[Xaqtly]

I have a question about the autofocus on my T2i. I've been noticing that when using autofocus and there are a lot of things in the frame to focus on it almost always ignores the center dot and focuses instead on things off to the side. I'm making sure to get a nice solid object to focus the center dot on, and it almost never makes a difference. it always wants to focus on things near the edge of the frame instead. I can't figure it out.

Anybody experienced this, and am I just missing something really obvious about how autofocus works?

[Brave Ulysses]

Originally Posted by Xaqtly I have a question about the autofocus on my T2i. I've been noticing that when using autofocus and there are a lot of things in the frame to focus on it almost always ignores the center dot and focuses instead on things off to the side. I'm making sure to get a nice solid object to focus the center dot on, and it almost never makes a difference. it always wants to focus on things near the edge of the frame instead. I can't figure it out. Anybody experienced this, and am I just missing something really obvious about how autofocus works?

don't have one, only have a canon point and shoot, but isn't there multiple AF settings that you can cycle through, I know my camera has a center point AF setting.

[PB PM]

You need to set the T2i to single point AF, rather than auto area. IIRC you need to press the * and turn the top command dial till you get to single point. It should tell you on the rear LCD if you have changed the setting.

[Xaqtly]

Originally Posted by PB PM You need to set the T2i to single point AF, rather than auto area.

Ah, ok. I'll try that. See, I knew it was going to be something stupid I was missing. That said though, is that normal behavior for auto area AF? Without knowing any better I just sort of assumed that when I target something in the center dot and focus, it should put a higher priority on the center than on the other AF points. It doesn't seem to work like that though. I believe for the way I shoot though, setting it to single point will definitely work better for me.

[PB PM]

I never trust auto area on any camera, the logic behind those systems are, illogical at best. I think they work best when people are in the frame. Also, most auto area systems look for the closest subject, rather than at any given area.

[Matsu]

Maybe it's just an AF thing. I don't seem to use AF tracking as much anymore. I pick my single point in the frame and shoot the thing I want, and this is mostly people and events. I get better framing when I'm not seduced by follow the tracking syndrome...

[PB PM]

It depends on how fast the subject is moving, but yeah AF tracking doesn't work all that well either. I use single point with AF-C for moving subjects, but prefocusing and letting the subject come into focus works better. It worked when people were shooting with manual focus cameras, and it still works today.

[Dorian Gray]

You've probably missed it, but Nikon had — for a brief few hours — a new AF-S Nikkor 50 mm f/1.8G lens on their site. The ever-vigilant Ken Rockwell got a page up before Nikon pulled their own page.

Thoughts? Seems like it might be up Matsu's street. The bokeh samples on the Nikon site looked really, really good for a fast fifty. We're talking Sigma 50 mm f/1.4 standards. Very interesting. Wonder if it'll be under $300?

[Matsu]

That's excellent news. My 85 f/1.8 is excellent, razor sharp. I'm keeping just for FX... eventually, sigh... It's illustrative of the general case for some smaller (still fast) but slightly slower primes. It might not have that last bit of subject isolation and nth degree wide open sharpness - but it's so close that it's only a couple of minor corrections away from matching the look you want, while being a lot less conspicuous.

It will be interesting to see if that 50 was a DX or FX lense. I don't really think a DX version is neccessary, this lense should be small and cheap even in FX guise. Anything under $300 = good value. Under $200 = must buy.

Edit, clerly states it's an FX. Which is the better move all around. Now we need a new AF-S 35mm f/2, followed by a 28 and 24, these I suspect will get progressively more expensive, but should all fall well below the $2K level of the 1.4 glass, while being quite a being easier to hide...

[Dorian Gray]

I'm guessing closer to $300 than $200, since the 35 mm f/1.8 is $200 but lacks the focus scale window and is DX only. But with its aspherical element (even if hybrid) and seemingly impressive large-aperture performance based on samples and the MTF chart, it would not shock me if this new lens is priced at $349. Of course I hope it's cheaper.

I'm almost tempted! I do like a good fifty.

[PB PM]

I suspect it will be somewhere between $200-300 US. They cannot put it too high because the 50mm F1.4D is still selling new for around $350. For a DX user the 50mm F1.8G might be more appealing, but a FX user might stop and think, F1.8 or F1.4?

[Eugene]

Hope it doesn't have the center hotspot issue that plagues my 50mm f/1.8D. Not sure if the f/1.4s have them.

[Matsu]

Depends on which DX camera you have. For the AF-S only models, the 1.8G beats either of the D lenses, but you need f/1.4 if the very last bit of light gathering is important, and that leaves only the 1.4G lens, which is a lot more expensive. It looks from the MTF charts that the 1.8 is sharper wide open than all other current Nikon 50s. If that's so, it might be the better buy regardless of FX or DX use. The 2/3rds stop light loss means almost nothing on modern FX, and even for DX, it's still 1.3 stops faster than any zoom, and sharper too.

Any linkies to the samples?

