My editorial comments on a bad review - mostly about AF

[bencoyote]

Advertisement (gone after registration)

In their editorial on the Leica Q they really seem to dislike the Leica T

http://www.dpreview.com/articles/7847172667/opinion-why-i-want-a-leica-q

 

I have an M with a 50mm Summilux and I know that part of the reason it isn't my "main" camera is that while the 50mm is my preferred lens choice, many times I want a different focal distance and right now, honestly, I can't afford it. As my budget and skills improve, I see myself growing into the M. I also find the T being smaller and lighter is easier to take when I go out riding my bike or running.

 

I am really surprised by how negative the reviews are of the T are. I believe that it has some faults but I really like it and feel like it is deeply misunderstood. Probably the biggest misunderstanding is with regards to AF. People describe it as slow. I also have an Olympus E-M1 and yes it is faster. However, what I learned from the T is how much I was dependent on AF when I didn't need to be and even when it got in the way. I would take 3-4 shots trying to get the right AF point with the UI but it really didn't matter because they were all beyond hyperfocal distance. In the E-M1 and other cameras that I have had, there really wasn't a way to know this.

 

Then there is the question of salience with AF. The camera really has no idea what you are going to focus on. For compact cameras with a large DOF and where people are less likely to be creative with composition, there are several heuristics that get it right a surprisingly large amount of the time.  With the E-M2 AF caused me to miss many shots because the camera would select the wrong AF point or I wouldn't be able to communicate to the camera, this is the thing that I want to focus on. In the fractions of a second that I have when trying to capture the decisive moment, there is no time compose, communicate to the camera the salient element that it should focus on, and then wait for it to focus. For capturing those really short time intervals that are decisive moments, you really need to use MF and some sort of zone focusing or be pre-focused. The rest is up to you as the photographer. The AF speed is fine and often times handy when I don't really care about capturing a moment. I would argue in those cases where people believe that AF speed matters, the photographers need to ask whether the problem is AF or communicating salience. It is the combined process of composing, communicating the salient objects in your composition to the camera, focusing on them and then, and then releasing the shutter. It is that whole process that needs to be considered not just how fast can the lens home in on maximum contrast within a box. 

 

Really, I think photographers need to consider why they are not using MF. It is a freaking APS-C sensor and the fastest native lens is 2.0 and it has fairly good high ISO performance, you can't zone or pre-focus that? That is a problem with the photographer not the camera. Start with the old National Geographic slogan "f/11 and be there". On the T is more like "f/8 and be there". 

 

Honestly, the T has some problems with MF but to me these should be easy to fix in firmware:

1) There should be some way that the focus scale can be made always visible even in AF mode just like grid lines and the histogram. I want to constantly be reminded about my usable DOF and I don't want to momentarily lose my composition with the zoomed in view of focus assist just so that I can see the DOF and where I am in the focus range.

2)  It would be nice if the focus scale worked a bit differently like marked the minimum and max in focus as well as the focal point. In other words if they are not too close together mark the beginnings and end of the green bar.

3) Use shading to indicate how in focus things are. It is not 1.5m is in focus and 1.48m is out of focus illustrate that with shading between green and white.

4) It takes a bit of thought to interpret what appears to be logarithmic scales. I'm not sure how to solve this one.

 

The T lenses on the other hand could use some improvement:

1) Move aperture to all lenses and at least on prime lenses engrave a focus scale.

2) Don't have free spinning MF rings. Give them mechanical stops and possibly a focus tab. Muscle memory and relative position is really important. e.g. 1.5m is 6 o'clock 2m is 7 o'clock My eyes are and visual system are busy composing the shot.

 

I honestly do not understand his idea of "cluttered minimalism". Compared to the menus on every other camera the T is amazing. I think the editorial author has additional explaining to do. Finding any feature on the T is trivial and clear (except for playback which is trick you quickly learn). I do believe that they need to add just a couple of refinements to the UI (see my previous posts).

