And finally what I forgot to say is that while the camera what a step up from the 1Ds2, it was only "incremental" and that some of the posts circulating on various forums were over the top. So we'll see - I guess depends what people expect.
KIDERAL wrote:
The 1DsII was announced September 2004.
The 50 1.0 was designed very similarly to the 85 so I would expect a 50 1.2L to be like the 85.
I think canon is having problems with throughput.. Digic III ... I don't think we will see a new fast camera soon.
Normally in new sensors, the 1.6x crop comes out first. issues of yield.
If a 30d sensor were made 1.0x then it would be about 21mp.
Interestingly, if the nikon 200d sensor was made 1.0x then it would be 28mp.
I got fooled on the release date of the 1DsII because Phil was ages late reviewing it! ;-)
What you say is interesting, as as far as I can gather the new 1DsIII is just a normal sensor, with no increase in DR or high ISO, and what you suggest about new sensors coming in at a smaller size would explain that, and also Canons' comment that 2007 is the year of DR - perhaps the 30D's replacements release will be the first with better high ISO and DR, or even the replacement 5D in Feb.
Regards,
DaveMart:
netexpress wrote:
I don't think Canon is having troubles with through-put. Coming from the computer industry I can tell you that moving that much data is trivial and Canon could do it any time they wanted. It is nothing new in IC design. Canon has the "know how."
Listening to Canon IR conference calls it is clear that Canon doesn't think they have any problems with the high end DSLR space. They think they can *double* their volume in this space effortlessly. That is huge. Whereas they think the P&S market is 90% saturated. I believe them on both counts.
What I think is going on is they have no meaningful competators in the 1Ds2 space. So why should they up the MP count faster than they need to? Just draw it out over the years so they can milk it and get a nice steady Y-O-Y revenue and profit growth like their investors want them to do. Things are moving up nicely for Canon. Why rock the boat? ...Show more →
Thom Hogan was saying that there might be problems getting data on to CF/SD cards - the fastest cards according to him would only take around 2fps for a 22MP image, so buffering requirements would be large.
I wonder if you would care to comment on that statement?
Regards,
DaveMart
I always assumed Galbraith measured megabytes (not bits). He shows 8mb per seond with CF. I believe SD cards are faster (no controller in each card). Anyways, I doubt flash memory is a problem. And what would a gig of buffer cost? Not much in money, perhaps too much in power usage and space.
Even if Canon could "only" do 22mp 5fps with better noise contror they will sell plenty of cameras.
Canon would cause concern with Canon-watchers (photogs and financial guys) if they didn't replace the 1DsII. That's why I'm confident it will happen. I'll see if my dealer will take a tentative order for a 1DsII replacement.
DaveMart wrote:
Thom Hogan was saying that there might be problems getting data on to CF/SD cards - the fastest cards according to him would only take around 2fps for a 22MP image, so buffering requirements would be large.
I wonder if you would care to comment on that statement?
Regards,
DaveMart
I think you answered your own question. Current CF cards are too slow and that needs to be addressed and it will eventually be addressed by innovative companies like Sandisk as the need arises but not directly by Canon. That’s not Canon’s problem. So Canon probably will increase the buffer size to compensate just as you are implying. Doing so will be trivial in terms of cost to Canon. If you analyze the semiconductor industry one thing because perfectly clear, production continues to far outpace sales. When supply outstrips demand prices continue to fall. So expect cheaper and cheaper memory. For example look at the current growth rate of the supply of wafers vs. sales (demand):
Supply is growing at over a 20% rate while sales are only growing at a rate below 10%. You do the math. Canon has nothing to worry about in terms of bundling more buffer memory with their flagship 1 series camera in terms of cost.
