Going crazy with my 70-300!

[pdk42]

OK, so I'll preface this by saying that I've now binned all the images so can't back this up with any examples, but...

I find myself sometimes extremely disappointed with the 70-300, but not always! It performs absolutely fine up to about 200mm, but above that I sometimes find that images are just a mushy mess. It only seems to happen on distant subjects and it's frustrating the life out of me. Maybe it's heat haze/atmospheric distortion, maybe my hand holding technique, I dunno. I guess I need to do some controlled tests on a tripod...

 

[dirk]

OK, so I'll preface this by saying that I've now binned all the images so can't back this up with any examples, but...

I find myself sometimes extremely disappointed with the 70-300, but not always! It performs absolutely fine up to about 200mm, but above that I sometimes find that images are just a mushy mess. It only seems to happen on distant subjects and it's frustrating the life out of me. Maybe it's heat haze/atmospheric distortion, maybe my hand holding technique, I dunno. I guess I need to do some controlled tests on a tripod...

Take your S5II and make test shots which others might be able to replicate. Write down the distance and daylight and maybe use subjects others have too. If I can replicate it, I will try the same with my 70-300. But I only have a Lumix S5mkII.

We will have no sun over the next 3 days here. Rainy and cloudy.

 

[AlanC]

You can check your hand holding technique with the I.S. Status Scope function: if the dot stays inside the inner circle then the stabilisation can fully correct for any camera shake.

 

[Pete_W]

OK, so I'll preface this by saying that I've now binned all the images so can't back this up with any examples, but...

I find myself sometimes extremely disappointed with the 70-300, but not always! It performs absolutely fine up to about 200mm, but above that I sometimes find that images are just a mushy mess. It only seems to happen on distant subjects and it's frustrating the life out of me. Maybe it's heat haze/atmospheric distortion, maybe my hand holding technique, I dunno. I guess I need to do some controlled tests on a tripod...

I get caught out by atmospheric haze with this lens at longer focal lengths, especially in warm weather.

I've just posted some photos with this lens in the March thread, and one of them is at 300mm.

I've looked at a few others and I can see blurriness at the edges in some photos particularly where the foreground on one side is not on the same plane as the other side. Maybe a bit of field curvature? It was windy today so I can't rule out that causing some blur.

I tried to upload another image here (not one I am planning to put up on flickr) but am running into file size problems for some reason - the forum won't even accept a 1.1mb file.

 

[pdk42]

Well, I think the problem is DOF, with a little bit of field curvature in play. At 300mm, even at f8, the DOF isn't that great. Being picky on CoC (0.005mm), DOF is narrower than I thought:

Focal length	300.00
Aperture	8.00
CoC	0.005
Hyperfocal dist	2250.30
Focus dist	100.00	500.00	1000.00	2000.00	3000.00
DOF	8.91	233.73	1107.47	16921.93	∞
Near	95.75	409.10	692.34	1058.89	1285.81
Far	104.65	642.83	1799.81	17980.82	∞

So, even focussing at 1km won't get things past 1.8km in focus. The hyperfocal distance is 2.2km. Even at f11, the hyperfocal distance is still 1.6km. Zooming out to 200mm, the hyperfocal distance at f11 drops to 700m, which is a significant difference.

In fact, looking at the equations, the hyperfocal distance is proportional to the square of the focal length, so it's easy to see how increasing focal length rapidly leads to very narrow DOF.

On top of that, there is definitely some field curvature going on which exacerbates the problem.

So, bottom line, at these long focal lengths, getting everything in focus on landscape type shots isn't easy. A narrow aperture is certainly sensible.

 

[dirk]

So, bottom line, at these long focal lengths, getting everything in focus on landscape type shots isn't easy. A narrow aperture is certainly sensible.

Sounds like the "negative" small aperture of the 28-200 zoom on the long end would be not that negative for your use cases

 

[Pete_W]

Well, I think the problem is DOF, with a little bit of field curvature in play. At 300mm, even at f8, the DOF isn't that great. Being picky on CoC (0.005mm), DOF is narrower than I thought:

Focal length	300.00
Aperture	8.00
CoC	0.005
Hyperfocal dist	2250.30
Focus dist	100.00	500.00	1000.00	2000.00	3000.00
DOF	8.91	233.73	1107.47	16921.93	∞
Near	95.75	409.10	692.34	1058.89	1285.81
Far	104.65	642.83	1799.81	17980.82	∞

So, even focussing at 1km won't get things past 1.8km in focus. The hyperfocal distance is 2.2km. Even at f11, the hyperfocal distance is still 1.6km. Zooming out to 200mm, the hyperfocal distance at f11 drops to 700m, which is a significant difference.

In fact, looking at the equations, the hyperfocal distance is proportional to the square of the focal length, so it's easy to see how increasing focal length rapidly leads to very narrow DOF.

On top of that, there is definitely some field curvature going on which exacerbates the problem.

