Is this front focusing?

AndyWest

AndyWest

Messages
8,400
Name
Andy
Edit My Images
No
Bought a Siggy 30mm f1.4 from the classifieds on here and generally i am very happy with the lens but i think it's front focusing a bit. Please can you tell me if I'm right of if it's not enough to worry about. I have taken lots of images and they all come out about the same. I have tried on objects as well as text. The lens is nice but i just want to know if I'm right or if I need to go to Specsavers!!

Image is at f1.4. Focus is on the word 'and'.
Sigma%20Test_1024pix.jpg


Also this one. Focus on the eye but seems to be at the side of the head.
Sigma-Test2_1024pix.jpg
 
Last edited:
Thanks

I'll give the chart a try.

Back in a tick......

Here is the chart test. Thanks for your input!

Sigma_Test3_1024pix.jpg


Deff Back focusing!! b****r!
Sigma_Test4_1024pix.jpg
 
Last edited:
The first two are back focused, and the second two are front focused :thinking:

Try this as a test subject - a large flat surface leaves much less room for target or user error. Shoot from a sensible distance, like 3-4ft. Artifically close targets can show up errors that don't appear at normal shooting range - there's usually a small element of compromise in getting the best overall AF accuracy, especially with zooms.

Make sure the box face is flat, standing upright, with the camera on a tripod dead square. Aperture f/1.4, centre point, AF expansion off, and aim just to one side of the ruler by a couple of inches or so. Take several shots, defocusing the lens each time. Zoom in to max on the LCD and see where the actual focus point sits on the rule.

IMG_3381-1.jpg
 
It looks back focusing, but check for smaller apperture, maybe f2 - f2.8 to see if the focus will get centered. If it gets centered, then the f1.4 slight back focus is just normal behaviour. Its up to you if you want to micro adjust - if you always shoot at f1.4 then it makes sense, but if you shoot at f2.8 or higher, then you dont need to microadjust.
 
This is my lens and it's performed faultlessly on my 20D at all apertures and at all distances so I'm sorry about this but also baffled as it's always been as sharp in normal use as it was in the test shots I posted in the for sale section.

I suppose there are a number of possible explanations...

1. It was ok and has suffered in transit.
2. Variation in camera bodies.
3. User error.

I'll obviously take the lens back and give a full refund and just out of interest I'll post some test shots in this thread when I get the lens back.
 
Last edited:
Don't worry at the moment. I'll do a few more tests and I'll also see how much it costs to put right as it is a nice lens. I would probably never use f1.4 so if it's OK at f1.8 then I will leave it and use it. It does seem odd that it seems to front focus and then back focus. I had the camera (40D) on a tripod and was careful with positioning and taking the shot.

As said I'll have more of a play :)
 
towershot said:
Cannot say fairer than that m8 :thumbs:
Click to expand...


Hey, without my reputation... I am nothing.

Andy, take a day or so and be safe in the knowlege that if you aren't happy I'm here and so is my bank account. Personally I think that lenses like this are for using at their widest aperture so please don't keep it if you are not happy with the f1.4 performance.
 
Last edited:
Andy, from what you've shown, the evidence is far from conclusive and, as noted, the error is inconsistent.

The image where there is least possibility for error, the test chart, is actually very close (2mm out?) and since the focus is within the depth of field, probably within tolerance.

You must have a flat target, square on, and don't shoot too close.
 
:D Ok mate cheers but the error is so tiny that I bet I would never had noticed normally. I would hardly use it at F1.4 anyway so if I do I can compensate. I may send it to Sigma if it does become a problem but I doubt it will. It's such a nice lens and gives really nice images that it's worth sorting.

Cheers

:thumbs:
 
HoppyUK said:
The first two are back focused, and the second two are front focused :thinking:

Try this as a test subject - a large flat surface leaves much less room for target or user error. Shoot from a sensible distance, like 3-4ft. Artifically close targets can show up errors that don't appear at normal shooting range - there's usually a small element of compromise in getting the best overall AF accuracy, especially with zooms.

Make sure the box face is flat, standing upright, with the camera on a tripod dead square. Aperture f/1.4, centre point, AF expansion off, and aim just to one side of the ruler by a couple of inches or so. Take several shots, defocusing the lens each time. Zoom in to max on the LCD and see where the actual focus point sits on the rule.

IMG_3381-1.jpg
Click to expand...

Weird. I have done this test and even at f1.4 it seems fine!?!?!?

