I'd like a eye control focus, solid grip, a little more resolution, and 1.3 crop like most others. But let me throw something out that no one else has brought up. How about the ability to override the IR filter, so IR shots are again possible with normal exposure times?

jeffmock wrote:
Burning incense...

I don't think there are going to be new bodies in the near term, they seem to be aiming for a 3-4 year design cycle on the bodies. [...]

Looks more like 2-3 years to me in the 1-Series bodies and even shorter intervals in the 20D catergory. Little summary of the release dates (from the camera museum)

1. Oct. 2000 - D30
   
2. Mar. 2002 - D60
   
3. Mar. 2003 - 10D
   
4. Sep. 2004 - 20D
   

1. Dec. 2001 - 1D
   
2. Nov. 2002 - 1Ds
   
3. Apr. 2004 - 1DII
   
4. Nov. 2004 - 1DsII
   

Philippe Arnez wrote:
Looks more like 2-3 years to me in the 1-Series bodies and even shorter intervals in the 20D catergory. Little summary of the release dates (from the camera museum)

1. Oct. 2000 - D30
   
2. Mar. 2002 - D60
   
3. Mar. 2003 - 10D
   
4. Sep. 2004 - 20D
   

1. Dec. 2001 - 1D
   
2. Feb. 2002 - 1Ds
   
3. Apr. 2004 - 1DII
   
4. Nov. 2004 - 1DsII
   

In a word : In-camera HDR*.

Canon has a patent describing how dynamic range expansion could be done on-chip but whether they are close to having it ready for demo is (still) anyone's guess!!

Adding another couple of stops' worth of range in the sensor would be "quite nice", and, for me for landscapes, it could be at the expense of fps or ISO range.

Andy

*If you've read the PS CS2 flyer from e.g. Martin Evening and others, 32-bits per channel (yes, per channel) processing is available in CS2 and will be in "more digital cameras .. in the not too distant future.." You will be able to create one humongous file from a bunch of bracketed shots (shot separately) and then convert to 8 or 16 bits per channel to compress the image's whole dynamic range into something presentable on screen or on paper - blending "easily" and controllably implemented.

Guys, I'm not one to push newer and better, because it's not the tool that takes good photographs, it's the photographer. But I can tell you that the 20D is one heck of a lot better tool than the 10D. I've had them all, since the D30 and this one is special the way the D30 was. The D60 and 10D were very marginal improvements (especially the 10D) over their predecessors (and in some cases not as good), but the 20D, wow. It's like a different class. And it's not the number of pixels. I also have a 1DS and will not be selling the 20D unless there's a real big leap in function/features.

Theresa



zonkola wrote:
JeffBowser wrote:
I'm hanging onto my 10D until the 20D replacement comes out at least. Every other, or even every 3rd generation is as fast as I am willing to upgrade. Upgrades for the sake of upgrading is a losing proposition.

I'm shooting on a 10D for exactly the same reason.

Both Mamiya and Hasselblad now have 22 megapixel cameras. The Hasselblad is pretty big and bulky and will probably only be useful in the studios but Mamiyas ZD as I understand it will accept all the 645 lenses and will handle like a DSLR. At 22 mp this may have some influence on a future Canon Camera. There are quite a few photographers out there with 645 lens that I am sure will be interested in this model. Estimated price is between $10 and $15,000.00 so it won't be for everyone.

Grant

Philippe Arnez wrote:

Looks more like 2-3 years to me in the 1-Series bodies and even shorter intervals in the 20D catergory. Little summary of the release dates (from the camera museum)

1. Oct. 2000 - D30
   
2. Mar. 2002 - D60
   
3. Mar. 2003 - 10D
   
4. Sep. 2004 - 20D
   

1. Dec. 2001 - 1D
   
2. Feb. 2002 - 1Ds
   
3. Apr. 2004 - 1DII
   
4. Nov. 2004 - 1DsII
   

One thing is that now that Canon has perfected the CMOS sensor you will always see the best "version" of it come out in a 1-series body before a 10 series.

the 1D was the exception because even though canon had the CMOS for the d30 it was still kind of primitive then - the 1Ds is what fixed that.

I think the two year interval on 1-series cameras will stick because the pros want tangible new featuers to justify an upgrade - the MKIIs are significant upgrades over the MKIs.

the 10D was something of an anomaly - only one year after the D60 - but that shows as the 10D has only minor improvements over the D60. the 10D was prompted as a lower-cost-to-produce camera over the D60 so that Canon could stay compeitive against competitor (ie NIKON) pricing (in the dSLR market you have to stay on top or be dead... look at pentax, olympus, minolta, fuji, kodak - they're all in the dSLR market now but because it took them so long they've fallen way way behind!)

unless a competitor does something serious to attack the 20D it will be the flagship non-1-series camera (I would not call it "consumer" or "prosumer" - it's a budget-grade "pro" camera - why? because your average consumer wouldn't be interested in purchasing one) until after the next 1 series comes out with significant nicer and better features.

what will the next series of camera have?