So, bottom line, at these long focal lengths, getting everything in focus on landscape type shots isn't easy. A narrow aperture is certainly sensible.

Thanks for the technical info Paul! That's really interesting.

Over time, I have become less concerned about edge/corner sharpness when viewed at regular sizes on a screen. If the main subject is sufficiently sharp, I am happy with the image. That long-range train photo I posted yesterday has smeary foliage on the bottom left-hand corner. There was a bit of swirly wind around the escarpment so it could have been that but I suspect it is just field curvature. In the end, it doesn't detract from the purpose of the shot and the client (basically me!) is happy enough.

Here is another shot at 300mm where the foreground looks even on both sides but you can see the heat haze kicks in right behind the main subject (loco 8201) and it's quite significant. The shot was framed this way (i.e. subject on the left) to avoid a large power pole that would have been right in the middle of the frame had I composed it differently. The problem with train photography is that sites are restricted and fenced off so you have to take shots from locations and angles that aren't ideal.

PN Depot at Port Kembla by Peter Watters, on Flickr

 

[pdk42]

Well, your roving reporter on 70-300 focusing here again...

Looking again at the various images from the 70-300, my conclusion is that the problems I'm having are definitely focus related since I'm getting very variable results. Sometimes I'm uber impressed with the images, other times I'm distinctly unimpressed.

The problem is much worse at the long end of the zoom range and at large subject distances (typically 500m or more). My suspicion is that the precision needed from the focusing elements at long distances and long focal lengths is very high. I reasoned that the focus type might matter, so I did some experiments with the S1R. I used five different focus techniques:

- SAF using the "1-Area" mode and the smallest target area
- SAF using the "Pinpoint" mode and the smallest target area
- CAF using the "1-Area" mode with the smallest target area
- SAF using the "Zone oval" mode with a big chunk in the middle of the frame
- MF

In each case, I took 10 shots of some trees about 500m away against a grey sky. Focus was re-acquired between each.

Here's a view of the scene (un-cropped image). Focus was on the branches at the top middle against the sky.

• Panasonic - DC-S1R
• LUMIX S 70-300/F4.5-5.6
• 300.0 mm
• ƒ/8
• 1/500 sec
• Pattern
• Auto exposure
• ISO 3200

The camera was mounted on a tripod and set as follows:

- ISO 3200 (to get a high shutter speed)
- f8
- 1/500s

I classified the resulting shots either as:

- Fully sharp
- Slightly blurred
- Unacceptably blurred

This is "fully sharp":

• Panasonic - DC-S1R
• LUMIX S 70-300/F4.5-5.6
• 300.0 mm
• ƒ/8
• 1/500 sec
• Pattern
• Auto exposure
• ISO 3200

This is "unacceptably blurred":

• Panasonic - DC-S1R
• LUMIX S 70-300/F4.5-5.6
• 300.0 mm
• ƒ/8
• 1/500 sec
• Pattern
• Auto exposure
• ISO 3200

"Slightly blurred" is somewhere in between.

These are my results:

Focus mode	Fully sharp	Slightly blurred	Unacceptably blurred
SAF using the "1-Area" mode	3	4	3
SAF using the "Pinpoint" mode	5	5	0
CAF using the "1-Area" mode	5	4	1
SAF using the "Zone oval" mode	6	3	1
MF	2	5	3

I found the results quite illuminating.

- SAF using the 1-Area mode is by far the worse of the AF modes. This mode has so far has been my "go-to" AF mode. I'll be changing that!
- SAF with pinpoint is the best, but not significantly better than some other methods
- CAF with 1-area works better than SAF 1-area, but the viewfinder shows a slight blurring when shooting which is a bit disconcerting (I guess DFD at work)
- SAF with an oval zone works better than with the 1-area mode (but obviously it's not great if you need a precise focus point)
- MF was very difficult. I used max magnification and had the ring set non-linear mode. But the slightest movement would change the focus a lot. I couldn't get consistently sharp focus.

So, my conclusions are:

- The 70-300 is actually quite a sharp lens at 300mm
- But focus accuracy is a problem at long distances
- The best mode is SAF + pinpoint
- Shooting several shots with focus re-acquisition is recommended to ensure at least one keeper

I'll repeat this at some point with the S5ii and see if it's any different.

 

[CharlesH]

The best mode is SAF + pinpoint

Very interesting, thanks for doing this. I quit using SAF quite a while ago and only use CAF. I don't remember why, probably some comparison testing with a Sony camera after their CAF got so good. And I've shied away from SAF+ on Panasonic cameras under an assumption that a little larger contrast area would be better. But your results look very good and I'll check out SAF+ with the 70-300mm.

 

[kbouk]

Pinpoint mode is the slowest but also the most accurate focus mode with all Panasonic bodies (m43 & FF) even in low light.

