I hope this isn't what i think it is

mxfun

mxfun

Suspended / Banned
Messages
3,242
Name
Sean Logie
Edit My Images
Yes
Did a wee test tonight ,and i don't like the look of what i discovered .Maybe i carried out the test incorrectly.

Is this frontal focusing ,if it is is there anything that can be done?
Test shot
img_3559.jpg


Right .. tried the test as was suggested

1.... 200 mm f2.8 iso 100 center weight mtr

img_3589.jpg


2... 200mm f5.6 iso 100 center weight mtr

img_35902.jpg


3...200mm f11 iso 100 center weight mtr

img_3591.jpg
 
You should retry the test with the chart perpendicular to the lens axis and at a distance of at least 10m to get a true idea.
The 45 degree chart has been discredited as a definitive test and all shots should be made at greater than 50x the focal length of the lens (prevents aberations impacting the AF)

Bob
 
Bob, is there a better way to test then, other than this chart? I've done this test with one of my Sigmas and not had good results so I'd be interested in another method of testing. When I shoot with it some shots look OK, others look miles off. :shrug:
 
The method recommended by one of Canon's top men is pretty much as I posted above.

Set the chart perpendicular to the lens axis and at a distance of 50x focal length or more...this prevents any small aberrations in the lens disturbing the AF. I use a moire pattern on a monitor instead of a printed chart. Use AF to focus and then manually adjust to see if the pattern can be sharpened...this is the test pattern I use but there are others.

Bob
 
As CB says, the 45 deg test can prove inacurate and I wouldn't trust it.

Much better to AF on a perpendicular flat surface/image as he suggests, with the lens wide open and at various apertures to compare. Always use a good tripod and remote shutter or on the timer. I wouldn't try viewing crops greater than 100% either as you can end up looking for problems that simply aren't there when printed out.

Good luck, and keep us posted with what you see this time. :)
 
When I first started photography I amassed a load of test sheets. The best one was about 3*2. It covered everything. In colour for colour shots. Dartboards, circles, LPMM, a printout for a given distance from chart.
 
Test charts are a waste of ink and do not work for a multitude of reasons.

IMHO this is the best focus test you can do is what I call the "Lexar test"* - using perpendicular targets. The CF cards are only 3mm think, so you can see instantly if you have any focus problems, if the left or right CF cards are sharper than the center one which you should aim at.

Ditch the charts and only use a good test target that is perpendicular to your sensor.


CONVAR124-web.jpg

​



*other brands of CF cards available!
 
puddleduck said:
Test charts are a waste of ink and do not work for a multitude of reasons.

IMHO this is the best focus test you can do is what I call the "Lexar test"* - using perpendicular targets. The CF cards are only 3mm think, so you can see instantly if you have any focus problems, if the left or right CF cards are sharper than the center one which you should aim at.

Ditch the charts and only use a good test target that is perpendicular to your sensor.


CONVAR124-web.jpg

​



*other brands of CF cards available!
Click to expand...
But I only have sandisk cards.:D
 
puddleduck said:
Test charts are a waste of ink and do not work for a multitude of reasons.

IMHO this is the best focus test you can do is what I call the "Lexar test"* - using perpendicular targets. The CF cards are only 3mm think, so you can see instantly if you have any focus problems, if the left or right CF cards are sharper than the center one which you should aim at.

Ditch the charts and only use a good test target that is perpendicular to your sensor.


CONVAR124-web.jpg

​





*other brands of CF cards available!
Click to expand...

I'll try this next .Cheers
 
really can't see how the 45* test chart can fail to work. The angle is completely irrelavent since the light is travelling perpendicular to the lens at the centre, by definition.
 
pearce_jj said:
really can't see how the 45* test chart can fail to work. The angle is completely irrelavent since the light is travelling perpendicular to the lens at the centre, by definition.
Click to expand...

No its not.

A typcial centre AF point looks like this:

..|..
..|..
----
..|..
..|..

Because its taller than the marks in the viewfinder, anything along that line is fair game. If it focuses at the bottom of the line, it'll look like "front focus". If it focuses at the top of the line, it'll look like backfocus.

Always use a perpendicular AF target, never ever ever use something at 45 degrees as you may get false results.
 
pearce_jj said:
really can't see how the 45* test chart can fail to work. The angle is completely irrelavent since the light is travelling perpendicular to the lens at the centre, by definition.
Click to expand...