[Matsu]

I wonder about the price. $349 might be about right. Now, If we see more AFS f/2 primes to replace the older AF-D stuff, there's a hint about those prices creeping up given newer optical formulas and technology (coatings, AFS etc...). I'd love to see an all f/2 line-up at half the price of the f/1.4s: half the speed, half the cost. Just keep them sharp wide open. And, make the 35 a little cheaper than the others just for me

[PB PM]

I doubt it will be $350, because as I noted that is the price of the 50mm F1.4D, which is still in production. If the price was the same, I'd take the 50mm F1.4D hands down. One does have to wonder how much longer it will be in the lineup though.

As for 1.4 only be 1.3 stops faster, that isn't the only factor in deciding between the two. F1.4 lenses have better bokeh, than F1.8 lenses due to the use of a 9 bladed aperture design. If the 50mm F1.8G has the same nasty bokeh wide open as the 35mm f1.8G and current 50mm F1.8D, it would be a little disappointing. I wouldn't be surprised if it does, since it uses the same 7 bladed aperture design of the 35mm F1.8G.

As for the samples, Nikon pulled the lens off it's website 4 hours after it appeared, so no links.

[Dorian Gray]

Olivier Giroux's thoughts, as perceptive as always, are worth reading here.

PB PM: you might take the 50 mm f/1.4D "hands down", but not everyone will. The AF-S f/1.8G will have much better wide-open sharpness and contrast, much better bokeh (bokeh is mostly affected by optical design; the shape of the aperture isn't the whole story), and a built-in focus motor. It will also have a price that makes sense for a new lens in 2011, rather than an old lens for which the development costs have long-since been accounted.

Consider the Sigma 50 mm f/1.4. It uses an aspherical element to better balance coma and spherical aberration, enabling superb bokeh — much better than any Nikon 50 mm lens. Properly designed, this new lens will offer similar performance to the Sigma, though at a lower price (it's slower and made of plastic).

The Sigma uses a glass-moulded aspherical element. I fear the Nikon will instead use a hybrid aspherical element, using a replica plastic layer over a spherical glass element to create the deviation from spherical. Hybrid aspherical elements distort with temperature changes, which limits the designer's freedom and ultimately aberration correction. That said, Nikon has great manufacturing expertise in the niche of large-aperture precision glass-moulded aspherical lenses, so there's an outside chance they used a moulded aspherical element in this new fifty.

Matsu: I couldn't extract Nikon's sample images from my browser cache, though I tried. They were too small to assess high-frequency detail, but they revealed a far better bokeh than the existing autofocus 50 mm Nikkors.

[PB PM]

Sure, the 50mm F1.8G might be sharper wide open, but how both lenses perform at F1.8, that will be the determining factor wont it? I do see your point though, and having an SWM focusing motor might be nice as well. The AF-D primes have been known to focus faster than the recent AF-S primes so that may also be a factor in some peoples choice as well.

[Matsu]

I didn't see the samples. Others noted that they looked good. I don't think that good/bad bokeh is as simple as more or less blur, or shape, or the shape/number of aperture blades. Though, these are all contributors, I think people have a few general misconceptions about what makes bokeh good or bad. It's the way the lens 'draws' or 'paints' that matters more. This is your DaVinci art lesson at work here, the way the conical propogations of light (in three dimensions) end up hitting the sensor (in two dimensions) is really the dominating factor. It will give you nice diffuse spots, or discs, or rings, or ripples. The first being the best, the last two often being the worst. If you have the first condition, it hardly matters at all what number or shape of aperture blades, since the points are dense in the centre and soft at the edges, the further you stray from that, the more visible the lense structure becomes in the image.

I think it might be better to talk about bokeh being messy or smooth. Bigger apertures have more blur, so they're more likely to seem smooth, but do they stay that way when closing down?

Look at the 4/3rds wide standard zooms, they have some very messy looking bokeh that gets especially nervous around irregular fine detail, like foliage. A lot of zooms aren't as good as primes I gather from numerous reviews, but some are good even at f/2.8 and f/4. That is they're smooth, if not as blurred, if any of this makes any sense? I'm probably not expressing it well enough...

[PB PM]

It makes perfect sense. My 300mm F4 has great bokeh, while the 50mm F1.8 is bad.

[Matsu]

I love this size comparison. This sort of stuff gets my vote for a dirt simple, small, FX body.

D5100 vs D700 size comparison from Nikonrumors

A couple of things: that 35mm f/1.4 is undeniably sexy, and it makes great pics, but it perhaps has a shade too much presence for a street rig.

The front faces of those two cameras tell the whole story though. The D5100 is small enough that it could disappear in my hands. The D700, like my D300, not so much. I'm not suggesting they can make an FX down to D5100 size, but something close to D90/D7000 size and some f/2 primes could work nicely...

[PB PM]

No doubt that the 35mm f1.4 would be a nice lens to shoot with, and having a D7000 sized FX body with a video mode to use as a backup to the D700 would be great.

I rented a 70-200mm F2.8G VR1 for a week, just returned it today. I think I love that focal length on FX, for event shoots! I rented the lens to shoot my sister's wedding, as a backup shooter, and I used it for 85% of my shots. The AF-S 60mm f2.8 and AF 20-35mm f2.8D also came in handy, with the latter working best for group shots.

