In camera image stabilisation

[Provokot]

On another site, I saw a discussion about the possibility of Canon putting IS into their bodies.

Most people were quick to dismiss this with the argument that Canon has IS in their lenses - where they make their real money. But is that sustainable?  Surely their competitors who don't have so much invested in IS lenses will have a powerful tool to prise customers away from Canon...

I see the race for pixel count as pretty much over. The race for lowest noise at High ISO seems to be nearing the finishing tape. The race for fast high-res. bursts has a few more laps to do but where will manufacturers go next?

Thoughts?

Regards, Paul

[mahleu]

I forsee the 1.6x crop canon's having IS in body unless there is a drastic reduction in IS lens price, if they don't they'll be losing business from people moving up from compacts who are used to it.

[Jonathan Wienke]

In-camera IS has the disadvantage of being less effective as focal length increases, which of course is where it is the most useful. With a 600mm lens, the sensor would have to move around in an area well outside the image circle of the lens to provide effective stabilization. On an interchangeable-lens camera, that simply isn't practical, as the lens mount would block the sensor's view of the lens. In contrast, putting the IS in the lens works just fine, as moving a fairly small, low-mass lens element accomplishes the same effect in a far more elegant manner. I can see value in a hybrid system where the body has IS used for short-to-medium focal lengths with non-IS lenses,  but for longer focal lengths only lens-based IS would be used.

[DarkPenguin]

I forsee the 1.6x crop canon's having IS in body unless there is a drastic reduction in IS lens price, if they don't they'll be losing business from people moving up from compacts who are used to it.
[a href=\"index.php?act=findpost&pid=162263\"][{POST_SNAPBACK}][/a]

Both canon and nikon have 18-55 and 55-200 lenses with IS.  So the point and shooters will be happy.

The people that can afford the really good glass would prefer to get the really good IS with it.  So they don't have a problem.

It is the people in the middle that get squeezed.  Those of us who would like IS with our Tamron 17-50 f2.8 but who do not want to shell out for a 17-55 f2.8 IS.

[Provokot]

In-camera IS has the disadvantage of being less effective as focal length increases, which of course is where it is the most useful. With a 600mm lens, the sensor would have to move around in an area well outside the image circle of the lens to provide effective stabilization. On an interchangeable-lens camera, that simply isn't practical, as the lens mount would block the sensor's view of the lens. In contrast, putting the IS in the lens works just fine, as moving a fairly small, low-mass lens element accomplishes the same effect in a far more elegant manner. I can see value in a hybrid system where the body has IS used for short-to-medium focal lengths with non-IS lenses,  but for longer focal lengths only lens-based IS would be used.
[a href=\"index.php?act=findpost&pid=162313\"][{POST_SNAPBACK}][/a]

Hmmm.. I am fairly sure that both Nikon and Canon and several other brands are holding/developing in-camera IS technology that eclipses that of their lenses... their competitors could very easily steal a march on them.  If Sony can do it...

...or maybe they'll just have to lower the price of their IS lenses.  I could live with that!

[DarkPenguin]

Hmmm.. I am fairly sure that both Nikon and Canon and several other brands are holding/developing in-camera IS technology that eclipses that of their lenses... their competitors could very easily steal a march on them.  If Sony can do it...

What do you base that on?

[Jonathan Wienke]

Hmmm.. I am fairly sure that both Nikon and Canon and several other brands are holding/developing in-camera IS technology that eclipses that of their lenses... their competitors could very easily steal a march on them.  If Sony can do it...

And does Sony have any long telephoto lenses available (400mm or longer), and if so, how effective is in-camera IS compared to lens-based IS of Canon or Nikon?

[mahleu]

I can see value in a hybrid system where the body has IS used for short-to-medium focal lengths with non-IS lenses,  but for longer focal lengths only lens-based IS would be used.
[a href=\"index.php?act=findpost&pid=162313\"][{POST_SNAPBACK}][/a]

I think that's what we'll end up with. As long as it can be turned off when neccessary i'll be happy.