JJ....the problem is that the AF sensors are bigger than the focus region on the chart (when shot from the correct distance).

Bob

Edit 'duck quicker than me.
 
puddleduck said:
Test charts are a waste of ink and do not work for a multitude of reasons.

IMHO this is the best focus test you can do is what I call the "Lexar test"* - using perpendicular targets. The CF cards are only 3mm think, so you can see instantly if you have any focus problems, if the left or right CF cards are sharper than the center one which you should aim at.

Ditch the charts and only use a good test target that is perpendicular to your sensor.


CONVAR124-web.jpg

​



*other brands of CF cards available!
Click to expand...

Im going to test my lenses doing this. I would never have thought about it but it makes perfect sense. Cheers Puddleduck :D
 
Janice said:
I don't actually want to try this... because if any of my lenses are out, I can't micro adjust them with my 40d and I would be angrier knowing they are wrong!!

I'd best stick as I am!! :D
Click to expand...


If i was you i'd leave well alone :D
 
Probably an example of test targets showing problems where, almost certainly, none exists.

For focus, I prefer a line of parked cars, shot at an angle. Focusing on the number plate and front grille from a sensible distance usually works well.

For sharpness, distortion and vignetting, shooting square-on to a line of shops or buildings, from as far away as you can get, isn't bad.

Real world and relevant.
 
Personally I prefer a line of AA batteries, 5, lined up diagonally.

Focus on the middle one and see what's most in focus. If you turn them round so that the writing is showing, really easy to see the sharpest battery.
 
barneyrubble said:
Personally I prefer a line of AA batteries, 5, lined up diagonally.

Focus on the middle one and see what's most in focus. If you turn them round so that the writing is showing, really easy to see the sharpest battery.
Click to expand...

What you are doing with such a close up test is exagerating the shallow depth of field so that you can easily see microscopic differences. There is no other reason for doing it like that. Which is fine, if you always use that lens at such short focus. And would be relevant for a macro lens.

But if that is for a 70-300mm zoom normally used at much greater differences, you are pushing it far beyond its normal operating range, exagerating errors which are quite likely within tolerance and invisible in normal use.

There is always a tolerance, set by the CoC depth of focus [not field]. It's there to help achieve good sharp focus quickly, every time. In normal use, this is much more useful than having the lens hunting back and forth in mircorscopic movements looking for absolute perfection.

What confuses me is that these tests are usually done by people with equipment that they currently own, and have been happily using for some time. These tests are not designed to show how good your kit is, but to push them into extreme circumstances where they are quite likely to disappoint. And guess what - they work! ;)

I always test my lenses. But I do it with some very simple real world snaps, and I only do it before I buy, not after.
 
Everyone always has a better way of doing it don't they.

In your rush to critisize, you don't actually notice I havent mentioned what range I use for which focal length. In fact noone has.

Glad you know the best way to do it.
 
barneyrubble said:
Everyone always has a better way of doing it don't they.

In your rush to critisize, you don't actually notice I havent mentioned what range I use for which focal length. In fact noone has.

Glad you know the best way to do it.
Click to expand...

I'm trying to be helpful and I'm sorry that it doesn't come over to you like that. I've explained why I believe some popular lens tests are flawed, and what I think are better alternatives. And for the most part those suggestions are also a lot easier as they don't have to be measurable, repeatable or comparable over weeks and years. You just need to know that your equipment is functioning right, or not.

I've been testing lenses, in every which way I know, for decades. And as a result of that experience, I think I know the best ways to do it. If we all had access to an MTF machine and optical bench, that would be good, but we don't. Pretty much only manufacturers have those things to hand and incidentally, they are always equipped with collimators so that tests can be made at infinity focus.

If I can share the knowledge I have for the benefit of others, I will. I'm not any smarter or better than anybody else, I've just done a lot of lens testing, and listened to a lot of knowledgeable people. People that design and manufacture lenses, and test them to death along the way. If there are better ways, then that would good and I'd really like to know (probably more than most for the reasons I've given) but a line of AA batteries is a small target, so it must be shot at close distance.

I'm not sure why you appear to be suggesting that this assumption is somehow wrong. If most lens tests were at least conducted at normal focusing distances, instead of being artificially close, then they'd be better for it (unless you are specifically testing for close up performance). In your own rush to criticise, you haven't said why you think I'm wrong, or why your own methods are better.

Best regards,

Richard.
 
