Camera/lens focus speed test (also about 20mm F1.7)

Started Mar 17, 2013 | Discussions thread
Anders W
Forum ProPosts: 18,388Gear list
Like?
Re: Update
In reply to micksh6, Aug 27, 2013

micksh6 wrote:

Anders W wrote:

micksh6 wrote:

Anders W wrote:

micksh6 wrote:

Anders W wrote:

micksh6 wrote:

Anders W wrote:

micksh6 wrote:

Anders W wrote:

micksh6 wrote:

The light was ISO 1600, 1/20s and F2.0 for the main target (yellow pig or chicken), slightly lower for secondary target (printed high contrast repeatable logo).
The reason for such low light is that with brighter light (1/30s, same ISO and F number) E-PL5 doesn't consistently turn on AF assist light and I needed AF light to measure AF speed. The AF light is covered by blue electric tape.

What light source(s) did you use to illuminate the scene?

A 6-head tungsten lamp over dining table on the left. It was on previous videos, 6 diffused bulbs on about 2 feet wide chandelier.

On the left? But it is clear from the videos (shadows), that the the light is coming in from the right. So what you are showing now is a mirror image of the actual scene or what?

On the left from the camera view. What you see on video is E-PL5 display tilted 180 degrees up (something that E-M5 can't do). In this case E-PL5 display shows mirrored view. It's useful trick in many situations, but here it helps because I wanted to capture both AF assist light and display.

I realized that what you showed was the E-PL5 display tilted 180 degrees but didn't know it would show a mirror image if you do that.

And if the light is constant, how do you explain the fact that the shadows are different. If you look at the shadow made by the "leg" under the shelf on which the primary target is standing, you can see that in the video with the 20/1.7, it is only about half as wide as the leg itself whereas in the video with the 17/1.8 it is nearly as wide as the leg itself.

The main focus target wasn't covered by any shadow. Camera metering showed the same light as well. But, I'm sure you will find lots of small details that could not affect test results by any means.

I think I (and others) can judge for ourselves what might and might not affect the results. So could you please just answer my question: Why is the shadow from the leg significantly different (nearly twice as wide in one case as in the other)?

What shadow and what leg? I got 2 videos for either of 17mm and 20mm. Could you point to a video file and approximate timestamp?

I mean the wodden "leg" extending downwards from the shelf the primary target is standing on and I mean the videos you linked to in your first update post, those that are meant to be fully comparable.

I moved setup slighly to make sure magnification is the same for all lenses, that's all that could affect test setup. Camera angle relative to main target (the yellow chicken) was the same and the target was at the same position relative to the light, it didn't move. This is all what should matter.

Magnification isn't quite the same either. The magnification of the wall segment of the scene (where the target is) is lower for the 20/1.7 than for the other two lenses. And if camera angle was the same and light was the same, how do you explain my observation with regard to the shadow?

First, in my setup it's impossible to change light position relative to target. The lamp hangs from ceiling and I never touched it. The shelf is a part of fireplace and belongs to the wall.

The magnification is the same for subject but since the wall is farther than subject the wall's magnification will be different. This may explain difference is shadow width.

The objects I was looking at for the purpose of determining magnification (the photographs, the curtain) are for all practical purposes in the same plane as the primary target. Consequently, the overall magnification is somewhat lower with the 20/1.7 than with the other two lenses.

As to the shadow, the difference in distance between the wall and the "leg" cannot explain why the shadow is significantly wider with the 17/1.8 than with the 20/1.7. If anything, you'd expect the opposite (wider shadow with the 20) if your explanation were valid. If lighting is constant, the only explanation left is that you shot with the 17/1.8 from a different angle than with the 20/1.7.

Take into account that E-PL5 LCD angle could be different. It wasn't at 180 degrees because it would be difficult to press magnification button. Also video camera angle relative to E-PL5 might change. Both of these could distort the scene on LCD, but none would affect results, of course.

To exclude this possibility, I compared, as already indicated, the width of the shadow with the width of the "leg".

As far as I can tell the angle was the same. But even if it changed this wouldn't make big difference. That's why I also shot couple of extra sessions, and when I misplaced AF box I also had to change shooting angle to compensate. Randomness of focus algorithm is more significant than tiny changes in scene.

So, if you are hoping that there is a magical "perfect" angle that makes 20mm lens fast, I'm afraid there is no such thing.

Look. If you want to make one lens look like it would be more prone to hunting than other lenses, it's very easy to arrange that by means of any of a large number of seemingly minor side conditions. As those have experimented with this themselves are already well aware, the changes with regard to target and lighting required to go from consistent AF locking, to border-line zone, to consistent hunting are extremely small.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Wow!New
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow