UPDATE
Aren’t patents fun? One thing I hate about patents is I can’t read them. I read through the 4 of these and still had no idea what they were.
One of our commenters did however.
“The first patent is describing a read out shift which only is useful in video recording. The patents 2 and 3 rely on the optical IS element doing the dirty work and the third describes a way of moving the relevant lens elements in an IS lens. So nothing, as in not a single thing, is implicating more than the presence of sensors that detect motion inside the body – the chore of correcting for that motion falls to the IS group.”
So there you have it. Thanks Karl. I love when these things are explained in english.
Canon is apparently top 5 every year for patents awarded. They were #2 in 2005 behind Big Blue.
In Body & In Lens Stabilization Patents
I haven’t had any time to read through these patents thoroughly, I only have my phone. However, I’m told the following 4 patents may be of great interest to Canon users.
20090231444 PDF
20090231452 PDF
20090232483 PDF
thanks Sean
cr
|
When you purchase through links on our site, we may earn an affiliate commission. Here's how it works. |
55 Comments
The first one apparently is intended to reduce noise inherent in image shifting IS. I see it is a tweak to existing IS methods which is intended to give a greater S/N ratio, and presumably better high ISO performance.
Its most likely to see use in camcorders first.
What if canon released in body IS on it’s cameras that could work in unison with it’s IS lenses to give you a couple more stops of IS…..that would be AWESOME!! Then I wouldn’t have to buy IS lenses unless I wanted 6 stops!! I’ve always hated Sony users for this one feature! Okay, I don’t hate them, but I’m VERY jealous. Then my Tamron 70-200 would be a whole new lens for portraits. I could actually hand hold it at 1/60th!
Alternatively, buy a flashgun or two.
If you can’t tell the difference, go for it.
Flash is not always the effect I’m looking for……..I usually like natural lighting when possible. I do have a couple flashes though already, in body IS would just give me more flexibility. I would only buy a body with IS if Nikon or Canon made it. Sony still can’t control their noise very well yet, and their features are still seriously trailing the other two.
Important to note here that applying for a patent is sometimes done to block competitors from owning or applying a useful idea. It doesn’t necessarily mean that Canon intends to introduce in-body IS to its system. But given that most mfr’s have either les or body IS,it could be true that Canon is planning to have in-body IS. and that would be a great news, if it materialize Canon becomes the first brand to have the two types of IS. And this would put Nikon in a difficult corner where they naturally have to respond to Canon when finding an in-body way would be difficult because of the existent patents of each in-body mfr.
Since all others do have in-body IS (*before* Canon), there are obviously no patents in Nikon’s way.
Some of what you describe occurs, but generally those are singular patents. What’s been described here is, IMO, an elaborate and thought-out system likely at least in the testing phase. If you pay attention to patents and their publication, they are very often timed to be very close to product annoncement (go back and look at most announcements which involved marked improvements or new types of products).
What’s going on on the 29th again?
It looks like this would still result in some electrical noise that will have an impact on high ISO performance. I wonder what “reduced” really means? Did they file this patent just to stop Sony from implementing something similar?
If I had to choose I would prefer high ISO to IS. You can’t stop action with IS. ….more of both is of course always better.
I guess if Canon did offer in body IS along with better high ISO performance than the 5DMK2, along with a F1 17->300 lens that was tack sharp for under $1K, I might be satisfied. …though I might feel deprived that they took away good stuff to argue about.
Maybe the 1DmkIV will have in body IS! That would be a HUGE announcement. That may justify Canon’s bold claim for Sept 29th,
“In the autumn Canon will announce one of their biggest and most important product launches in the company’s history”
The first Canon body with IS would fit that bill I think. Then if it had 12fps at 18mp, and the new focus system 7D with 45 pts instead of 19. Hmmm……
From Sony folks I’ve talked to, this is actually very similar to the way early Sony in-body IS worked, so I doubt that this was a block to keep them from going there. More likely I think that Canon and Sony are squaring off right now for a real match over quality DSLRs, each out for blood in the market. Great news for us, of course, as it’s doing far more than Nikon has to get corporate off their hindquarters (see: 7D).
