Olympus new killer feature for sports...

This was discussed elsewhere in the thread. It's not just ground speed that matters. A child running toward you at close distance is harder to track than a plane.

There's a good explanation here which is fairly technical but short:
http://forums.dpreview.com/forums/read.asp?forum=1021&message=34713361

Julie
My colleges at the European Championships in Beach Volleyball last summer, told me they could get sharp images shooting players moving towards them. For me, with E-5, this has not been possible.

--
Raymond
http://www.blurb.com/bookstore/detail/1244032#author-bookshelf (my books)
http://osloray.wordpress.com/ (Image blog)
http://www.bildebank.com
For me it was. My E-5 was surprisingly good in C-AF, even when using slower lenses like the 35-100mm f2. I shot fast planes coming towards me, and got very high keeper rate. I also shot rally, motocross, and the same, very good C-AF. Amazing actually, because by reding these forums it seemed the Oly cannot do C-AF.

The key for me was to use Diamond AF, AF-Lock ON, Release Priority OFF, and let the camera ,,catch,, the focus before releasing the bursts.
 
--Yes but with 4/3 then lens with same speed will be 1/2 to 1/3 the size and much lighter to cart around! With the high good high ISO performance of these cameras and even the more recent Pany ones the very fast lens speed is much less necessary than it used to be. I can see future super teles for m4/3rds being much smaller and potentially less expensive with equal IQ. My hopes at least.
Do you really expect a dramatic size differential between a m4/3 400mm f2.8 lens and its traditional AP-C/FF counterpart? Or even a 300mm f4 m4/3 prime vs the current traditional iterations?

The size of those super tele-primes is more relevant to light gathering properties than the size of the camera's sensor.

I don't see anything longer than perhaps a 300mm prime for the m4/3 genre and I wouldn't even hold my breath for that one ;)

--
Holmes
http://holmes.zenfolio.com/
 
if they keep the same apertures as full Frame (it will have the 2 stops penalty for a smaller sensor but they have been taking 1/1000 iso 6400 f2.8 sports pictures so you dont need iso 51200 or something to get good sports pictures (people walking needs iso 6400 1/125 and f2.8 and people standing still only 1/60 3200 f2.8)

So iso 3200 looks very good in the omuser samples... iso 6400 looks like 1600 on current cameras (since most sport shooters shoot jpgs with more chroma noise at 6400 than what olympus om-D jpgs have I think it will be fine)

4.2 frames per second might not be enough... I dont know... there is Panasonic 35-100 f2.8 announced so Indoor sports may be possible in 2012 with m43..

So for outdoor natural and artificial you probably need a 150mm f2.8 (300 f2.8 equiv.) and probably a 200 f2.8 (i believe it will be a 200mm f4 for size constraints). Those would have a tremendous size advantage compared with aps-c or m43 and could be a second camera for FF shooters
 
--
john carson
 
I hadn't noticed at first, but the E-M5 has this 3D tracking feature, coupled with a doubling of the sampling rate.

I think that the EVF should follow with little or no black out.

That might also be the reason for even faster AF.

It would mean that m4/3 can do fast sports relatively easy. Did any of you notice?
No, you're the only one who noticed ;) Of course we did--that's huge! Well, potentially huge anyway. Continuous AF is the one major weakness of m4/3 and all mirrorless systems bar the Nikon 1. If Olympus's solution makes continuous AF usable that's a giant step forward for me. I doubt it would turn CDAF into a sports photography favorite, but going from unusable to usable is big.

Olympus are talking up the AF, saying that it's still the fastest in the world (single AF of course), but Mr. Terada from Olympus has also mentioned that they have improved continuous AF. I'm cautiously hopeful.
 
A simpler explanation is that it has to do with apparent motion in the frame. The size, speed, and direction of movement by the subject doesn't matter, but the change in what's seen by the camera does.

The camera doesn't care if the "plane" is a model of an SR-71, or the real plane in flight. All it cares about is how long it takes for the image of the plane to move from one side of the sensor to the other, one pixel to the next. Assuming both the model and plane have the same apparent size in the frame, I can move a model 10' from the camera faster than the real plane can fly across the camera's FOV while at 10,000' AGL. The camera doesn't care which one is actually moving faster, just which one is moving through it's field of view faster. That's also why a panning shot requires less of the camera's AF system, since the apparent motion of the subject is essentially nil.

Maybe I'm wrong about all of that, but it makes sense to me when thought of in that way.
This was discussed elsewhere in the thread. It's not just ground speed that matters. A child running toward you at close distance is harder to track than a plane.

There's a good explanation here which is fairly technical but short:
http://forums.dpreview.com/forums/read.asp?forum=1021&message=34713361

Julie
My colleges at the European Championships in Beach Volleyball last summer, told me they could get sharp images shooting players moving towards them. For me, with E-5, this has not been possible.

