Focus shift - can live view focus stopped down?


WaypointCharlie

Link Posted 24/10/2019 - 08:58
I recently picked up a 18-270mm to update my ancient 18-250mm 'travel zoom' for my K-5 IIs. I realise it's not going to be up to the quality of a set of primes, but sometimes convenience is more important to me.

One thing I notice, as I did with the 18-250mm, is how much the focus varies with aperture. The effect varies with focal length and probably focus distance too.

For instance, at 53mm and at a distance of 1.4m focus is about right at f/4.5 but I get front focus at f/8.0. This is presumably due to lens aberration and because phase detect focussing is always performed wide open, regardless of aperture setting.

However, I get the same with live view using contrast detect. I had hoped contrast detect focussing would be perfomed stopped down, but it appears not to be.

Is there a menu option somewhere to enable contrast detect focussing to be perfomed stopped down?

Otherwise I guess it's a case of learning the lens and where best to focus when composing.

pschlute

Link Posted 24/10/2019 - 15:06
Odd. There is a thing called focus shift which may explain this but it usually only applies to fast lenses 1.4 or so.
The actual focus plane should remain the same whatever aperture you choose. Can you post some large scale examples
Peter



My Flickr page

WaypointCharlie

Link Posted 24/10/2019 - 17:42
I've just dusted off my tatty old adjustment chart and took these two shots. Focal length 58mm. Distance about 60cm. I'm not sure how well the image will look at forum resolution.

It's clear that f/8.0 is focussed in front of f/4.5, even though I didn't refocus betweeen shots. At this distance it's f/8.0 that looks about right and f/4.5 is back focussed (by about 12mm?). The error will probably change with focus distance but I think f/4.5 will always be behind f/8.0.

As you say, this is the sort of focus shift you might expect with a large aperture spherical lens. However, I've noticed this characteristic before with some of the Pentax zoom lenses. It means it's pretty pointless attempting to program micro focus adjustments on these lenses!

I was hoping there might be a way of focussing with live view contrast detection, but this seems to focus wide open too. Or is there an option somewhere to focus stopped down?

pschlute

Link Posted 24/10/2019 - 18:44
You need to do about 10 shots at widest aperture to get an accurate representation. So camera on a tripod, remote or 2 sec delay. Natural light if possible.

Make the camera focus each time from infinity alternating with closest focus distance. Then compare the results. This will tell you if your AF system needs adjusting.

I am afraid those picture are too small to draw any conclusions.
Peter



My Flickr page

WaypointCharlie

Link Posted 24/10/2019 - 19:52
There's no point in making micro AF adjustments. The problem is focus shift. I notice it with the lower grade Pentax zoom lenses. Is there no way of focusing stopped down using contrast detection?

Yes, of course I used a tripod and 2 sec delay and natural light. I also tried acquiring focus from infinity and near, although that's irrelevant to the focus shift with aperture issue. In any case, I didn't re-acquire focus bewteen the f/4.5 and f/8.0 shots - the difference is entirely due to aperture related focus shift.

Here's a higher res image on flickr. I'll leave it up until tomorrow morning. The focus shift should be obvious (further back when wide open).

flickr link

Nigelk

Link Posted 24/10/2019 - 20:22
I can see what you mean Wayward but sadly have no solution.
I used a newspaper with a black line on the margin to do a similar focus test. I soon realised that the focus of the zoom lens I was testing had a very variable bias back and forward of the focus point dependant on focal length and aperture. I quickly abandoned the idea of making in camera focus adjustments, it would have been pointless.
I wonder if newer bodies would give a better result or if is is all the lens. I think you need to buy a KP to find out.

WaypointCharlie

Link Posted 24/10/2019 - 20:34
I don't suppose newer bodies would be different. I suspect they still acquire focus wide open and won't perform any tweaks for different focal lengths, apertures and focus distances. Fortunately in reality I rarely do much that's focus critical anyway, as I'm sure you'd agree!