I’m also going to guess that if the noise from the electronic part of the IS system is significant that it’ll either be disabled in ISO expansion, or separately enabled / disabled through CFs.
I suppose anything’s possible, but to the engineers out there ….
Is it practical to combine lens-based IS with sensor-based IS so the effect is cumulative? (Say, 3-4 stops from lens-based IS plus a few more stops from sensor-based?)
That would be an obvious winner.
Or would you be forced to switch from one IS system to another based on which lens you are using?
Still a winner, but not a home run.
I dont think they would ever have full time in body IS, as you said im sure it would be either a c/fn or easily assigned to a button on the back of the camera or maybe you would set parameters for it to kick in (eg under a certain shutter speed, over a certain lens length etc)
I thought it had already been established that the 29th announcement is not going to be EOS related?
Has anyone taken the time to read the patents? Not just the “There is a patent pertaining to in body-IS?”
Well if you had, then you would have noticed that these are
a) video stabilizing techniques (the first patent) or
b) descriptions of the sensors in an classical IS lens setup…
The latter could well be the patents on which the new hybrid IS is based upon…
Ummm, yes. That’s why I submitted them.
And you didn’t understand that the patents 2 and 3 have nothing to do with a in-body IS system as they are tied to an optical correction element – normally called an IS group in a lens?
While you could be right, I happen to disagree as to what it’s describing based in large part on the block diagramming and language used. I will happily acknowledge that this may well only be intended for use in P&S bodies based on its descriptors, but that it could be used in DSLRs is, IMO, equally likely.
Is canon a patent troll?
Initially everybody was convinced that two independent stabilization systems are likely to hurt the effectiveness of each other if shooting with both systems activated, because they cannot coordinate their action in a way that would maximize the effect, this was the theory, but a new test suggests that may not always be true, see here:
http://www.1001noisycameras.com/2009/09/cool-stabilization-test-with-surprises-lens-vs-sensorshift-pentax-k7-and-olympus-ep1.html
One thing I am certain of is that it describes systems that have nothing to do with an in-body stabilization system. I am a professional software developer and reading block diagrams and flow charts is part of my daily work. The first patent is describing a read out shift which only is useful in video recording. The patents 2 and 3 rely on the optical IS element doing the dirty work and the third describes a way of moving the relevant lens elements in an IS lens. So nothing, as in not a single thing, is implicating more than the presence of sensors that detect motion inside the body – the chore of correcting for that motion falls to the IS group.
OK
I actually opened the first patent and few other and got the feeling that they are not talking about in-body IS at all, but because I am a biologist and don’t understand nuch of tech-talk, I said to myself (wishful-thinkingly maybe), “I must have misunderstood becauce this kind of douments is too difficult for me”. But regardless, there has always been rumors of canon adding in-body IS, unless they make all their lenses stabilized, IS will always be badly missed on non-IS lenses.
some people didn’t get the memo
reading flow charts and patents are not related at all.
If you knew how to read patents then it would be a different story.
which is weird, since they make the nikon sensors…
I seriously doubt the IS is for DSLR’s, probably for P&S’s.
It’s not weird, the sensor in the Nikon D3x is optimized for low light performance, the sensor in the Sony Alpha is optimized for dynamic range. The Sony has about 1 stop of extra dynamic range than top Nikon or Sony. Actually only the Fuji Super CCD cameras compete with the Sony in terms f dynamic range.
could this be IS for point and shoots or video and not SLRs?
Is there any clue that these technologies might be in their bigger sensored cameras?
Is there anything that might point to something like Micro Four thirds Type cameras or like the Leica X1?
or are all of these improvements on current things already out there?