 

 

[AlanG]

Really, I think photographers need to consider why they are not using MF. It is a freaking APS-C sensor and the fastest native lens is 2.0 and it has fairly good high ISO performance, you can't zone or pre-focus that? That is a problem with the photographer not the camera. Start with the old National Geographic slogan "f/11 and be there". On the T is more like "f/8 and be there". 

 

 

I think that was never really the case and is a very dated way to look at things today. Perhaps if you are just doing fast "street" grabs you can zone focus. But today many of us have higher requirements because we examine our images more closely. I routinely study an image at 100% to fine tune focus when using a 17mm lens at f8. Yes zone focusing this manual lens gets me very close but I can see very slight focus fall-off if I don't nail it. Yes maybe others will overlook this but I am using high res gear and good techniques to get the most out of the gear.

 

I can't directly comment on the T but if other cameras do a better job of focusing, that is important to many users and for many uses.  Telling people they should use an AF camera manually seems silly.  Cameras compete just as much with technology as with design and lens quality. An oof focus image made with the best lens is still oof. 

 

I just bought my 13 yo stepson a $250 Panasonic Lumix FZ70 that has a 20-1200 equivalent lens. I tested it and was surprised how well it works for this incredible range when used within its limitations. (Don't expect it to compare to $5,000 primes on a full frame.) Like many other cameras today, it processes out the C/A and other lens defects automatically so this lens performs pretty well. I got sharp images handheld in sunlight at 1200mm... something that would be very difficult for me to acheive with a 35mm DSLR.  So I can't see any reason to make excuses for a much much more expensive T.

[bencoyote]

I don't think that I'm making excuses for the T as putting forth the hypothesis that AF isn't really as useful as people think it is.

 

When you take into account the whole process of getting the camera to lock onto the right focus point, or communicating the focal point, and all the shots you miss because the camera's AF picked the wrong thing or because the salient detail isn't the highest contrast object within the box or ...

 

First you remove when AF really doesn't add much:

1) You have landscapes mostly you are out near f/11 or f/16 and want everything in focus anyway. Plus virtually everything is way out beyond the hyperlocal distance.

2) I have flowers and lichen and detail things like that. This is zoomed in and up close where you have minimal DOF and you really care about what you want to get in focus. In these situations, it is a problem telling the camera exactly what you want to have in focus. It ends up being quicker and easier to put the camera on MF and move in and out to get exactly what you want.

3) The decisive moment. There is no AF system that is faster than having your lens pre-focused and having a wide enough zone that you get your subject into that zone. Maybe other people are much better than I am with AF systems but the salience problem really rears its head here. The process of communicating "focus there" (mostly this) and then letting it focus (less this) takes too long. 

4) Many times the focus from the previous shot is almost exactly what the focus from this shot is going to need to be. Going through an AF cycle in these cases just gets in the way and slows things down.

 

Those are cases where AF really isn't helpful and gets in the way. So you throw those out and what is left.

There are many cases where AF is nice to have. I like it; I use it; I actually use it quite a lot. However after thinking about it a lot, I've come to believe that in the remaining cases AF speed really isn't terribly important.

 

I think photographers need to consider how much value add AF really is making to their photography and the situations where they are using it and then seriously answer does raw AF speed really have any impact here. I think more often than not, the challenge really is more about selecting the AF point and the appropriate DOF -- salience. When that becomes the limiting factor, the question becomes would MF be more appropriate here. This is why I believe that the actual value of lightening fast AF is over rated.

[AlanG]

I don't think that I'm making excuses for the T as putting forth the hypothesis that AF isn't really as useful as people think it is.

 

Yes one can focus manually for all kinds of purposes. I bet the majority of people who buy a T or similar camera use AF almost exclusively.  That's why reviewers say what they say. AF performance, for better or for worse, is a key feature for most people getting sharp images.

[earleygallery]

The AF is below par - from what you're saying. 

 

The T is an AF system. It should have class leading AF given that it is a Leica, and its premium price tag.

 

I've not used a T but if the AF is not up to muster then yes, I would say you are making excuses. 

[bencoyote]

The AF is below par - from what you're saying. 

 

The T is an AF system. It should have class leading AF given that it is a Leica, and its premium price tag.