 

[RuleOfThirds]

For my 70-300 pictures in the March thread I used the animal detection.That worked quite well for me. Maybe we can ask Panasonic to add tree branch detection and train detection as well . But seriously, is a roof top, or some kind of man-made-structure (distant church tower?) not a better option then moving branches? Probably has more contrast too. Ofcourse in nature there is no such thing... Today and next couple of days it will be dark and rainy, so can not test it for you, probably not this week anyway.


I only have S5ii, so can not test on soley on DfD. Did you try it on your s5ii as well ? Hmm I don't know if the 24mp can resolve those branches at that distance at all...

 

[pdk42]

I repeated the test with the S5ii. These are the results:

Focus mode	Fully sharp	Slightly blurred	Unacceptably blurred
SAF, 1-area	0	5	5
SAF, pinpoint	9	1	0
CAF, 1-area	0	4	6
SAF, oval area	10	0	0

So, a complete fail for both SAF and CAF using the small 1-area zone.

Pinpoint and SAF oval area both did very well.

 

[pdk42]

For my 70-300 pictures in the March thread I used the animal detection.That worked quite well for me.

For subjects at relatively close distances I've not had a problem. I suspect slight focus shifts only mean, say, that the ears of the dog get focus and not the eyes which the user would probably put it down to "just one of those things"; whereas with subjects (like trees) at close to infinity any small drift makes the whole image look awful.

Maybe we can ask Panasonic to add tree branch detection and train detection as well .

Ha ha - that would be an interesting feature.

But seriously, is a roof top, or some kind of man-made-structure (distant church tower?) not a better option then moving branches? Probably has more contrast too.

I think in principle that trees against a grey sky are about the best target for AF - certainly for SAF.

Ofcourse in nature there is no such thing... Today and next couple of days it will be dark and rainy, so can not test it for you, probably not this week anyway.

Dark and rainy shouldn't stop us getting the camera out!

I only have S5ii, so can not test on soley on DfD. Did you try it on your s5ii as well ? Hmm I don't know if the 24mp can resolve those branches at that distance at all...

See the post above for the S5ii equivalent.

 

[RuleOfThirds]

Dark and rainy shouldn't stop us getting the camera out!

True, but I have the tendency to melt in rainwater. It runs in the family I guess. But storms and bad weather can give incredible skies….

 

[RuleOfThirds]

So the S5ii is better and worse with AF. Because you had 11 unacceptable focus. vs 5 from the S1R if not looking at MF.

But good to know and thank you what focus mode to use for landscape. I never tried the oval. Normally I use one area…

 

[Quentinquirelino]

I repeated the test with the S5ii. These are the results:

Focus mode	Fully sharp	Slightly blurred	Unacceptably blurred
SAF, 1-area	0	5	5
SAF, pinpoint	9	1	0
CAF, 1-area	0	4	6
SAF, oval area	10	0	0

So, a complete fail for both SAF and CAF using the small 1-area zone.

Pinpoint and SAF oval area both did very well.

Interesting. Could it be that the focus area for 1-area was quite difficult for the camera for some reason? I can't figure another reason, why the 1-area modes should be much worse, no matter if SAF or CAF.

Maybe you also should try other CAF modes.

Thank you so far for the comparison.

 

[AlanC]

In SAF mode, have you tried changing the size of the AF area?

I ran into what I'm guessing is the same issue with my S5 and 70-300, typically with the lens at 300mm when my subject was around 150m-250m away: I'd end up taking 4 or 5 photos to get one that was properly sharp. What I found was that a larger AF area will focus faster, but a smaller area will be more accurate. In a way this agrees with using the oval area because the individual AF areas within the oval are slightly smaller than the minimum size you can set manually.

 

[pdk42]

In SAF mode, have you tried changing the size of the AF area?

No, but that’s a good point. I’m using it on its smallest size - because I want to know what I’m focusing on!

I ran into what I'm guessing is the same issue with my S5 and 70-300, typically with the lens at 300mm when my subject was around 150m-250m away: I'd end up taking 4 or 5 photos to get one that was properly sharp. What I found was that a larger AF area will focus faster, but a smaller area will be more accurate. In a way this agrees with using the oval area because the individual AF areas within the oval are slightly smaller than the minimum size you can set manually.

Glad I’m not the only one. Unreliable SAF is a cardinal sin in my view. I don’t mind if the AF struggles, or even fails - but giving focus confirmation when the image is blurred can absolutely ruin a unique shot that can never be repeated.

 

[AlanC]

The only other thing I can add is that the problem does seem limited to medium distances at the long end of the zoom range. When I took some photos of Sandsend bay from Whitby beach, a distance of around 2 miles, they were all consistently sharp. And when I've used the lens for macro photos at 300mm, close to MFD, it's worked consistently well.

 

[Pete_W]

Maybe we can ask Panasonic to add tree branch detection and train detection as well

I know you're joking but I'd like the train detection!!!