Even with the buffer issue aside, look at Canon’s competition in the 22MP – 39MP market: Leaf, Phase One, Sinar, Hasselblad/Imacon. All of these 22MP – 39MP backs are slow writing to CF cards. Even with a Sandisk Extreme III it takes the Leaf Aptus 75 33MP digital back 9 seconds to turn and then shoots 25 frames in 60 seconds (2.5 seconds per frame). The Phase One P45 39MP digital back takes three seconds to turn on and then fires 21 frames in 60 seconds (2.85 seconds per frame). The Aptus 75 doesn’t burst faster than 2.5 seconds per frame. The P45 can burst for 7 frames at 1.5 seconds per frame.
Note carefully I’m talking about *seconds per frame* and not *frames per second.* This is slow! Can Canon do better with a bigger buffer and do it cheaper to boot? You bet. They don’t have to do much to kill the competition in this market in terms of speed. They have resources that dwarf Leaf, Phase One, Sinar and Hasselblad/Imacon combined.
And I’m saying this as someone with a lot of money tied up with Hasselblad
I am also wondering. wondering if the camera-designers of canon are only able to improve sensors. who really needs 22mp ? and who on the other side would like better dr, af, ergonomics and dust protection. since these are the things i care - and everybody i talk to, thinks the same - the development of canon cameras over the last years has been very disappointing.
Incidentally if you next question is with bigger and bigger sensors coming out each year and with iPODs, etc all increasing the demand for flash memory isn’t the increasing demand going to cause the price to skyrocket on compact flash? Can they make compact flash cards big enough? The answer again is prices will drop and sizes will get plenty big. Supply growth continues and will continue to outstrip demand growth. Just look at spot wholesale prices we’re seeing this year for NAND devices that we use to build compact flash memory:
The prices are plummeting. This is why a lot of insiders in the semiconductor industry don’t buy semiconductor stocks. They know just how toxic this stuff really is.
philllie1 wrote:
I am also wondering. wondering if the camera-designers of canon are only able to improve sensors. who really needs 22mp ? and who on the other side would like better dr, af, ergonomics and dust protection. since these are the things i care - and everybody i talk to, thinks the same - the development of canon cameras over the last years has been very disappointing.
Canon have not been too keen on Lithium batteries for their pro cameras (can't think why - they should buy Sony :-) )
So they have been using the relatively bulky NiMiH batteries.
My guess is that they have been leaving a re-design until they can decisively reduce the wieght and size of the batteries.
The joke is wearing a bit thin though.
Regards,
DaveMart
netexpress wrote:
Incidentally if you next question is with bigger and bigger sensors coming out each year and with iPODs, etc all increasing the demand for flash memory isn’t the increasing demand going to cause the price to skyrocket on compact flash? Can they make compact flash cards big enough? The answer again is prices will drop and sizes will get plenty big. Supply growth continues and will continue to outstrip demand growth. Just look at spot wholesale prices we’re seeing this year for NAND devices that we use to build compact flash memory:
The prices are plummeting. This is why a lot of insiders in the semiconductor industry don’t buy semiconductor stocks. They know just how toxic this stuff really is. ...Show more →
Nah- got a much better question than that! ;-)
How much do you think it would affect costs to go to 16bit in increased storage, mainly to cope with better DR, as Canon have said that they hope to increase that in 2007- but not just the memeory, the cost of the processing power and so on?
Regards,
DaveMart
DaveMart wrote:
How much do you think it would affect costs to go to 16bit in increased storage, mainly to cope with better DR, as Canon have said that they hope to increase that in 2007- but not just the memeory, the cost of the processing power and so on?
Regards,
DaveMart
I thought that DR is a matter of sensor sensitivity, not 12 bit vs 16 bit in-camera processing? Or are you asking about buffer costs under the premise that Canon has figured out the sensor problem?
I would think (guessing, of course) that the increased cost of a more sophisticated sensor with more DR, outputting in 16 bit instead of 12bit, would be much, much more than the increased costs of any changes to support/storage chips.
