AI Servo Test CC Please

Sam · July 7, 2011

I got the 7D specifically for the improved AI and higher frame rate to shoot fast moving subjects.

I went out this morning and practiced with the neighbors black lab. This was only to test the AI. I wasn't overly concerned with anything else.

I tried a variety of settings and it seems that using the AF point expansion, and leaving the IS on worked the best. I made a lot of user mistakes, but with practice I know I can improve the keeper rate.

In this posted example I used the Canon 70-200 mm 2.8L IS lens.

This was taken at 1/400 F8 ISO 800 @200mm. Many other images were taken at higher shutter speeds, up to about 1/2000.

The first is the uncropped image. The second is the cropped image.

For those of you using the 7D for this type of photography shooting sports, animals etc with the 7D (or other cameras) feel this is representative, or should I expect better? Remember I am only referring to focus. Composition, exposure, subject, etc is for another day.

Thanks!!

Sam

1.

2.

EDIT: The images are a bit crisper on my computer. Say about 12% better.

pathfinder · July 7, 2011

You were using AI Servo, weren't you Sam?

I found I like AI Servo with a choice of three AF points, either center weighted or right of left of center, or above or below center.

I never found great success with a Single AF point and AI Servo, but the choice with three simultaneous AF points active really worked well for me shooting whales in Alaska.

You were not using Single Shot AF I assume.

In bright sunlight with a dark dog, your AF should just crackle!! Were were shooting gray whales in overcast skies from moving boats, and three point AI Servo worked quite well with my 7D and a 100-400 IS L - which is about two whole stops slower than your f2.8 IS L. Two stops more light should make AF much, much faster too.

Sam · July 7, 2011

Pathfinder,

I was using AF point expansion (manual mode). I chose the center point and the 4 additional focus points were one on each side of the center point.

x
x X x
x
I was also using burst mode and using the back button for AI focus and tracking.

The one question I have is: In your opinion is this image sharp / clear / in focus enough to satisfy you? Is this typical results? Assuming I develop the skills to get say a 70% keeper / in focus rate with the quality displayed here, would this be acceptable or would you look for better image quality?

Thanks!!!

Sam

pathfinder · July 7, 2011

I did find micro adjusting can pay real dividends on the 7D also

THis image isn't terrible, but maybe not what I would think is ideal when I look at the cropped image. Your 70-200 F2.8 IS L should be critically sharp I would think. I am, however, probably more critical than most clients, Sam.

Has it had any sharpening in RAW processing or in camera processing at all? That can make a significant difference in the apparent sharpness of an image as well.

This image was shot with a 7D and a 100-400 IS L, which most would agree is not glass of the quality of your 70-200 f2.8 IS L.

Here is Neva after her mud bath - same camera and lens I believe

My 100-400 did have AF micro-adjusting for my 7D as well ( by me ) - it did make a noticeable difference for that lens combination. Also, correcting for Chromatic aberration in ACR can make a noteworthy improvement as well which as part of my RAW workflow

Bear in mind my images have been sharpened and lens corrected in ACR as well, however.

Since you posted the images and asked if they are sharp, I suspect you already are suspicious your self, Sam. Correct??

Sam · July 7, 2011

Pathfinder,

I want you to be critical. I am critical of my own stuff, but can't see with others eyes. This ain't about no stinking clients!!

If I like it the clients will like it except in some odd ball circumstance.

This is about learning and getting better. I will not be happy with average or satisfactory.

These images were shot in jpg, and have been sharpened etc. You are seeing pretty much the best image quality, sans Herculean processing.

I wanted to try shooting in jpg. That should be fine for determining if the focus is good. Also I could try 8 frames a second for an extended time. It be like a machine gun.

The eagles head and beak look good. The wings are out of focus DOF, which is normal.

The dog looks great. Was he moving? Was this taken in a burst series using AI Servo? Or was he sitting there?

I guess I am trying to confirm my thoughts but don't want to chase ghosts.

I will go ahead and do a focus test with the 7D and 70-200 2.8 L IS, and make any micro adjustments needed.

Let you know in a day or two.

Sam

pathfinder · July 7, 2011

No, Neeva ( the Lab) was not moving, but I do have lots of moving targets that I did shoot - It was just a dog image like yours was a dog image.

If your images are sharpened jogs SOOC, I would expect a bit better - yes.

In my gallery shot in Alaska, there are a number of images of whale tails that were all shot in high speed frame rate, with AI Servo, as in camera jpgs with my 7D and 100-400. They were not shot in RAW because I got tired of buffer overflow too often. They were sharpened significantly in LR3

Check your AF micro adjustment

Sam · July 8, 2011

Jim,

You need to give the poor dog a bath.