[aaykay]

And does Sony have any long telephoto lenses available (400mm or longer), and if so, how effective is in-camera IS compared to lens-based IS of Canon or Nikon?
[{POST_SNAPBACK}][/a]

I personally feel an in-lens IS would be more effective than an in-camera IS (with sensor shift), primarily because the IS response can be tailored to the FL.  I also think that Sony will be introducing in-lens IS for their longer lenses, including their 70-200 f/2.8, which is being replaced with a new one....and have the in-camera IS automatically switch off, when the camera senses a stabilized lens as having been mounted.

I think Sony has quite a few longer lenses coming up (including a 600mm), as per their mockup during PMA 2007 (picture below).  The mockup also includes 2 camera bodies, with the one on the left (smaller one) already being sold as the A700 (as of October, 2007) and the one on the right, with the large mirror box and no built-in flash, is supposed to be the Full-frame pro-oriented model (A900?) expected during PMA 2008.

One of the lenses shown during the mockup has already broken cover as the Carl Zeiss Vario-Sonnar 24-70 f/2.8  SSM and there is also a rumored Carl Zeiss Vario-Sonnar 16-35 f/2.8 SSM or a Carl Zeiss Vario-Sonnar 14-24 f/2.8 SSM.  In addition to that, Sony is pulling their 70-200 f/2.8G from the market and is replacing it with a rumored Carl Zeiss Vario-Sonnar 70-200 f/2.8 SSM.  Obviously all of the above, in addition to the longer primes, will be supplementing their current lens range that also includes a Carl Zeiss Planar 85mm f/1.4 and the Carl Zeiss Sonnar 135mm f/1.8.  

[a href=\"http://www.dpreview.com/articles/pma2007/Sony/IMG_4806.jpg]http://www.dpreview.com/articles/pma2007/Sony/IMG_4806.jpg[/url]

I think their strategy of bringing out ultra-premium Carl Zeiss FF lenses (with expected mind-blowing performance corner-to-corner) is a sound one, in terms of establishing an immediate credibility in the market, when their pro-oriented FF model (essentially a more robustly built 5D) hits the road.

[Jonathan Wienke]

I personally feel an in-lens IS would be more effective than an in-camera IS (with sensor shift), primarily because the IS response can be tailored to the FL.  I also think that Sony will be introducing in-lens IS for their longer lenses, including their 70-200 f/2.8, which is being replaced with a new one....and have the in-camera IS automatically switch off, when the camera senses a stabilized lens as having been mounted.

That would be the sensible way to make it work...

[Farmer]

And does Sony have any long telephoto lenses available (400mm or longer), and if so, how effective is in-camera IS compared to lens-based IS of Canon or Nikon?
[a href=\"index.php?act=findpost&pid=162343\"][{POST_SNAPBACK}][/a]

I've used the KM7D with 100-300mm and 70-210mm natively and with 2x t/c and 500mm reflex with the AnitShake as it was called then.  Actually, have even used it with the 500mm reflex with 2x t/c without terrible results.  Performance is good - a genuine 1.5 to 2 stops allowing hand held at shutter speeds below 1/focal length.  The limitations were the t/c and the reflex lens.

The higher end Canon lenses perform better, but then so would higher quality lenses if I'd had them for the KM7D.

The Sony system is meant to be improved from the KM by all accounts - I'll have an Alpha 700 in the new year so I'll be able to see then.

Where the IS lenses certainly have an advantage is things such as tripod mode, that isn't available with the Sony system and some people much prefer to see the stabilised image in the view finder (I have no preference, having used both).

The main issue has always been how in-camera would function with a FF sensor.  The 1.5 crop obviously has a lot of real estate to play with and it would seem prohibitively expensive to put an oversized sensor in place in order to provide in-camera stablisation.  I support we shall see if the rumours are true and the Alpha 900 is FF.

[Jonathan Wienke]

For example, for non-interleaved RGB recording, with floating sensor and automatic triple exposure, one can get each image pixel with all three RGB sensors (time multiplexed 3-CCD or 3-CMOS).  With automatic 6-expsosure, one can take the same image pont with multiple-redundancy for error or noise correction.

Due to the GR-GB checkerboard layout of the sensor, you need 4 shots to record true RGB, not 3. Some medium-format backs do this, but only when tethered to a computer. The computer has to take the multiple frames and combine them; the back doesn't have the CPU horsepower/memory to do so by itself.

[AJSJones]

... The key is more in the repetitive/data-transfer rate of the multi-exposure. It would be much easier for landscape, close-up and anything not moving fast.

