[Yakim]

[quote name='boren' timestamp='1300208815' post='6822']

Most user reviews in dyxum.com say that the Tamron 60/2.0 is slow to focus. The Canon 60/2.8 on the other hand is known to be one of the faster-focusing macro lenses around (at least according to user reviews in fredmiranda.com). Testing the Tamron on the EOS-7D will give you a better point of comparison.



The kit lenses are a similar story. One of the users at dyxum.com had this to say about the Sony 18-55 SAM: "Extremely poor auto-focus (slow and unreliable in anything but good light)". This is from a user who also used the following lenses: Sony 16-105, Sony 18-70, Minolta 17-35, Minolta 20-35, Minolta 24-50. He has enough experience to compare its performance. Another user had this to say about the lens: "slow to almost none af in low light". Among the 44 review at fredmiranda.com I couldn't find a single one that complained the the Canon 18-55 IS is slow to focus. In fact, quite a few of them raved about its fast and accurate AF.



Bottom line: SAM motors are very different than SSM motors. The former are a marketing trick. They are slow and noisy and were only released to the market by Sony as a cheap answer to "those in the know" (e.g. salesmen and Canikon FUD-spreading fanboys) that claim screwdrive AF is a disadvantage. Tamron isn't known for fast focusing lenses and the 60/2.0 is a typical example, if not the worst. Both Canon lenses you've used are much better, and this means that your findings, as interesting as they may be, are not really valid. At least not as a comparison of the AF capabilities of these cameras.

[/quote]



1. I guess it all depends on your PoV. As I have many lenses with ring-USM, IF design and non-rotating front filter I can not define the AF of the 18-55 IS as fast. However, that of the 18-55 SAM was much worse.



2. The problem was less AF speed per-se but the inability to lock focus fast. In simple words: They hunted a lot. That is why I suspect that more than it's the lens' fault, it's a body related issue e.g. AF algorithm. The fact that I saw the same issue with both lenses only strengthens my suspicion.



3. I agree with you that the best way of testing this is with the same lenses in different mounts e.g. Tamron 60/2 in SA and EF mounts but as I said, I'm no lens tester. I'm just a user. All these lenses and bodies were bought from my personal budget for my personal use. On paper, the A33 had many advantages. In practice, it disappointed.



4. I must confess that the Tamron 60/2 in SA mount was a dream combination of mine for a very long time. I love fast aperture and I love stabilization. The 7D + 60/2.8 combo gives me none of these. Again, I witnessed that theoretical advantages do not always translate to practical ones.

[Brightcolours]

As promised, I made a few sequences of cars moving at around 50 km/h, trying to stay as close as possible to what imaging resource in their review of the Sony A55v described they did to test AF tracking.



Their description:

Sony A55v with Sony 70-200mm f2.8 G SSM lens, one of Sony's top lenses.

Focal length used: 150mm

Shooting distance: "perhaps" 70 feet (about 21 meters) to "perhaps" 15 feet (about 5 meters).

Subject: A car traveling about 30 mp/h (about 50 km/h) towards the camera.



What I did:

Canon EOS 450D with Canon 70-200mm f4 L USM. Not the Canon 70-200mm f2.8 L IS USM II, as it is the cheap f4 lens that I own. I do realize that the f2.8 lens focusses faster AND that it offers the f2.8 advantage to the AF sensor, but it is all I have available to me <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/tongue.gif' class='bbc_emoticon' alt='' />.

Focal length used: 145mm (lens only has 70/100/135/200mm markings, I ended up actually using 145mm.. no big deal, as the crop factor of the cameras is a bit different anyway, and 145mm is pretty close to 150mm.

Shooting distance: starting at 20 meters.

Subject: cars traveling at about 50 km/h towards the camera.



The IR crops can be found on this page:

http://www.imaging-resource.com/PRODS/AA55/AA55A.HTM

Crops specifically:

http://www.imaging-resource.com/PRODS/AA55/zA55focustrack.jpg



Here are my results (also 100% crops):





What we notice right away is that my crops are bigger, and the car logo's are bigger too. The Sony has a higher resolution (14mp versus 12mp), so the only conclusion can be that they shot the car from further away than they suggest in their texts. Advantage for the Sony: car further away means less taxing job for the AF.