Nice try with the KP!

pschlute

Link Posted 24/10/2019 - 22:28
Interesting. Does the effect still occur at longer camera to subject distances and at different zoom settings ?
Peter



My Flickr page

WaypointCharlie

Link Posted 25/10/2019 - 11:50
It gets weirder. I've just tried at 270mm and found that the camera focuses slightly differently as you stop down. This suggests the firmware attempts to compensate for the focus shift.

I never realised the camera took account of the aperture when focusing. It's cleverer than I thought. Only problem is, it compensates in the wrong direction! At f/16.0 it sets the focus slightly closer than at f/6.3 (wide open), yet what we want is for the focus to be set slightly further back. Ah well!

It's not good for your health to investigate these things too deeply.
Last Edited by WaypointCharlie on 25/10/2019 - 11:51

Nigelk

Link Posted 25/10/2019 - 13:19
"It's not good for your health to investigate these things too deeply."

Wise words softly spoken

So apart from all the aforementioned, are you pleased with your new convenience travel zoom?

pschlute

Link Posted 25/10/2019 - 13:25
WaypointCharlie wrote:
It gets weirder. I've just tried at 270mm and found that the camera focuses slightly differently as you stop down. This suggests the firmware attempts to compensate for the focus shift.

I never realised the camera took account of the aperture when focusing. It's cleverer than I thought. Only problem is, it compensates in the wrong direction! At f/16.0 it sets the focus slightly closer than at f/6.3 (wide open), yet what we want is for the focus to be set slightly further back. Ah well!

It's not good for your health to investigate these things too deeply.

The firmware will not be adjusting focus to account for "focus shift".

Are you testing this scientifically ? At least 10 attempts for each setting otherwise the vagaries of AF accuracy (ie not very) come into play.
Peter



My Flickr page

JAK

Link Posted 25/10/2019 - 14:12
It may well be worth testing this on another body too.
How do you find the accuracy of TTL centre spot focussing with your K-5iis?
John K

WaypointCharlie

Link Posted 25/10/2019 - 14:16
pschlute wrote:

The firmware will not be adjusting focus to account for "focus shift".

Are you testing this scientifically ? At least 10 attempts for each setting otherwise the vagaries of AF accuracy (ie not very) come into play.

Yes, I'm testing reasonably 'scientifically'.

I first noticed it at a distance of about 0.6m with a focal length of 270mm. Using a tripod I focussed at f/6.3 and set the camera to manual focus. I fired a shot at f/6.3.

I then set to f/16. The camera wouldn't fire because it considered it to be out of focus. When I enabled auto focus the camera refocussed slightly nearer.

I then returned to f/6.3. The camera then refocussed slightly futher back again. I am making no changes to the focus or camera position.

I have repeated this several times - always the same.

Like you I didn't realise the firmware took account of the aperture when focussing, but it does. At least with the 18-270mm lens at 270mm.
Last Edited by WaypointCharlie on 25/10/2019 - 14:29

WaypointCharlie

Link Posted 25/10/2019 - 14:29
Nigelk wrote:


So apart from all the aforementioned, are you pleased with your new convenience travel zoom?

Yes, it's an improvement on my old 18-250mm which feels like it's about to die. It may not be the finest quality glass but competition judges don't seem to notice! Pity it's not weather sealed though.

I had vowed not to buy any more Pentax gear but it I couldn't resist.

pschlute

Link Posted 25/10/2019 - 14:37
I have no explanation, but doubt the firmware has anything to do with it.

At 60cm you are only just inside the minimum focus distance, try at 15 feet or so. Also get the camera to focus each time from the extremities (infinity or closest distance). It does make a difference. I have calibrated 14 lenses with my K1 and they all behave slightly different. I put the K1 AF accuracy to be 70%. I doubt the K5 is any better.
Peter



My Flickr page
Add a Comment
You must be registered or logged-in to comment.