JRMDC wrote:
I thought that DR is a matter of sensor sensitivity, not 12 bit vs 16 bit in-camera processing? Or are you asking about buffer costs under the premise that Canon has figured out the sensor problem?
I would think (guessing, of course) that the increased cost of a more sophisticated sensor with more DR, outputting in 16 bit instead of 12bit, would be much, much more than the increased costs of any changes to support/storage chips.
Absolutely, increased DR is dependent on the sensor, not how the data is packed.
But Chuck Westfall indicated on the old Rob Galbraith forums (sorry, no link, as that site has changed) that Canon saw more bits as necessary to record the info an increased DR camera would provide.
Furthermore, I believe ( again, no link ) that they have said that 2007 was to be the year of increased DR, so presumably they feel that they can crack the sensor side of things.
I have to admit that perhaps I was over-estimating the difficulty of dealing with the data rates, as you would only be going from 12 bit to 16bit, but so many relattively inexpert people (just like me!) seemed pretty awstruck at the thought of moving 22MP at 8fps that I had wondered if it would be a major difficulty to simultaneously increase DR.
In fact, I now believe from what I have picked up on this thread that Canon will increase DR on a lower camera than the 1 series, perhaps the 5D or 30D replacements, and perhaps next year, although it is possible that they have already made a move on increasing high ISO capability in a 1DsIII.
Regards,
DaveMart
JRMDC wrote:
I thought that DR is a matter of sensor sensitivity, not 12 bit vs 16 bit in-camera processing? Or are you asking about buffer costs under the premise that Canon has figured out the sensor problem?
I would think (guessing, of course) that the increased cost of a more sophisticated sensor with more DR, outputting in 16 bit instead of 12bit, would be much, much more than the increased costs of any changes to support/storage chips.
I would tend to think so too. The increased cost of handling the data flow due to increased DR assuming a higher bit width would be trivial as a percentage of the total cost of a DSLR. The cost of designing and producing a CMOS sensor that has increased dynamic range vs. the current sensor is a question that I’m not even sure that even Canon has a handle on yet. However, I would hazard to guess that the cost of re-design needed to increase DR could fit snugly into their total R&D expenditures and hardly be noticed. If you have some idle time visit Canon’s Investor Relations page. It has fascinating statistics and the amount they allocate for R&D is huge compared to most companies: http://www.canon.com/ir/index.html
.That bodes well for all of us Canon owners in the future. But making one last point, the whole idea of using CMOS is that it is much less expensive not only to produce but also to re-design compared to other competing technologies once you get past the initial investment. So they have an advantage there by virtue of using a CMOS that other companies don’t have.
trueimage wrote:
why are the cards not helping? its the camera
according to rob galbraith's cf database, a 1Dmk2 can do about 9 MB/s RAW
ultra II cards do 9 MB/s
extreme 3 do 2x that
so why is it the cards?
Sounds like Thom had not allowed for the extreme 3!
As for the extreme 4, perhaps delaying the 1DsIII until next yar would mean that they can specify that speed and economise on buffer, although I am sure that they would not delay for that reason alone.
Regards,
DaveMart
trueimage wrote:
why are the cards not helping? its the camera
according to rob galbraith's cf database, a 1Dmk2 can do about 9 MB/s RAW
ultra II cards do 9 MB/s
extreme 3 do 2x that
so why is it the cards?
The camera's don't read /write as fast as the cards are capable of. The new Ex IVs have a different file system that allows for 40mb/sec. Some MF cameras have updated firmware that uses this format and should boost speed. No published numbers yet. I would be very surprised if any new 1series body does not use the EXIV's speed...15mb/sec or faster is my guess.
wjmeyer wrote:
Canon does not have an L series lens in their EF-S line do they. I'd be curious to see if they don't introduce an "L" series lens for EF-S to see if they can establish more of a professional following for their 1.6x crop bodies.
The top EF-S lenses are L quality or very near L quality optically. They are not as built as robustly, which can be an issue for photojournalists, but less of an issue for other photographers.