Sam

pathfinder · July 8, 2011

Neeva just loved diving into mud bogs for sticks. She was fearless.

kdog · July 8, 2011

pathfinder wrote: »

This image was shot with a 7D and a 100-400 IS L, which most would agree is not glass of the quality of your 70-200 f2.8 IS L.

I've found the 100-400 to be considerably better in the 201-400 range.

pathfinder · July 8, 2011

You know Joel, you may be right. I cannot really remember using my 100-400 at anything but 400, or very close to 400. If I need less than 300, I would switch to my 100-300 IS L, or Sam's 70-200 IS L.

Sam · July 8, 2011

Good news bad news............................

Recently I have only used the 70-200 2.8 L in difficult lighting conditions. I haven't been ecstatic with the results, but chalked it up to the shooting conditions. I tend to use primes whenever I can.

For the tracking test I only used the 70-200 2.8L.

I spent the better part of today with a Lens Align, my 5dII and the 7D.

The darn 70-200 2.8L can't get a clean image at any focus point. Moving or still!!! Micro adjusting doesn't help. It just moves the least fuzzy area to the front or rear. It does the same thing on my 5D II.

Technically it be broke.

Monday it will be off the the Canon doctors.

Then the AI tracking saga will continue!!!

Sam

pathfinder · July 8, 2011

I am sorry to hear that, Sam.

But it is good to know that it can be restored to white lens goodness.

Sam · July 13, 2011

Latest update:

FYI: I sent my 70-200mm 2.8 L IS in to Canon Last sat. It arrived on Monday. Canon did whatever and shipped it back on Tuesday, and it's scheduled to be delivered today (Wednesday). Hard to ask for more!

I went out to a local foot race on Sunday, and my results were better with my 24-105mm L 4.0 IS, but not great. I am looking for great. Is that too much?

Yesterday I went to a local park and took some shots.

These are examples of what I would call success. Although I am still not satisfied with the overall keeper rate, camera settings and technique.

Suggestions and comments are requested. I am hell bent on learning how to capture sports / motion.

I will update this with my 70-200 2.8 L results when they are available.

Thanks!!!

The first example is an uncropped image of a basket ball about to hit / or bouncing off the rim. Looks good. 1/1000 f5.6 AP flash fired.

The second is a 100% crop of the same shot. To me this is good.

This is the 5th image in a burst. AI Servo used to achieve focus and track.
Full un-cropped: 1/320 f5.6 no flash

Here is the 100% crop.

This last example I another bike rider. this is the fourth image in a burst.
1/800 f5.6 no flash.

Here is the 100% crop of the same image.

pathfinder · July 13, 2011

These look great on my iPad, Sam.

The strings of the net and rim may be a touch sharper than the ball, but that would not be that unexpected, since they are closer and brighter. Well done.

Brandon Hanley · July 13, 2011

looking at 100% crops nothing is going to be as sharp as you want it to be. loved the ball on the rim shot btw.

Sam · July 13, 2011

Thanks for the feed back.

Added in: I should have stated that I am not concerned with exposure, composition, backgrounds, etc. on these images. The only priority I have is getting consistent sharp images. When I can do that I will address the other issues.

As to 100% crops not being sharp................

Here is the kind of quality I am seeking to achieve with sports / motion subjects.

Un-cropped image fro 5D I.

Here is the 100% crop.

I not only want to see the dog. I want to see the flea on his back. When I get to that point I want to see the flea's eyebrows!

Sam

Sam · July 13, 2011

The saga continues.............................

The nice FedX man just delivered the lens! Yeah! Box appears undamaged. Yeah!

Lens cap was crushed into the top ring where the filters screw in resulting in the top ring being dented to the point where a filter can not be screwed in.

Upon close inspection I did see a slight bump / wrinkle in one corner of the box, but the box damage I saw should not have resulted in any visible damage to the lens.

Lens is damaged! Boo! Waiting for return shipping label. Boo!

How come nothing seem to go as planed? Sigh..................:D

Sam

pathfinder · July 13, 2011

Wow, Sam, that sucks.
I have shipped camera bodies, lenses and flashes to Canon factory service in NJ, and a lens to Tamron service, as well, and all escaped unscathed. Sorry to hear about your misfortune.

Sam · July 13, 2011

pathfinder wrote: »

Wow, Sam, that sucks.
I have shipped camera bodies, lenses and flashes to Canon factory service in NJ, and a lens to Tamron service, as well, and all escaped unscathed. Sorry to hear about your misfortune.

In the scale of life, it's just a little thing.

Sam

kdog · July 14, 2011

Where did you send your lens? That lens sounds like it was dropped before it was put in the box. Very disappointing behavior.