HoppyUK said:
What confuses me is that these tests are usually done by people with equipment that they currently own, and have been happily using for some time.
Click to expand...

Well, perhaps the reason is that they have just bought a new camera body, as I have, and want to confirm that everything works together fine. I'm sure in your experience you have come to realise that focus is not just a function of the lens, but a joint one between the lens and camera body.

HoppyUK said:
I always test my lenses. But I do it with some very simple real world snaps, and I only do it before I buy, not after.
Click to expand...

Jolly good. How is that going to help with AF Fine tune in my New Nikon Camera Body?? Unless it's controlled and repeatable it's worthless for AF Fine tune.

HoppyUK said:
I've been testing lenses, in every which way I know, for decades. And as a result of that experience, I think I know the best ways to do it.
Click to expand...

For you - in your opinion. See above.


HoppyUK said:
If I can share the knowledge I have for the benefit of others, I will. I'm not any smarter or better than anybody else, I've just done a lot of lens testing, and listened to a lot of knowledgeable people.
Click to expand...

So if you've done a lot of lens testing you'll realise that one test won't work for every situation with every lens.

HoppyUK said:
I'm not sure why you appear to be suggesting that this assumption is somehow wrong. If most lens tests were at least conducted at normal focusing distances, instead of being artificially close, then they'd be better for it (unless you are specifically testing for close up performance). In your own rush to criticise, you haven't said why you think I'm wrong, or why your own methods are better.
Click to expand...

Firstly, lets get some things straight.

I did not critisize your method till now. I was sarcastic to you based on your assumption that your test is so much better for me than my own test. It's not. It may be better for you. Your 40D doesn't do AF Fine tune. My camera does. I need repeatable and controlled shots. I didn't say a single derogatory thing about your tests and assumptions till this post.

You say artificially close. I can do this test from 5-10m away with a couple of my lenses. Is that artificially close? Again, you have assumed things that are simply not the case. You also say - normal focussing distances. For Who? You or me? What's your normal style of shooting. Mine is close up portraits a lot of the time.

How do you know I'm not a macro shooter (which I'm not) and only shoot macro.

Anyway, apologies to the OP for taking this so far off topic.:bonk::bonk:
 
barneyrubble, sorry if maybe the tone of my post upset you, I don't mean to be blunt and I apologise if it sounded like that, but the content still stands. Basically what I'm saying is that if you test equipment in a manner that is too far removed from its generally intended application, then it is more likely that you may find something wrong. And in that case, if you then do something like correct the AF, it could be out for the situations you actually bought it for.

This much is obvious, and you don't test a Ferarri off road, find out it's not very good, and fit a set of knobbly tyres. Nor do you test a Landrover on a race track. So don't put a general purpose lens through an ultra close up test and expect optimum performance. Chances are it will still do a pretty good job, but it may not. The technicians at camera companies that handle complaints about incorrect AF or poor lens sharpness, will tell you that quite often it is the customer's test method that is at fault, more than the equipment.

Looking at your battery test shot, it is fine for testing close up focusing. I have said as much already, but I wouldn't recommend it as a reliable test for more general photography and I certainly wouldn't do any fine AF tuning without also doing a much greater distance test.

You then go on to say that you can also shoot a line of batteries for a meanigful test from 5-10 metres away. You must have a very long lens indeed for that, 600mm at least. Not your everyday lens, and not the kind of lens you would generally use for close ups or portraiture as you describe. It's the kind of lens you might use to shoot the moon, which I see from your profile you have done. That is at infinity focus, and if you've adjusted your AF on the basis of a line of batteries across the room, there is a good chance that it's now out of tolerance for that distance.

Richard.
 
It's not a question of winning and it's not a contest. I would like to think that this is a productive thread about getting the best result.

But I will offer a comment on why I think close up tests are popular for checking AF. When you shoot normal subjects at normal distance, you often can't tell if the AF is out or not. It all looks pretty sharp around the focus point anyway. So you create a more extreme test with miniscule depth of field to magnify any problems, and that's where errors can creep in.
 
Canon Bob said:
JJ....the problem is that the AF sensors are bigger than the focus region on the chart (when shot from the correct distance).
Click to expand...

Also AF favours the closest point to the camera so tends to show front focus where there is none or even perfect focus when in fact there's a small amount of back focus.

Angled charts are bad :nono:
 
You must log in or register to reply here.
Back
Top