 

I've not used a T but if the AF is not up to muster then yes, I would say you are making excuses. 

 

I'd say that it is much more subtle than that.

 

I started out with a loaner Nikon J1 but I didn't like that. Then I went to a Lumix GX1 and caught the bug and wore that thing out. Then I got the E-M1 based upon specs and size in comparison to a DSLR. The AF on that is remarkably fast. Evidently, the AF on the E-M1 is class leading in the mirrorless segment at the time it was released. It is still lightening fast. However, like most AF systems, you have to tell the camera what to focus on or you have to let it pick. When you let it pick by itself, it is a crap shoot and it seems to frequently not pick the right thing. However, if you reduce it to a smaller area it is more likely to pick the right thing to focus on. But you have to communicate where you want to have it focus. Mostly this means moving the focus point with up down left and right buttons. You can also do touch to focus but then you can't use the EVF then. I found that it was always a game to get the camera to focus on the "right" thing. It was like:

"no that's not it"

"not that"

"no - not the grass behind the flower, the flower"

Single point focusing did help sometimes and so did pinpoint focusing but it is always a game of figuring out how to get the camera to focus on the right thing. 

 

The other thing was I'd see a shot starting to develop and I'd want to get a shot ready. The last thing that I'd focused on was along the right but now the thing that I wanted was on the left and so I'd be frantically clicking left to right to change the position where the camera chose to focus to get it in place in time. Once it was in the right place, the camera locked focus practically instantly.

To help out with this, I programmed one of the buttons to reset to the center. This at least gave me a reference point so that I wouldn't have to have the camera all the way up to my eye before I could select the focus position. While I'm moving it up to my eye I could hit the shutter, then hit that button, and maybe get a couple of clicks in the right direction before it got up to my eye.

 

The other thing I did was focus center, compose, release.

 

I really didn't make use of MF much except when I got very frustrated with the AF and couldn't convince it to lock onto the thing that I wanted. I tried astrophotography with E-M1 and found it practically impossible. I could sometimes lock onto a star but it would frequently want to AF on something. When I used MF there was no scale and there was no standard with lenses as to which way was infinity or any indication that I was there or if I needed two more turns of the focus ring.

 

With the T it has the multipoint focus mode and I never really spent much time with it trying to figure out how to convince it to find the "right" point out of the field. It felt like the same game as with the E-M1.

Then there is the touch AF. This works fine when you are holding the camera. However if you are using the viewfinder the touch screen is not usable to move the focus point. Bright light -- not the best you want the viewfinder.

The two focus modes that really work well are single point and spot. Both of those are focus center, compose, release. When you have time to compose the shot these are totally fine. However, every single shot you have to move the camera around and recompose because every single shot the camera wants to refocus.

All of the things above about the T are the same with the recently best-in-class E-M1 except you have the ability to move the focus point with the arrow keys while using the viewfinder.

 

It was all of that which led me around to start working with MF mode extensively. I started realizing that most of the time was spent communicating to the camera what I wanted in focus. It didn't matter that on the E-M1 it would lock onto a point in .125s while on the T it was .25s Most of the time was me telling the camera, that is the thing that is important to me.

 

I've watched my friend who shoots a Canon 5D M3 and my mother with her Lumix GM5 struggle with the same thing. Focus - nope, focus - nope, focus - nope. Or damn missed the shot - the camera focused on the wrong thing.

 

Unlike the E-M1 the T's MF mode seems intelligently thought out. It gives you the focus scale that indicates DOF and all the lenses focus the same direction. Seeing where the lens was focused and the DOF gave me a much better understanding of the realities of hyperfocal distance and zone focusing. I realized the things that I said above:

1) landscape you don't need to focus everything is at infinity

2) flowers and lichen and details I can see what I want in focus and nail it with MF

3) You really need the DOF scale to do zone focusing and know how much you will get before you compose the shot.

4) For grabbing a decisive moment no AF system can beat being already focused.

5) When you are working a shot most of the time your distance to subject is a matter of fine tuning vs. going through a whole AF cycle.