OK maybe it's just me eh!!:cuckoo: :D
 
AndyWest said:
Weird. I have done this test and even at f1.4 it seems fine!?!?!?

OK maybe it's just me eh!!:cuckoo: :D
Click to expand...

With f1.4 this is quite often the case. It certainly was with me! ;)
 
AndyWest said:
Weird. I have done this test and even at f1.4 it seems fine!?!?!?

OK maybe it's just me eh!!:cuckoo: :D
Click to expand...

LOL ;) A possible answer is that the AF point is larger than shown in the viewfinder.

First shot, it could have picked up on the line behind. On the monkey, it could have been drawn to the area on the right because it's brighter. With the plants, if the AF is in two minds, it will always go for the closer area.

And as I say, the one with the least potential for error, the test chart, is very close and probably acceptable.

AF tests are also very prone to user error unless a flat target is used, and after all we're looking at no more than a few mms. That can look like a lot on a chart shot at close range, but in reality the much more likely cause of error is going to be slight subject movement, or movement by the photographer, or just not nailing the AF point accurately.

That Sigma 30 1.4 is a lovely lens :) I used to have one and while I didn't use it much (after I got a 17-55 2.8) it was great when I did.
 
Last edited:
HoppyUK said:
LOL ;) A possible answer is that the AF point is larger than shown in the viewfinder.

First shot, it could have picked up on the line behind. On the monkey, it could have been drawn to the area on the right because it's brighter. With the plants, if the AF is in two minds, it will always go for the closer area.

And as I say, the one with the least potential for error, the test chart, is very close and probably acceptable.

AF tests are also very prone to user error unless a flat target is used, and after all we're looking at no more than a few mms. That can look like a lot on a chart shot at close range, but in reality the much more likely cause of error is going to be slight subject movement, or movement by the photographer, or just not nailing the AF point accurately.

That Sigma 30 1.4 is a lovely lens :) I used to have one and while I didn't use it much (after I got a 17-55 2.8) it was great when I did.
Click to expand...

Thanks everyone. I think it's probably fine and I'm just going to enjoy it while the sun is shining!! :thumbs:
 
May be a silly question, but are you focussing and then recomposing your shot? If so, that would explain the problem.
 
No. I used a tripod and remote.
 
That test chart has come in for a lot of crticism in the past and generally isn't recommended to use. Better option is the SpyderlensCal which gives you a parallel target to aim at first then the correct scale angle to read.

Link

http://www.youtube.com/watch?v=t-JrUQO_Z5U

Realspeed
 
realspeed said:
That test chart has come in for a lot of crticism in the past and generally isn't recommended to use. Better option is the SpyderlensCal which gives you a parallel target to aim at first then the correct scale angle to read.

Link

http://www.youtube.com/watch?v=t-JrUQO_Z5U

Realspeed
Click to expand...

That's true. The problem with an angled target is that, by definition, the focus spot is at different distances. Some of them are hopeless (eg the one given away with Amateur Photographer magazine recently) though with the one used above that's minimal.

The main problem with shooting A4 sheets of paper like that is you have to shoot them very close, artificially close, that and immediately puts the lens right at the edge of it's operating tolerance window. The result is that you sometimes see errors that don't exist at greater distance and if you correct for them, it throws the focus out for normal shooting.

The other problem which seems to affect tests like that is basic user error. It seems easy enough, but often folks just get it wrong. Not being exactly in line with the target is a comon mistake.

The SpyderlensCal is actually identical to the cerial box method :) It is much less prone to error because it is flat, and when shot at a more sensible distance the potential for slight errors in set up shooting angle are substantially reduced.

While I'm on this particular soap box :D I'll mention another AF test method which works better with some lenses, particularly wide angles, that are harder to adjust using the cerial box method, because the depth of field is inherantly quite large. I've not mentioned it before, because you need AF microadjust which the OP's 40D doesn't have.

Very easy indeed. Pin a target to a wall - something with fine detail, like a magazine page or £10 note. Make sure it's dead flat, and set the tripod up central and square. Keep the distance sensible - a 'typical' shooting distance, and not artifically close, like a few feet at least but not the other side of the room.

Shoot at lowest f/number, and run through the full range of AF microadjustment from -20 to + 20 at increments of 5. Zoom in on the LCD and flick through the images to see which one is sharpest. Fine tune from there :thumbs:
 
You must log in or register to reply here.
Back
Top