“What if canon released in body IS on it’s cameras that could work in unison with it’s IS lenses to give you a couple more stops of IS…..that would be AWESOME!! Then I wouldn’t have to buy IS lenses unless I wanted 6 stops!!”
Nobody but nobody actually needs six stops of stabilisation, if you think you need it there is something seriously wrong with your technique; additionally six-stop stabilisation basically has no practical use. Let’s think about it – six stops from say 1/100s is 1 second, if for whatever reason you want to take a picture of something with such long exposure the object itself my be rather still and you will be better off using a tripod anyway. There are certain limits to how technology can and should be used, 2-, 3- ever 4-stop stabilisation makes a lot of sense as it enables to drop shutter speed while taking pictures with longer lenses, anything more than that is simply and unrealistic and impractical overkill coupled with poor technique, if you think you need it, you probably need either more light (flash) or a faster lens.
Karl Günter Wünsch should know, he is an expert…
“One thing I am certain of is that it describes systems that have nothing to do with an in-body stabilization system. I am a professional software developer and reading block diagrams and flow charts is part of my daily work.”
No direct indication of a M4/3 or the ilk in anything currently published, though sensor sizes are directly elucidated. No EVF mentions save for the P&Ss. And the technology may well be intended for P&Ss, though I think any distinction between DSLRs and video at this point would be arbitrary (who thinks the 1DIV will *not* have video?).
*sensor sizes are NOT directly elucidated
Nor should they be if they want decently broad patent. Plus I hope everyone’s analysis is based on the claims, not the written description.
*That*’s the important part? ;)
“their biggest and most important product launches in the company’s history”
Caught on the road here: http://i629.photobucket.com/albums/uu19/cameratribe/CanonLLens.jpg
The biggest L lens yet!
(reference to the updated version of the post) Regarding quantity of patents, not sure Big Blue will stay first for many year. Our rules for patent (I work for Big Blue) has changed a lot this year… I guess tired we are (stock holders) tired spending too much money on “vile” patents (believe me, there is a lot that are not going anywhere). Samsung should be first within the next 2 years. Where Canon will stand… I wish they will focus on quality instead of quantity.
IBM is one of the largest patent trolls in America. Which rules regarding patents have IBM changed this year? They may have a huge effect on industry legal tactics.
The first patent appears to be a variation on a CCD image sensor that scans in two dimensions (horizontally and vertically). What this patent is describing is a method for reading out the image data from the sensors without smearing an image whenever the imager has capture a lot of light info (a common problem with CCD’s – whenever a bright light is aimed at the sensor [like a flash light etc.] then the light streaks across the image as a result of how the image data is read out of the sensor).
So it may be that Canon is researching the use of either CCD’s and eliminating this effect – or incorporating the two dimensional readout system used in CCD’s into CMOS imagers.
Pretty cool.
If anyone ever tells you “I am certain that it describes” in the same sentence of “patent” then you should stop listening. The court will construe the patent as narrowly as the claims allow, and the written description is simply a statutorily mandated aid to assist the PTO and courts should they choose to rely on it. Claims are king, and they are construed in their own court hearing called a Markman hearing. So NOBODY knows exactly what is claimed until that hearing ends, And even then it can be appealed.
I just read the claims of the second patent. The claims are what courts rule on, NOT written descriptions.
The independent claim is for “a correctable unit moving orthogonal to the optical axis.” This is not limited to moving glass, just as it is not limited to a moving sensor. “Unit” can describe either. The dependent claims, which serve as insurance against infringement by previous broader claims, do not seem to limit this to a glass moving system.
Bottom line: Canon could use this to protect a unit that has a sensor mounted to it. If somebody really wants to present a coherent argument against why this does not cover an in body sensor, please cite prior air or failure to enable through the written description. That would be more convincing.
PS – I’ve got two engineering degrees, am studying patent law, and worked at a patent litigation firm. However, I am not qualified to give legal advice, and my previous statement was simply a non-legal opinion.