6) When you zoom with a manually focused lens the focus doesn't shift 

 

So it never really matters if the lens locks focus in .125s vs. .25s. The real time is taken up with communicating to the camera what the "right" thing to focus on is. So in what way is the AF system on the T sub-par?

I think from the perspective of the reviewer it comes down to something like:

1) I mostly use a DSLR and so I prefer the EVF to using the touch AF

2) When looking through the EVF without arrow buttons to move the focus point, I don't know how to make AF lock onto the thing that I want. 

3) I don't like focusing on the center then composing because it interrupts my flow when working a shot.

 

The thing that the T taught me was not to be afraid of MF because it is well implemented and not an afterthought and that in many situations MF is faster and more accurate.

 

I feel like AF speed in the camera is one of those marketing things that camera makers can easily compare between each other and so they do. However, once you get beyond a certain point, other factors become more important when actually using the camera. My argument is that it has more to do with effectively communicating the focus point to the camera.

 

What are other people's thoughts? What problems do other people have with the AF on the T? How does it compare to the AF on other cameras? Is it really that bad and I've just totally figured out how to work around it by using single point and MF almost always?

[dpitt]

Advertisement (gone after registration)

+1 on the MF side here.

I have an X2 which is probably worse in AF speed than the T (never had the chance to compare)

But I think the MF mode on the X2 is a big step in the right direction.

 

I found that I struggle with the same issues as the OP when trying to use any AF camera. While I fiddle with the buttons to get the damned thing to focus on what I want, the decisive moment is long gone!

Also, when using the AF camera it is very easy to lose shots because you did not see the camera switched focus to some subject you did not want… Very often I came home thinking I had perfect shots and when uploaded to the computer I found that focus was not at all on what I thought it was in the viewfinder when taking the shot.

 

Bottom line is that I find I lose less shots when doing manual focus on a camera like the M9 than on a fast system like the Canon 5D. Also, when focus failed because of wrong subject, the shot is often useless. When focus fails because of lack of time to focus, the shot may still be close enough to be acceptable.

 

Maybe I am slow and old fashioned, but I'd rather blame myself for not being accurate or fast enough than having a camera decide for me.

 

So, as long as the T or X or any camera takes pictures that are sharper and better than anything else in this price/weight/size category, I do not think Leica should have spent more money on getting the AF system more perfect.

Indeed, if Leica would give me the option to buy a X without AF for a few hundred euro's less, I would go for that one.

[euston]

I got the T mainly to use with lenses that require an adapter - manual focus only, therefore.

 

But I have one native T lens - the 11-23 zoom. I use it on auto focus exclusively - usually the spot setting - and I don't have problems with focusing, even in dim light. It may not be the fastest AF in all conditions but, for my kind of photography, I don’t need it to be fast as long as it’s accurate, which it is.

[rsh]

I like the T, and yes, it could be better.  if Leica introduces a T2, I am confident it will have built in EVF and faster autofocus.  In the meantime, I will continue with it as it gives me two great zooms that compliment the M and the faster M primes, especially for travel.  The 11-23 and the 55-135 are terrific lenses.

 

Additionally, the T lens mount is larger than the R Mount and offers possibilities for a future full frame auto focus camera with zoom lenses and longer focal length lenses as well as the use of current M & R lenses with adapters.  I have to believe that the T is the first step in this direction.

 

Instead of looking at Leica Products with a glass half empty attitude, I tend to look at them with a glass half full attitude.  In 45 years, I have had my Glass half empty moments, but they have quickly disappeared.

[Peter L]

 AF, not a great shot, however, shot in windy conditions, swaying hummingbird feeder, could not have done this with MF

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[Guest VVJ]

I own a T and I really it.  That being said I also pre-ordered the new Q with built-in EVF and from what I can read good AF.

 

A Leica T with built-in EVF and good AF would have been one heck of a camera but unfortunately...

 

I still love the T though but honestly if the Q delivers the T's future is uncertain...

 

What concerns me most is that Leica seems to have gotten it right with the Q but does not seem to have put the same effort in the T...

 

Perhaps there will be a new T body, perhaps not, only time will tell...