--
Raymond
http://www.blurb.com/bookstore/detail/1244032#author-bookshelf (my books)
http://osloray.wordpress.com/ (Image blog)
http://www.bildebank.com
For me it was. My E-5 was surprisingly good in C-AF, even when using slower lenses like the 35-100mm f2. I shot fast planes coming towards me, and got very high keeper rate. I also shot rally, motocross, and the same, very good C-AF. Amazing actually, because by reding these forums it seemed the Oly cannot do C-AF.

The key for me was to use Diamond AF, AF-Lock ON, Release Priority OFF, and let the camera ,,catch,, the focus before releasing the bursts.
--
http://453c.smugmug.com/
 
While m43 is still behind DSLRs today, CDAF speed tracks Moore's law, meaning that advances in processing power allows for continued improvement of CDAF. I think we are one more "doubling" away from prime time performance.

Global shutter is still a big hurdle. When that arrives then we could see mirrorless advance beyond DSLRs.
--
SLOtographer

"If we limit our vision to the real world, we will forever be fighting on the minus side of things, working only too make our photographs equal to what we see out there, but no better." -- Galen Rowell
 
according to robin wong who tried it: ( http://robinwong.blogspot.conz )

"If you have tried shooting with high frames per second burst continuous modes on higher level DSLR models such as Canon 7D (8FPS), you will know the limitation on the optical viewfinder, because the mirror flips up and down so fast that it blocks the view through the optical viewfinder, and you cannot see what you are shooting while you are bursting the 8 frames per second all the way. However, the E-M5 does NOT have a mirror, hence you can fully see what’s happening while you are shooting at high speed, even at the 9 frames per second continuous drive mode. Surely, this will be advantageous in many situations where you need to be fully aware of what you are shooting and not doing guess works based on the composition of the first frame. Yes, I have tried the full 9 frames per second shooting on the pre-production E-M5 unit, and I can fully see what I am shooting at all times, while the shutter was machine-gunning away. "

--
Shoot the Light fantastic
 
While m43 is still behind DSLRs today, CDAF speed tracks Moore's law, meaning that advances in processing power allows for continued improvement of CDAF. I think we are one more "doubling" away from prime time performance.
I have always worked under the assumption that CAF + tracking was a virtual feature, too slow, and rigid.

However it seems that Oly set up to crack the problem, not only by doubling the sampling rate and processing, which has positive effects on the EVF lag too, but by using some predictive algorithm when the subject moves in the depth direction.

In Sony's a55, I read in the LL, this worked quite well in a rodeo.

Since the Oly processor is dual core, filling the buffer is independent from other calculations, therefore the hope that it might be fast enough this time.

Faster processing might also benefit face and eye recognition too. Shape recognition is at work here, in tracking it might be something else, like colour or contrast variation.

Am.

--
Photostream: http://www.flickr.com/photos/amalric
 
Dreaming is fine, but Oly has been slippery with their words before.

When the E3 came out, they advertised the 'world's fastest AF' and the lowest light sensitivity (-2ev) and one of the most accurate AF modules on the market. That combination would have have created an industry leader.

It turned out two were true but didn't work together. For example, the E3 had one of the slowest focus times in low light of any camera in that class.

As to the accuracy, to be kind, the E3's AF was 'quirky'.

In bright contrasty light, it worked fine.

As to the glowing words given to the tracking AF of the OMD, I'm suggesting we wait until we see how Olympus's words translate into reality before we make declarations.

Let's hope they are not being slippery again.
 
However it seems that Oly set up to crack the problem, not only by doubling the sampling rate and processing, which has positive effects on the EVF lag too, but by using some predictive algorithm when the subject moves in the depth direction.

In Sony's a55, I read in the LL, this worked quite well in a rodeo.
Sony's A55 uses phase detect, not contrast detect.

--
john carson
 
However it seems that Oly set up to crack the problem, not only by doubling the sampling rate and processing, which has positive effects on the EVF lag too, but by using some predictive algorithm when the subject moves in the depth direction.

In Sony's a55, I read in the LL, this worked quite well in a rodeo.
Sony's A55 uses phase detect, not contrast detect.

--
john carson
Oh, yes, how could I?

However thanks to AI and FAST AF and processing, Oly might have achieved in CDAF what was before the preserve of PDAF.

Funny that we should be taught the error in our ways in a preventive manner by Fuji adepts and 4/3 old geezers.

Are they so 'be afraid, be very afraid'? As for operational speed, I wouldn't accept a candy or a crumb from Fuji.

LOL

Am.

--
Photostream: http://www.flickr.com/photos/amalric
 
I'm not sure what he means when he says "you cannot see what you are shooting." I don't know about the 7D, but the D2H does 8-9 fps, and the blackout time is so miniscule that I can't say I've ever found it a problem, even shooting very fast and unpredictable action.