(educated guesses)

-the 1DmkIII will have more than 12 megapixels - to smoke the D2X - and a faster than 8.5 fps - and may or may not do the full-frame thing.
-the 1DmkIII will have more than 22 megapixels - to smoke the Mamiya ZD or whatever else comes along - and will still be full-frame
-the next 10 series will be a "lite" version of the 1DmkIII.

EF-S will stick around because the DoF on an APS-C camera is suffiecient to get good bokeh even with cheap lenses. I would not be surprised to see a 1-series with EF-S... although if that happened Canon would release an "L" Lens in EF-S with a really wide angle to compensate for that. Canon will be releasing more and more EF-S lenses over the next few years for 20D's and Rebel XT's.

Full-frame is only really needed for wide angle but the 10-22 solves that. a brighter L version would not be a major feat for Canon to put out. In the film days you only ever went with less than 16 if you wanted a fisheye...

As for dynamic range, the technology on the CMOS chips is so good that physical size won't make much of a difference nowadays. I do see RAW file technology improving to the point of keeping bracketed shots in the file info so your final images would have super range on them. I also see in-camera processing becoming smart enough to automatically bracket a shot and produce a combined RAW or JPEG file on it's own. I also see even less noise at high ISO levels - the 20D is a vast improvement over the 10D - now the 1600 and 3200 ISO modes are fully usable on the 20D wheras on the 10D they were more of a last resort option.

As the cameras get higher res the lenses will have to get better (and they will cost just as much as the do now for good L glass!)


I find that with technology that devices get better - even way better - but prices only drop slightly.

One thing is for certain - the futuer for dSLRs looks very bright indeed.

Do a search on previous threads. I'm sure there is one almost identical to this one, but they were talking about the camera after the 1D and 10D.

I also agree that it's too soon. Probably at least another 12 months before another pro body comes out. I think they already milked the technology they created in this round to create the set of cameras you see today.

Rebates aren't genuine price decreases. They're a marketing gimmick that doesn't reflect any change in the actual wholesale price.

Rebate offers spur sales, but in fact, such a small percentage is actually redemed that most rebate offers are handled by insurance companies. The manufacturer/dealer pays a premium to an insurance company, and the insurance company makes the payout to the customers who actually submit a proper redemption form. The manufacturer gets more sales, the insurance company gets paid a premium, and doesn't make too much of a payout to customers.

A rebate is no indication of anything going on on the production side.

I would not hang too much on the release cycles up 'til now. Unlike others, I think Canon DSLRs are post-adolescent, and I think the cycles will slow in terms of real performance/image quality upgrades. They may continue to release new versions, but the changes will be incremental (USB 2.0 to Firewire, or a new DIG!C).

Earlier, Canon had three major issues to overcome in their DSLRs:

1. Compared to 35mm, the DSLR cameras had very slow handling characteristics.
3. Compared to 35mm, the DSLR cameras lacked in quality up to A3 enlargements (larger than that was not a primary issue).
4. Compared to 35mm, image quality varied too radically across the DSLR line.

It was necessary to overcome all these issues as quickly as possible, yet continue to sell cameras while doing so. Thus we saw short cycles and radical improvements in either image quality or handling with each cycle. As of the release of the 350XT, all three of these issues have been met. We are effectively "there" in terms of matching 35mm performance for most purposes.

The last major "handling" issue will be an 8fps 24x36mm 1D camera at $5000, which at least two Canon Japan execs have said is coming. The Canon-Spain execs hinted at a possible "October surprise" of a completely new camera in the lineup, but their hint was extremely specious.

This doesn't mean Canon won't continue to improve their cameras, but as we've seen with computers, the improvements won't be those that will greatly change the capabilities of the camera (think of the changes from the F-1 to the F-1n to the New F-1, or the various AE-1 models as examples).

Philippe Arnez wrote:
jeffmock wrote:

1. Oct. 2000 - D30
   
2. Mar. 2002 - D60
   
3. Mar. 2003 - 10D
   
4. Sep. 2004 - 20D
   

1. Dec. 2001 - 1D
   
2. Feb. 2002 - 1Ds
   
3. Apr. 2004 - 1DII
   
4. Nov. 2004 - 1DsII
   

Michael, the chart for MP is very consistent with the release date, a continuous linear increase. However, do you anticipate the chart may took a different function since from 2000 to 2005 Canon have been comfortably enjoying the easy competitions but this may all change soon?

Son,
I think I remember reading, either some discussion here and/or an interview with Canon directors that suggest Canon and perhaps Nikon have technical improvement paths that are not necessarily driven by competition between them. We should all remember that product planning MUST start several years before its planned introduction so Canon have to have a continuous improvement plan and I don't think they can rest on their laurels unless they can see Nikons 3-5 year plan while they are making theirs. Consequently I believe that the trend graphs do hold some merit.