[pico]

Instead of looking at Leica Products with a glass half empty attitude, I tend to look at them with a glass half full attitude.

Whether the glass is half-full or half-empty depends upon whether you are the customer or the manufacturer.

[earleygallery]

 

 

What concerns me most is that Leica seems to have gotten it right with the Q but does not seem to have put the same effort in the T...

 

Perhaps there will be a new T body, perhaps not, only time will tell...

 

The T was designed by Audi. Leica designed the Q in house. 

 

Why didn't they use Audi to design the Q? I can only think the T was a 'project' - if it was a roaring success it would have been a blueprint for future models. 

 

So, why is the Q such a conventional design?

[jaapv]

 AF, not a great shot, however, shot in windy conditions, swaying hummingbird feeder, could not have done this with MFL1010724.jpg

Others can...:

 

http://www.wildlightphoto.com/birds/trochilidae/calypte/anhumm06.html

[pico]

So, why is the Q such a conventional design?

 

Is it conventional? To me it satisfies the relationship to Leica's traditional camera liberated of its rangefinder, retaining control placement, and rounding out some corners. That's good design.

[LexS]

I was very interested to read the article on the Q. After thinking it through, however, I realised that this was because I am a technician: a superspeed AF and so on. Is that really important, this thread in general states that the picture is made by the photographer and not by the AF speed !

After having used SLRs all my life, from time to time trying out a Leica Rangefinder: I could not get used to seeing much more through the rangefinder in comparison to the through-the-lens viewer of the SLRs.

Now i own a T, and I am very satisfied with the system. The AF is speedy more than enough, the handling is very simple (the 1.310 firmware update is perfect).

Now I am at the point that I decided not even trying out the Q when it is available. The IQ with the T is soooo good, and soooo Leica. I am happy (and my wallet too).

The photos of the hummingbird in this thread are wonderful, I hope to be able one day to make such good pictures also, the T can, now it is on me.

LexS

[tom0511]

IMO AF of the T is quite good, not class leading but really fine...except with the telezoom.

I am also someone who likes the T system quite a bit.

[earleygallery]

Is it conventional? To me it satisfies the relationship to Leica's traditional camera liberated of its rangefinder, retaining control placement, and rounding out some corners. That's good design.

 

It's about as conventional as it gets - shutter dial, aperture ring, traditional styling. Why wasn't the Q designed like the T (i.e. mostly touch screen controls, aluminium hand polished unibody etc). 

 

I can only assume that the T project hasn't been a roaring success. 

[dkCambridgeshire]

It's about as conventional as it gets - shutter dial, aperture ring, traditional styling. Why wasn't the Q designed like the T (i.e. mostly touch screen controls, aluminium hand polished unibody etc). 

 

I can only assume that the T project hasn't been a roaring success. 

 

Oh! … You've got the sales figures ?

 

Please publish.

 

And … Please be aware that the Q would probably have been 'on the drawing board' before the T was launched.

 

dunk

[Advertisement]

Hallo,

Look here: Leica L-Mount

[bencoyote]

And … Please be aware that the Q would probably have been 'on the drawing board' before the T was launched.

Yes that is a very important point. And the next generation of the T is probably fairly far along in its design at this point as well. 

 

I think that more than being a harbinger of the end of the T & X lines, the design of the Q gives us some very good clues about what the next M is going to be like.

 

It seems to me that Leica can either diversify or refine the T. I doubt that they will drop it this early. Diversity would continue the T mount and offer variations on the T for different user interface preferences. One more like the X or Q would be an obvious direction to go. Refinement would be something like taking the current concept and giving it a bump in specs to increase processor and sensor performance. Doing both simultaneously wouldn't really be that hard. A third option which I've lobbied for would be also offer a "CLAU" service where they do the normal CLA and also upgrade the logic board and sensor. That would be consistent with the T as the product where they innovate and try new ideas. 

 

BTW does anybody know what happened to N, O, & P?

It seems like we have M,_,_,_,Q,R,S,T_,_,_.X? (X is kind of special and so I'm less surprised by the missing U,V, and W)