[{POST_SNAPBACK}][/a]

If Canon can solve that problem with multiple data pipes off the sensor (to read images so quickly that it allows 3 data sets in, say, 1/250 or 1/500 initially) they could actually use their [a href=\"http://patft.uspto.gov/netacgi/nph-Parser?u=%2Fnetahtml%2Fsrchnum.htm&Sect1=PTO1&Sect2=HITOFF&p=1&r=1&l=50&f=G&d=PALL&s1=6388709.PN.&OS=PN/6388709&RS=PN/6388709]patent[/url] to get sequential R, G and B values from the same photosensor element (like a Foveon but using a temporal rather than spatial array) that would be kewl!  However, such read rates seem to be a long way off...

Andy

[Provokot]

What do you base that on?
[a href=\"index.php?act=findpost&pid=162328\"][{POST_SNAPBACK}][/a]

I base that on the knowledge that these companies have vast sums of cash for R&D and no doubt a steady trickle of ultra clever boffins coming through their doors with nifty pieces of patented technology available at the right price...

[AJSJones]

It actually not as hard you might think. To continoue multi-exposure in high speed is one thing, but to get fast but triple (fix number) exposure is much easier.  I'll leave it for your mental exercise if you can access to eletroinc enginnering knowledge.

[{POST_SNAPBACK}][/a]

Well, my poor mental exercise capability helped me find a Cypress " high speed CMOS [a href=\"http://www.sensorsmag.com/sensors/Sensor+News/Cypress-Introduces-High-Speed-SXGA-CMOS-Image-Sens/ArticleStandard/Article/detail/466473]sensor[/url] " announced in October of this year, running 500 f/s for 1.3 MP, so ~ reading at 650 MP/s.  For a 1/500 sequential RGB exposure(as Canon's patent envisions), you need to read 3 x 1/1500 off the chip (one each for RG and B right?)  For a 6MP sensor (which for still objects might represent the info of a 12MP Bayer) this would be 9,000 MP/s.  Still a ways off, I think

[BernardLanguillier]

My view is that Nikon and Canon will have to follow.

It seems that building FF pro zoom lenses (24-70 f2. with built in VR/IS is still technically very challenging since neither Canon nor Nikon have managed yet to release such lenses. It probably is related to the size of the lens that would need to be moved to stabililize such lenses.

This will probably be the main driver for these companies to adopt in body VR.

Another possible reason is that it could in theory be possible to use both lens VR and body VR to have even larger gains. My guess is that Canon and Nikon have not released such bodies yet because they have not found yet a good way to mix in body VR with lenses VR. The challenge is of course to keep selling VR lenses after the release of VR bodies...

Regards,
Bernard

[elkhornsun]

Canon and Nikon already do this with their P&S cameras where it makes sense. For DSLRs better results are obtained with IS in the lens.

[Jonathan Wienke]

And it is no fundamental that IS body can't be superior to IS lens.

Not true. Lens IS will work at any focal length, while body IS is limited by how far the sensor can travel. This means that body IS becomes less effective as focal length increases, which is where it is needed most. Lens-based IS does not have this limitation; it is equally practical and effective at any focal length.

[DonWeston]

John - is there some reason for us laymen, why future technology could not address this in some way? One might think it is a frequency issue to the movement with increasing focal length as much as just the distance the sensor has to move? But that is from a layman's perspective...  With all the coding built in these days, any reason why the camera could not adjust its sensor movement to the specific requirements needed for a particular focal length, within the limits of a particular sensor...just wondering??

[AndyF2]

Another aspect is focus accuracy.  A camera with lens IS and a fixed sensor should have more reliable or more accurate initial focusing than a floating sensor.  

With both systems, the focus detectors are not on the image sensor itself, they're on the path split off by the mirror.  The lens focus will be adjusted so the image will be in-focus at the plane where the sensor is expected to be.  On a camera without sensor IS, the sensor is always at that plane, it's never shifted.  

On a camera with sensor IS, the sensor can be moved in several axis, and it would seem there is a possibility the sensor will not be dead center and parallel to the focus plane when focus is adjusted, and may therefore be slightly off.

This would certainly be something the sensor IS designers have solutions for, but it seems more controllable to have the sensor in a known position and control the image path upstream of the sensor and focus so both see the same results.

Andy