JRMDC wrote:
Since all L lenses fit those cameras, the question becomes what unique EF-S lenses are needed that are not covered by current L glass. (Keep in mind that and EF-S specific lens of any focal lenght and aperture is not going to be much smaller than the EF equivalent, as much of a lens' size is determined by the size of the front element, which is determined by focal length and aperture and is independent of the sensor size it is to project on.)
This is wrong. In the wide angle range, EF-S lens can be and are smaller, not only smaller than the equivalent, but smaller than the same focal length. Wides and wide zooms must have a retrofocus design (reverse telephoto) to clear the mirror. The 1.6 cropped cameras have correspondingly smaller mirror boxes, which means that a 35mm full-frame lens must have a retrofocus design, but the EF-S equivalent would have a normal design, hence smaller. I'm not sure where the cross-over point is, but I think it would be around 30mm in crop and about 45mm in full-frame. Further, even when in the retrofocus range, the rear element can be closer in EF-S than in EF, meaning the EF-S lens has a size/weight advantage.
Compare at equivalent focal lengths (and same aperture f/2.8) the EF-S 17-55mm IS versus the EF 24-70mm at 645g vs. 950g.
Compare at almost the same focal length and f/2.8 the EF-S 60mm macro vs. EF 65mm macro, but 335 g vs. 881 g.
I do agree with the thesis that the EF-S line is ripe for some sharp primes. A 50mm f/1.2 EF-S would make some sense and could be done for about 1/2 to 2/3 of the price of an EF 50 f/1.2 L, but the howls of disappointment from all the full-frame partisans would be intense! Likewise a 32mm f/1.4 EF-S would be cool and could easily be 1/2 to 2/3 of the cost of the 35mm f/1.4 EF L. Personally I'd love to see a 14 or 16mm f/2.8 EF-S sharp corner to corner (22 or 24mm equiv.), though I'd love even more to see an EF 21mm f/2.8 or f/3.5 sharp corner to corner for about $1200 that would give the Zeiss a run for the money.
Many rumors here and elsewhere are nothing more than wishful thinking, however, reflecting personal desires and not actual information.
Monito wrote:
This is wrong. In the wide angle range, EF-S lens can be and are smaller, not only smaller than the equivalent, but smaller than the same focal length. Wides and wide zooms must have a retrofocus design (reverse telephoto) to clear the mirror. The 1.6 cropped cameras have correspondingly smaller mirror boxes, which means that a 35mm full-frame lens must have a retrofocus design, but the EF-S equivalent would have a normal design, hence smaller. I'm not sure where the cross-over point is, but I think it would be around 30mm in crop and about 45mm in full-frame. Further, even when in the retrofocus range, the rear element can be closer in EF-S than in EF, meaning the EF-S lens has a size/weight advantage.
Compare at equivalent focal lengths (and same aperture f/2.8) the EF-S 17-55mm IS versus the EF 24-70mm at 645g vs. 950g....Show more →
Thanks for the correction, I just keep on learning!
Monito wrote:
Compare at almost the same focal length and f/2.8 the EF-S 60mm macro vs. EF 65mm macro, but 335 g vs. 881 g.
I find it hard to believe that this extreme weight differential is solely due to EF vs EF-S. Is there a difference in macro ratio, or something else?
Monito wrote:
Compare at almost the same focal length and f/2.8 the EF-S 60mm macro vs. EF 65mm macro, but 335 g vs. 881 g.
I find it hard to believe that this extreme weight differential is solely due to EF vs EF-S. Is there a difference in macro ratio, or something else?
New comment: I just checked, the 60 is a 1:1 macro, the 65 does 5:1, that will explain part of the sizeable differential. I understand the general point, however.
? and who on the other side would like better dr, af, ergonomics and dust protection. since these are the things i care - and everybody i talk to, thinks the same - the development of canon cameras over the last years has been very disappointing.