Canon fit and it's fine. The ONLY thing I have found is that in "Macro" mode it's focussing is not particularly accurate, so I use live view and manual focus. Other than that, for the money, it's great.

Obviously, the Canon is better, but when I bought mine, the Canon was 3 times the cost, and not 3 times better !
 
You must log in or register to reply here.
Back
Top