Not to say that the Oly feature isn't a great acheivement. It's exciting to have 9 fps in an Oly camera, though it's unfortunate if it doesn't work in C-AF. Then again, depending on how well the C-AF works or doesn't work, that could be a moot point.

Julie
according to robin wong who tried it: ( http://robinwong.blogspot.conz )

"If you have tried shooting with high frames per second burst continuous modes on higher level DSLR models such as Canon 7D (8FPS), you will know the limitation on the optical viewfinder, because the mirror flips up and down so fast that it blocks the view through the optical viewfinder, and you cannot see what you are shooting while you are bursting the 8 frames per second all the way. However, the E-M5 does NOT have a mirror, hence you can fully see what’s happening while you are shooting at high speed, even at the 9 frames per second continuous drive mode. Surely, this will be advantageous in many situations where you need to be fully aware of what you are shooting and not doing guess works based on the composition of the first frame. Yes, I have tried the full 9 frames per second shooting on the pre-production E-M5 unit, and I can fully see what I am shooting at all times, while the shutter was machine-gunning away. "
 
I'm not sure what he means when he says "you cannot see what you are shooting." I don't know about the 7D, but the D2H does 8-9 fps, and the blackout time is so miniscule that I can't say I've ever found it a problem, even shooting very fast and unpredictable action.
agree entirely...........ive never really noticed any mirror blackout at 8 fps on my old d2h either.

I'm looking forward to some future real-world testing of this camera..........looks promising!!
Not to say that the Oly feature isn't a great acheivement. It's exciting to have 9 fps in an Oly camera, though it's unfortunate if it doesn't work in C-AF. Then again, depending on how well the C-AF works or doesn't work, that could be a moot point.

Julie
according to robin wong who tried it: ( http://robinwong.blogspot.conz )

"If you have tried shooting with high frames per second burst continuous modes on higher level DSLR models such as Canon 7D (8FPS), you will know the limitation on the optical viewfinder, because the mirror flips up and down so fast that it blocks the view through the optical viewfinder, and you cannot see what you are shooting while you are bursting the 8 frames per second all the way. However, the E-M5 does NOT have a mirror, hence you can fully see what’s happening while you are shooting at high speed, even at the 9 frames per second continuous drive mode. Surely, this will be advantageous in many situations where you need to be fully aware of what you are shooting and not doing guess works based on the composition of the first frame. Yes, I have tried the full 9 frames per second shooting on the pre-production E-M5 unit, and I can fully see what I am shooting at all times, while the shutter was machine-gunning away. "
 
If at this summers Olympic games we see lots of Olympus OM-D's instead of Canon 1Dx's, Nikon D4's and D800's, then I guess thats all the more proof anyone would need that its a viable sports photographers tool, good enough for the biggest stage in the world.

--
http://www.millsartphotography.com
 
I'm not sure what he means when he says "you cannot see what you are shooting." I don't know about the 7D, but the D2H does 8-9 fps, and the blackout time is so miniscule that I can't say I've ever found it a problem, even shooting very fast and unpredictable action.

Not to say that the Oly feature isn't a great acheivement. It's exciting to have 9 fps in an Oly camera, though it's unfortunate if it doesn't work in C-AF. Then again, depending on how well the C-AF works or doesn't work, that could be a moot point.

Julie
according to robin wong who tried it: ( http://robinwong.blogspot.conz )

"If you have tried shooting with high frames per second burst continuous modes on higher level DSLR models such as Canon 7D (8FPS), you will know the limitation on the optical viewfinder, because the mirror flips up and down so fast that it blocks the view through the optical viewfinder, and you cannot see what you are shooting while you are bursting the 8 frames per second all the way. However, the E-M5 does NOT have a mirror, hence you can fully see what’s happening while you are shooting at high speed, even at the 9 frames per second continuous drive mode. Surely, this will be advantageous in many situations where you need to be fully aware of what you are shooting and not doing guess works based on the composition of the first frame. Yes, I have tried the full 9 frames per second shooting on the pre-production E-M5 unit, and I can fully see what I am shooting at all times, while the shutter was machine-gunning away. "
i also think Mr Wong is talking out of his @rse on the subject of Dslr blackouts, but then again i'm not promoting Olympus.
--
working as intended
 
i also think Mr Wong is talking out of his @rse on the subject of Dslr blackouts, but then again i'm not promoting Olympus.
Well, he's hardly the first person to mention it. I guess it's possible this was a bigger problem on older DSLRs.
 
We'll know soon enough.

But I don't expect miracle in sports. 9Fps only works in single AF. When you use continuous it is 4+fps.
I'm not understanding why is single AF faster. Does it not re-focus in between shots? If it does, what is the advantage of CAF if SAF is faster?
 

Keyboard shortcuts

Back
Top