What we notice about the IR crops:

first 4 are "green", pointing to back focus. Last 4 are purple, pointing to front focus. 5th, middle crop appears to be best in focus.



My crops:

Lexus BMW-3 series clone.

1st back focus.

2nd in focus.

3rd in focus.



VW fox

1st: back focus.

2nd: in focus.

3rd: back focus.



Audi A4:

1st: in focus.

2nd: back focus.

3rd: in focus.



The A55v does 10fps, the 450D a bit over 3 fps. So, a series of 3 with the 450D is within the same duration of the 9 crops from the A55v.



When I compare the crops, to me it looks like the 450D does at least equally as good, even though the light conditions were a lot less bright, the lens only was an f4 lens, and the shooting distance was closer.



This all the more points to that the IR phrase "We shot side by side with the Sony A55 and a Canon 7D, and felt that the A55's AF performance here was pretty similar to that of the much more expensive Canon." should make one a bit skeptical, especially since people who have gone from a 450D or alike to a 7D to comment on the superior AF tracking abilities of the 7D.



My conclusion: The IR "claim" should not be used as evidence of the AF tracking of the A33 and A55v. The crops provided for sure are not impressive.

[Guest]

Not sure if this proves anything (apart from that Canon camera prefers Lexus), but just out of interest:



Judging by those crops on the IR site, my math (which can be wrong obviously) tells me that car at the beginning of the sequence was around 26 meters away, which actually *should* help the AF, making the DOF wider.

[Brightcolours]

[quote name='Lomskij' timestamp='1300306161' post='6884']

Not sure if this proves anything (apart from that Canon camera prefers Lexus), but just out of interest:



Judging by those crops on the IR site, my math (which can be wrong obviously) tells me that car at the beginning of the sequence was around 26 meters away, which actually *should* help the AF, making the DOF wider.

[/quote]

How did you calculate that? The 70 foot comes to 21 meters here (using a calculator as I am not familiar all that much with feet <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/wink.gif' class='bbc_emoticon' alt='' /> ), and I can not, from the crops, determine 26 meters. A foot is 0.3048 meters.



The Toyota (IR) obviously was further away than the Lexus (BC), and that should help the Toyota crops in 2 ways: DOF is a bit deeper and the AF has it easier when a subject is further away.



But, on DOF, I shot with f4 (lens is not faster. This makes the AF having a harder time, though. I also am not entirely sure about the aperture used with the Sony... You can not actually choose the aperture, and if it is like with the video mode with AF, the aperture is actually set to f3.5. Which then does not make a big difference to f4.



All in all, I can not conclude, from the IR crop samples, that the A55v (and therefor the A33) have a very capable AF tracking implementation.

[Guest]

[quote name='Brightcolours' timestamp='1300307391' post='6885']

How did you calculate that? The 70 foot comes to 21 meters here (using a calculator as I am not familiar all that much with feet <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/wink.gif' class='bbc_emoticon' alt='' /> ), and I can not, from the crops, determine 26 meters. A foot is 0.3048 meters.



The Toyota (IR) obviously was further away than the Lexus (BC), and that should help the Toyota crops in 2 ways: DOF is a bit deeper and the AF has it easier when a subject is further away.



But, on DOF, I shot with f4 (lens is not faster. This makes the AF having a harder time, though. I also am not entirely sure about the aperture used with the Sony... You can not actually choose the aperture, and if it is like with the video mode with AF, the aperture is actually set to f3.5. Which then does not make a big difference to f4.



All in all, I can not conclude, from the IR crop samples, that the A55v (and therefor the A33) have a very capable AF tracking implementation.

[/quote]

Well, it was rather approximate calculation - the size difference between the first crop and the last crop is 2.1x, so taking that car's speed was constant at 14 m/s and those frames were all taken in 1 second, you have to start approx in 26m from camera to make it look 2.1 times larger in 1 second (as it was tele lens used, it has to be fairly linear, or I'm wrong here?).



Regarding the aperture - bokeh fringing is quite apparent, so it has to be rather wide.

[Guest]

Funny "test." Different cars, different lighting, different cities, and probably even different countries!!, and he comes up with his "conclusion." IR tested 7D and A55 side-by-side, and there is no reason to claim that they are lying except bias. A55 AF tracking is above average, acccording to IR side-by-side test.