Funny story. Shortly after buying my 20D and my first few lenses, I had some AF issues and Canon decided I should send in my body and three lenses to get checked out. I was living in NY at the time so drove the whole mess to Canon. A few days later when it was ready, I was standing at the service window waiting for the tech to find my gear. He eventually came trotting in quickly juggling my 20D body and all three lenses in his arms, and practically had to sprint to the bench to set them down before he dropped one. I almost had a heart-attack. They actually had a stack of baskets nearby for carting gear around but he elected not to use one. I guess when you're around expensive camera gear all day you get pretty cavalier about it.

ziggy53 · July 14, 2011

Sam wrote: »

Thanks for the feed back.

Added in: I should have stated that I am not concerned with exposure, composition, backgrounds, etc. on these images. The only priority I have is getting consistent sharp images. When I can do that I will address the other issues.

As to 100% crops not being sharp................

Here is the kind of quality I am seeking to achieve with sports / motion subjects.

Un-cropped image fro 5D I.

Here is the 100% crop.

I not only want to see the dog. I want to see the flea on his back. When I get to that point I want to see the flea's eyebrows!

Sam

What lens did you use to make this image?
(Nevermind, thanks for posting the link to the image and keeping full EXIF. It appears that you used the Canon 24-105L, at 105mm and f11.)

Did you use a tripod?

Sam · July 14, 2011

Ziggy,

This was taken so long ago I can't remember. In reality if I took it last week I might not remember.

Looking at the EXIF data I believe I used a studio strobe and most probably I did use a tripod. Without the use of a strobe the shutter speed would have much been longer.

Sam

ziggy53 · July 14, 2011

Great Sam.

What was your technique for the 'dog running towards you' image? How much time did you give AI-Servo to latch, for instance?

Sam · July 14, 2011

ziggy53 wrote: »

Great Sam.

What was your technique for the 'dog running towards you' image? How much time did you give AI-Servo to latch, for instance?

Again...the neurons don't fire like they used to.

That said, as I recall, I was trying to lock on to the dog and follow him while I shot. I used the back button for focus. This wouldn't have been the first in a series.

It was shot with my 70-200 which according to Canon after the first repair, was out of spec. So I am going to go ahead and ignore anything shot with this lens until it is repaired.

I had better luck with my 24-105 at the park, but need to work on my technique.

April suggested shooting full manual mode as apposed to aperture priority. The theory being not to let the camera think about making any auto changes on the fly and slowing down the focus tracking.

That give me a thought.....................(5 minutes later. Went outside and preformed a quick and dirty rest.) Yes when using AI and AP once you lock onto a subject if the lighting conditions change the camera will adjust the shutter speed on the fly. I don't know if this will affect, (delay), the AI tracking or shutter lag time, but at the rate things move it wouldn't take much to put the subject slightly out of focus.

My next test will involve using my 135 2.0 L, shooting in manual mode and using as wide an aperture / ISO as needed to raise the shutter speeds to say 1/1200 to 1/2000 and see if this helps.

I have noticed that I have been ISO spoiled with my 5D II versus the 7D.

Sam

ziggy53 · July 15, 2011

The Canon 7D does have 2 processors and I believe that during focus operations one of the processors is dedicated to AF operations (before capture). It still takes around 1/2 second for AI-Servo to sample the rate and vector of motion of the moving subject. If you plan on that time with a half-depress of the shutter button or with a depress of one of the rear buttons (depending on the CF IV-1 settings) then either the first or second image should be in sharp focus and approximately 80-85%* after that should be in sharp focus in good light, depending partly on subject matter, your ability to hold the selected AF point on target, AF sensitivity settings, etc.

The AF section of the 7D is coupled to the metering section to some degree and both low-light levels and difficult lighting can impact on AF acquisition. (I realize this seems to somewhat contradict what I said above, but the processors and software, while capable of autonomy of duty, are still linked in tasks through their programs.)

I do recommend double-checking CF III-2 and making sure it's set on "0".

The following is an excellent primer on (older) Canon AI-Servo design and operation (not necessarily directed at Sam, but anyone trying to follow the logic):

http://photonotes.org/other/ai-servo.html

*Best case scenario.

Sam · July 15, 2011

Ziggy,

I just got the camera out and double-checked CF III-2 and made sure it was set on "0".

Thanks for the link. I think I understand the limits of AI Servo a little better. I always thought it could track a jet aircraft at near the speed of sound.

Who knew it had severe speed limitations?

No wonder all the UFO photos are fuzzy.

I am going to the zoo and see if I can find a turtle.

Sam

AI Servo Test CC Please

Comments