I love having lenses that I can shoot wide open and come home with tack sharp images. That's a good thing, because the lighting in the lodge was terrible at best. I wanted to shoot at around 1/125-1/160s, so I set the D700 to auto ISO (200-3200) and shot without thinking about the camera most of the time. Didn't have to use my SB-800 till after dark with the D700.

[Matsu]

Nikon could make a huge splash with a D9000 type camera (D7000 size + FX).

For the D700, how do you like pushing it to ISO 6400? The D300 isn't at it's best past ISO 1600, but shooting RAW, I now routinely take it to ISO 3200, and just do my best to have the subject properly exposed and sharp.

I've decided that my next camera has to give a convincing performance at twilight light levels. So ISO 6400-12800, even just to keep the shutter tripping a stop faster when everything's moving fast.

[PB PM]

ISO6400 is still better than ISO3200 on the D300, but I don't like to use noise reduction (in camera or in post) unless I have to. The character of the noise from the D700 is different from that of the D300, you cannot really compare the two. ISO12800 on the D700 would be okay if you used third party noise reduction software, rather than what comes in programs like Aperture or Lightroom. I'd be okay using ISO12800 on the D700 before I pushed the D300 above ISO3200, let's put it that way.

ISO3200 sample from the wedding.

[Matsu]

No question it's cleaner, by 2 stops, maybe a bit more if you account for the way its colors stay true at higher ISOs.

[Matsu]

While we've been doing much jabbering on about Nikon and full frame, I've recently read a couple of very interesting reviews. One is about Sigma's HSM OS 17-50 f/2.8. Like a couple of recent Sigma prime lens designs, its frame centre resolution edges out the best from Nikon and Canon. Reviews put it in prime lens territory even at maximum aperture. It's centre reaches impressive lp/mm results at all focal lengths and apertures starting from f/2.8 to f/8.

Between this, the 30mm and 50mm f/1.4, Sigma is on it's way to producing a very high quality lens line-up. I want to see these things in front of an SD1, they might help the X3 sensor rewrite the rules for APSC resolution...

[PB PM]

Sigma has some impressive figures, problem with Sigma is quality control. If you get a good copy you will get some great glass, if you don't you could be testing four or five before you get one. I've been looking at 85mm f1.4 lenses and it is a tough choice between the Sigma 85mm f1.4 HSM and the Nikon 85mm f1.4D. Some are saying the Nikkor has better bokeh while others say the Sigma. I think I'd want to test them both before making a choice.

[Matsu]

To the discussion earlier in the thread about mirror-less and DSLR etc... Olympus seems to be now suggesting that it will continue to produce E-x and E-xx 4/3rds DSLRs while micro43 will replace their low end E-xxx.

Lots of interesting speculation on that... hmmm... Could be that they just can't get EVF based focus to work as fast as they need want. Or, the advantages of the mirrorless system being mainly reduced mechanical complexity and lower cost, less so the actual size - especially when fitted with consumer pleasing but slow large ratio zooms. Other developments in the marketplace? Is Panansonic eating all the m4/3 pie, so we have a slight retreat back to 4/3rds? Olympus seems a bit confused these days...

[PB PM]

Really, Olympus had said that the E-5 was going to be the last DSLR they were going to make. Do you have a link?

[Matsu]

They did but...
Supposedly from an interview with an Oly exec.

Olympus has been missing a bit too much for their own good lately. They mis-judged sensor size early. The industry settled on APSC, and now they're working with a sensor that's just a bit over half the size, about 60% or so, I haven't done the math. Equivalent lenses, where they exist, aren't appreciably smaller - ie, f/2 zooms vs f/2.8 zooms.

The products have their own unique qualities, but some of the design considerations, it's obvious now, were too conservative, especially the telecentricity requirements. Yes, perhaps a digital sensor of 1/4 the size is capable of equaling 4X larger film. They're probably right, the best 12MP and 16MP 4/3rds at mid-high ISO meet and exceed 35mm color film above ISO 400. There was a Kodak paper on this somewhere way back...

But it doesn't matter, the world has moved on. APSC comfortably equals 35mm film, and 35mm digital exceeds it in almost every way, and it's what enthusiasts and professionals want, not to mention spec whores...

If they do keep the 4/3rds mount alive, they ought to consider a larger sensor using the same mount. Basically just like Nikon does with DX and FX: make them inter-operable, let the bodies crop down automatically when a 4/3rds lens is mounted to a 35mm body. Nikon and Canon would have 35mm and 1.5/1.6 crops and Olympus would use 35mm and 2x crops, something no less sensible, and with a clear upgrade path to larger cameras.

[PB PM]

Considering that rising stars in the camera industry, like Panasonic, are using 4/3s, I doubt sensor size is the issue.

I think Olympus is in a tough spot, which is nothing new for them. You can look at their market share and think they are in trouble, but they've held that same market share for 10-15 years now. Going to a bigger sensor would force them to start from scratch in terms of lenses, considering that Olympus hasn't made any lenses for larger sensors in about 10 years since the OM line was dropped.