For the next upgrade to the 1Ds family I think that Canon already have the pixel. They can shrink the 1Ds2 pixel from 7.2um to 6.4um and get a 20.5MP sensor. I'm not sure if that is a big enough step from 16MP though. To maintain throghput that may also have to upgrade their Digic II processor as the file size will increase.

For the 1D family if they use the 6.4um pixel in place of the 8.2um the 8MP sensor will increase to an impressive 13MP. Personally I think this may be more likely as it would be a real step up and the 1D2 is an older camera than the 1Ds2. The same holds true for a Digic upgrade, and as this is the PJ/nature camera it will be key to ensure no performance is lost.

Mike

Added in data showing the size of the pixels for each release

Just curious as to whether most people really need more than 8mp?

I don't care about whether it is a 1.6 or 1.25 crop factor but if it is the 1.6 - put a better viewfinder in it!! Give me a viefinder as far from the 350 as you can. Lose the dial and stream the ergonomics of the prosumer canons so that they are closer to the 1 series. It is too much of a mental jump from one to the other way of working the controls. I will be happy then but if I want everything and the cake too I would want eye control as well. Then there would be no reason left to buy a newer camera after that. I'm waiting

Personally, I'd like to see a much lighter 1 series FF body & lenses, something like a carbon 1ds mk2, and some new wide glass. 22mp will be fine, but without new glass whats the point?

Andy

Why upgrade at all? just kidding! My other camera is an A2 film body that I bought in 1997. Still works perfect when I want to use Velvia! I did have to replace the command dial though. So as I see it, I will still be using my 10D five years from now. Well, unless the shutter goes and the cost of replacing the shutter....no, wait, shutter replacement cost is still only a fraction of the money I am saving on film and processing...yup, I guess it is true, I am a mizer, and I'll still be using a 10D when the 50D is the cats assss.

In my opinion, Canon has to improve and add
1. The dynamic range of image sensor, this makes difference from slide and DSLR.
2. Brightness and shown information of 20D's viewfinder
3. The wifi 802.11g
4. Memory Buffer size for raw
5. Writing speed for flash memory
6. Wireless remote release
7. Wireless control of external flash same as KM 7Hi

Now, the my 20D works great, if Canon does the above modification, I'll put it on my purchase list.

Just my two cents.

On the prosumer end, anything more than 6mp seems to be overkill. Yes, an 8mp sensor will look better than a 6mp on a 20x30" print, but the vast majority of prosumers will never go over 11x14".

Here are the things my wife would ask for as an entry level Dslr user, who is now using a Rebel XT:

1) Higher ISO's to take indoor/evening shots of children. Iso 6400 or 12800 will be great when the sensors can do it.
2) A moderate price 18-125 or 18-200 lens (the Tameron and Sigma don't count because they focus much slower and less accurate than the Canon lenses).
3) Fast focus that is accurate.
4) An even more intelligent light metering system in full auto.

More pixels if you want (I don’t think I need them, but I don’t mind).
Every better low light/high ISO performance.
Better low light focusing.
Improved dynamic range.
ISO in the viewfinder.
Auto-ISO mode. I set the minimum shutter speed, preferred aperture and a limit on ISO. The camera sets the ISO to meet my preferred aperture. When my ISO limit is reached, aperture is sacrificed to maintain shutter speed.

Now moving a bit out of the box….
How about the ability to alter ISO pixel by pixel?
ISO set to 100. The sky is too bright, the ground is too dark. Exposure is correct to the highlights and, based on a pixel by pixel histogram, the system alters the ISO of the dark areas to achieve correct exposure. I get to set a limit of 1 or 2 stops for this action.
This is over simplified. A reasonably good algorithm would be necessary to prevent very flat images.
Call it an electronic neutral density filter.

Being able to vary ISO from shot to shot is one of the most powerful features of digital. Designs should take better advantage of this fact.

smaniscalco, that's an awesome idea about ISO's. Wish I thought of that . You should try selling your idea to the mfg's :-)

I have a long history of brilliant ideas...that someone else already thought of.
It seems likely that Canon has thought of this already. It may be too costly to implement, or thought to be of too little interest to the customer, or available in the next model.
Time will tell.

The Canon patent I referred to above is basically a ND LCD on top of the sensor, so each pixel in ther LCD can be progressively darkened once enough photons have been collected - the more photons collected, the darker it the LCD pixel gets and that sensor pixel doesn't blow out. If not many photons are arriving, it doesn't get dark so shadows continue to accumulate valuable photons. The density of the LCD is controlled on a per pixel basis and the "tone curve" for this would be applied in reverse by software later to re-expand the dynamic range. Getting each LCD pixel to have a the same and reproducible tone curve is required for this to work but may be tough in practice....

Andy