[Brightcolours]

[quote name='oneguy' timestamp='1300357692' post='6896']

Funny "test." Different cars, different lighting, different cities, and probably even different countries!!, and he comes up with his "conclusion." IR tested 7D and A55 side-by-side, and there is no reason to claim that they are lying except bias. A55 AF tracking is above average, acccording to IR side-by-side test.

[/quote]

I do not see how "different cars" and "different cities" would make the test invalid. I actually shot 3 different cars, and as you can see, the results are relatively consistent.



The light was different, yes. In the IR test's advantage.

The distance was different, yes. Because the distance IR said they did it in was incorrect. And that too was an advantage for the A55v.



So, 2 things to the advantage of the A55v, yet the results are WORSE than the results for my 450D. And then there is the fact that I had an f4 lens, also an advantage for the A55v's AF sensor.



So. worse reults, but conditions that were in its advantage.



Now about IR:

- They claimed the most OOF shot was camera moved, not OOF. The green LoCA shows it was severely back focused. The fact that it was a shot from a sequence of 9 within one second, shows it is nonsense to claim that shot was "moved" (other shots had to be moved then too). The fact that the lens was wide open, and the light was bright, and the focal length was only 150mm, indicates camera shake was never a reason for the image to be that soft. And the fact that the A55v has in body IS, also makes camera shake very unbelievable. So, the camera shake excuse was a lie.

- They claimed they shot from 21 meters, from my shots (starting at 20 meters) it is clear that was false.

- They claimed they "felt that" the A55v's performance was pretty similar to that of the 7D. "felt that". They did not find that, they use "felt", as some obfuscating politician.

- They never show any comparison, yet they claim a comparison. I have never encountered that in any source, be it in print or online.



Many reasons do doubt the claim from IR, not in the least the shown crops from the A55v with 4 back focused shots and 4 front focused shots, with 2 shots very OOF.



You dismissing my test as "funny" shows one thing, bias. A shame, I had hoped for a more serious response.

[Yakim]

[quote name='Brightcolours' timestamp='1300360485' post='6898']

You dismissing my test as "funny" shows one thing, bias. A shame, I had hoped for a more serious response.[/quote]



Still? You're optimistic by heart I see.

[Guest]

IR tested 7D and A55 side by side and concluded the result were comparable. You did not. You don't even own the camera and have done "ZERO" tests. There is no reason to doubt the review (by someone who has hasn't even seen the camera) except due to bias.

[Brightcolours]

[quote name='oneguy' timestamp='1300429344' post='6925']

IR tested 7D and A55 side by side and concluded the result were comparable. You did not. You don't even own the camera and have done "ZERO" tests. There is no reason to doubt the review (by someone who has hasn't even seen the camera) except due to bias.

[/quote]

I have made very clear all the doubts that that silly paragraph in IR's "review" raise.



I even went through the trouble to replicate the so called "test" with my own camera, which resulted in more difficult circumstances for my camera (apparently, the IR reported distance was wrong, so I shot closer by than they did, closer by is harder on the AF... and the light I shot under was a lot lower).



I explained why the claim of "camera shake" for one of the crops is just plain nonsensical. I pointed out that the LoCA colours green and purple tell on the back and front focus of 8 of the 9 crops, so even debating whether something is sharp or not has little merit. A very poor result they show, for a not very taxing (for the AF) test.



If you want to believe that a Canon EOS 7D can not do better than that, fine. That is your problem. I showed that my 3 year old 450D does a better job at tracking focus, and anyone with a little bit of real knowledge about the 7D knows that it will better the by IR shown A55v crops anytime.



So that you keep repeating your rhetoric only shows one thing: a strange bias and a need to defend the A33/A55v against whatever is being said.



Anyway,this all started with me disagreeing with the colclusion of the photozone review that the A33 was a good choice for sports photography, because its AF tracking is not upto the job to match the 7fps (6 and 10 fps for the A55v) in making it a good sports camera (and that the slide show during the high FPS modes shows where the subject was before, not where it is now).



DPreview finds the AF tracking to be lacking, IR's crops find the AF tracking to be lacking. I showed that the 450D, even with a lesser lens, is better capable of tracking a subject moving towards the camera.