Sigma 50/1.4 - soft copy? - DPChallenge Forums

Canon RF 24-70mm f/2.8 L IS

02/02/2013 06:39:24 PM · #1

I finally decided to play the Sigma lottery. The first unit was clearly a returned one and had front-focus issue. The second copy i was sent seems OK but i just never seem to get a sharp enough shot at below 2.8.

Here is a shot at f/1.8:
//www.pbase.com/mejaz/image/148590651/original.jpg

And here is one of the best one at f/1.4:
//www.pbase.com/mejaz/image/148590653/original.jpg

For both shots, i used the center focus point placed on the eye (camera: 5DII). Is this something i can adjust using microfocus? I am disappointed.

My "kit" lens takes sharper images - granted at f4:
//www.pbase.com/mejaz/image/148590670/original.jpg

Message edited by author 2013-02-02 18:53:29.

kirbic

Canon EOS R5

02/02/2013 06:48:18 PM · #2

Looks like missed focus to me, particularly the 1.8 shot, looks like front focus. Have you tried micro-adjustment?

Mike

Canon EOS-5D Mark II

Tamron SP 24-70mm f/2.8 Di VC USD Canon Mount

02/02/2013 06:58:11 PM · #3

Micro adjust.

LanndonKane

Nikon D800

Nikon AF-S Nikkor 16-35 f/4 ED VR

02/02/2013 07:25:44 PM · #4

definitely looks like missed focus

02/02/2013 08:39:05 PM · #5

Ah, so its the old Sigma issue - front/back-focusing going on here. Thanks everyone, i will try to microadjust but i am ready to return it and maybe wait for Sigma 50 Art series?

Canon EF 24-70mm f/2.8L USM

02/02/2013 11:27:26 PM · #6

Seems like i have to take it to the max - at +20, it seems to be good. Probably will send this in to Sigma for adjustment.

Giles

Canon EOS-5D Mark III

02/03/2013 06:03:43 AM · #7

oh a new lens, its not sharp enought thread hehe ;)

no lens in the world os going to be tack sharp at 1.4/1.8

can you do a proper test on a tripod in daylight at a reasonable aperture 5.6-11 and post a pic up

click on it to view full size, 1 Dx 50mm sigma 1.4 1/60 second noose is in rest is outside its depth of field

f1.4

AB8T8619 by gilesbert, on Flickr

f10

AB8T9316 by gilesbert, on Flickr

your first one was Exposure: Auto exposure, 1/41 sec, f/1.8, ISO 200, slow for hand held push the iso up

your kit lens one 24-105L (its no kit lens!!)

Camera: Canon EOS 5D Mark II
Lens: Canon EF 24-105mm f/4L IS Shot at 93 mm
Exposure: Auto exposure, Program AE, 1/83 sec, f/4, ISO 250

f4 i s alwas going to be sharper than 1.8 apples and oranges

Message edited by author 2013-02-03 06:07:58.

Canon EF 24-70mm f/2.8L USM

02/03/2013 06:10:51 PM · #8

Giles, i agree 1.4 will be soft - but to miss the focus totally is another thing. My sigma seems to be focusing close to the focus point now after +20 MA. Check out the results:

Sigma 50/1.4 @f/2.0 1/40 ISO 200:
//www.pbase.com/mejaz/image/148611042/original.jpg

Your first pic viewed at large 2048 shows front-focusing as well - you can see the hairs closest to the lens are somewhat getting in focus while rest is out of depth of field. I wanted the lens to focus on the eye if i put my focus point on the eye, not on the nose and that is the issue with this lens.

Also yes you are right about 24-105/4L as not really a kit lens, hehe. Its amusing to call it so - its impressive for a zoom. So is the 70-200/4 (i have non-IS).

Message edited by author 2013-02-04 00:26:03.

Giles

Canon EOS-5D Mark III

02/04/2013 02:32:02 AM · #9

your still testign at 1/40 of a second though which isn't going to show a sharp image handheld

try this method here

//www.dpchallenge.com/forum.php?action=read&FORUM_THREAD_ID=1182196&highlight=micro%20adjustment

or get a test chart, from flikr post

napocapo69 says:

I was a bit scared of the reported "front focusing" issues, despite the other wonderful features of this lens, so that I've performed quite a bunch of test shots to understand what happens at wide ( 1.4-1.8 ) apertures.

My statement, is that most of the reported "front focus" issues are wrong.

I'm not saying that this problem does not exist, but that is applicable to only a minor share of the lens.

Indeed, the real issues that likely many of us are experiencing are the combination of the "shallowness" of the focus plane and the sharpness behavior of the lens at wide apertures; as reported in some reviews, as in "dpreview" if I'm not wrong , the lens is softer in the center than in other areas (moving radially from the center).

I suggest, when verifying the "front-focus issue", to secure the focus plane is parallel to the subjects and ensure that subjects have enough contrast. Done that, if you get "in focus" areas at a different distance from where you expected (once again do not confuse focus with sharpness), then you most likely have a focus issue.

Why am I writing this? For a simple reason. Let's imagine that you shot, as i did, with the subjects not parallel to the "narrow" focus plane and you try to focus in the center of the lens (usually this is the case); then you look at the images and apparently the center will look in focus so-so, not so bad....but there is another area not too far from the center that "has better focus" ...is it? Actually is not properly the case; most likely the "better focused" area lays on the same focus plane of the center but has better sharpness due to the lens softness issue (at wide aperture), and the final result is that apparently the lens has focused a closer (or more far) subject.

In practical terms, the lens has to be used with care at wide apertures, to avoid weird results and may be problematic for usage in "macro" shots at wide apertures.

Apart from this (and some issues with purple fringing in particular light situation), the lens could provide outstanding results in most situations.

Hope it helps.
Posted 27 months ago. (permalink)

i was incorrectly testing on an angle, i tested at your settings 1.4 1/40

focus by gilesbert, on Flickr

Message edited by author 2013-02-04 02:32:24.

Ann

Olympus OM-D E-M1

Olympus 12-40mm f/2.8 Pro

02/04/2013 07:09:50 PM · #10

The other thing is that phase detect autofocus, while fast, isn't totally consistently accurate, and at 1.4, the DOF is so thin that the autofocus might just be missing. The LensRentals guy wrote a blog post about this. They were testing a Canon 50 1.4 and 1.2 with a 5D Mk II, and they found that 60% of the time it nailed the focus, 20% of the time it missed focus but the target was still acceptably sharp, and 20% of the time it completely missed focus. He definitely recommended using micro adjust/auto fine tune to correct consistent front/back focus issues, but said that phase detect autofocus is going to miss sometimes. It's built for speed, not accuracy.

The other thing is that the focus sensors are actually quite a bit bigger than the little rectangles you see on your focus screen. If the target (the eye, in this case) just barely fits the little rectangle, the camera probably doesn't have quite enough info to focus with, which will increase the chance of it missing.