The last patent deals with image stabilization through rotation of the sensor or lens. It corrects for the only thing currently NOT corrected on IS lenses, the rotation of the body. Current systems correct for ‘pitch’ and ‘yaw’ as well as image shift up and down on the newest macro lens. The ‘roll’ of the camera is what this system seems to address. It’s all in the patent, silly.
Actually shift is the only thing currently stabilized, as I don’t believe any IS lenses have been released yet with angular compensation – unless the new EF-S ones which came out with the 7D are? But you’re quite right that the ‘roll’ aspect is the key in this patent.
The EF 100mm f/2.8 L Macro is the first Canon lens to have that “hybrid IS” system.
Meanwhile I imagine correcting for roll with optical element(s) would be quite complicated.
I read that Pentax has managed to add stabilization to camera rotation in addition to shift, (in their K7 or K20), which of the 3 axis of rotation they were able to stabilize,I am not sure. The best stabilization system is a three-legged one.
Maybe it’s IS for photocopiers
The point here is I feel there is a kind of maturing on how patent are handled. Being the first in patents quantity will not necessary put you first market wise. But, focusing on your best patents and be first on the market will keep you ahead of the competitors. Canon is a bit like Big Blue, being very well vertically integrated (having their own R&D and fab, from silicon to final product, with no one in the middle that you need to depend on). Patents could protect your technologies, but if you spend to much money on your technologies and don’t do enough money with it, what is your final gain. So, the new rule is something like this: “less patents with a greater focus on getting more money from your best one”. That should be our new trend. R&D is freaking expensive, we should not stop it, but use it more wisely. There is nothing new here… but living it is another story.
I believe Canon made a mistake lately (let say within the last 3 years)… they under estimated Nikon and thought they were kind of alone. They played the game of minor updates in their newer model (50D for example) and opened the door to Nikon. I’m Canon and I know they have excellent product, but it’s not that clear who has the lead right now (yes in quantity sold, but I mean leadership). They are getting better, 5DII is a good example. Next 1D(1Ds) should confirm that as well. 7D on paper is not bad either (still questioning the 18MP though, but we will see).
If we come back to Big Blue, we still have the lead in what we do best, but it’s a freaking race to stay first… end of comments
Being a patent troll isn’t about R&D. Its about prospecting turning into anti-competitive conduct. IBM is notorious for doing this, to the dismay of many in the industry and joy of law firms. Just search “IBM patent troll.”
They even tried to patent being a patent troll. Chilling effect on innovation I’d say.
LMAO…. particularly useful in California and Japan.
The first patent seemingly to me(too long to finish) is to read image sensor vertically and horizonally asychonously, then average the two value to remove noise(noise values are also averaged). We can expect super clean image if applied, but I worry about color accuracy though.
The second one is an active image stablizing system which an analog system readout the vibration, 2 ADC(?)convert that to digital signal, after calculating that signal, a movable part (in the lens or not it didn’t say, very likely is in the lense) is driven to compensate.
People has measued that Canon camera’s noise is mostly generated during the Read phrase, so the first patent might be able to greatly improve image quality.
The 3 patent is almost the same as 2nd one might be more detailed.
The 4th patent is the machinery of patent No.2 and 3, which looking like a powershot or camcorder lense to me.
Groundbreaking…literally.
Oh, man, Johnny’s gonna LOVE that one.
One thing’s for sure: given the exploded diagram and nature of lens groupings as visible in very detailed patent drawings on the 20090232483.PDF – this is certainly no SLR lens.
Given the nature of others’ comments regarding video feed, it could be a highly-integrated video component of a point-and-shoot variety, or a new video camera about to be released.
Anyway, it’s non-SLR, so I’m not that piqued.
Many people have alos become fans of Paul Smith Shoes since they came into the market. I also one of them , I very like Paul Smith Wallets and <a Paul Smith Belts,It designs very good.