Home

Please support mflenses.com if you need any graphic related work order it from us, click on above banner to order!
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

Canon EOS-1D Mark IV announced
Post new topic   Reply to topic
Home   Official News and Announcements
View previous topic :: View next topic  


PostPosted: Tue Oct 20, 2009 8:26 am    Post subject: Canon EOS-1D Mark IV announced Reply with quote
http://www.engadget.com/2009/10/20/canon-eos-1d-mark-iv-announced-16-1-megapixels-45-point-autofo/

    * 16.1 megapixel APS-H CMOS sensor
    * ISO 100 to 12,800
    * 45-point AF (39 cross-type)
    * dual Digic 4 processors
    * 10 fps
    * 1080p HD video
    * optional wifi
    * USD 5000


PostPosted: Tue Oct 20, 2009 9:13 am    Post subject: Reply with quote
With ISO Expansion, this camera will give ISO 50 - 102400 (no this is not a typo). Not sure how usable is the ISO 102400 though.


PostPosted: Tue Oct 20, 2009 9:43 am    Post subject: Reply with quote
mistake, please erase thanx

Last edited by hexi on Tue Oct 20, 2009 4:47 pm; edited 1 time in total


PostPosted: Tue Oct 20, 2009 9:44 am    Post subject: Reply with quote
In Germany the recommended retail price will be approx. 4.700,00 € (500 € more than the first 1DMKIII retail price)

Will wait and see what will happen, especially with the AF.

Wink


PostPosted: Tue Oct 20, 2009 10:51 am    Post subject: Reply with quote
I don't care for Video...
Nikon makes advertising for 3Ds with this only argument : video.
I'm a fan of photo, not a fan of video... Confused

Else, this new 1D MkIV is a APS-H format. What are the pros and cons of the APS-H (coeff. x1.3 / FF) compared to APS-C (coeff x1.6 / FF) and FF ?


PostPosted: Tue Oct 20, 2009 11:41 am    Post subject: Reply with quote
The target market for the 1D line is news and sports photographers. Fast AF, rapid fire and a big buffer are important. The slightly smaller sensor probably helps with those features.


PostPosted: Tue Oct 20, 2009 1:17 pm    Post subject: Reply with quote
PaulC wrote:
The target market for the 1D line is news and sports photographers. Fast AF, rapid fire and a big buffer are important. The slightly smaller sensor probably helps with those features.


Yes, and Canon choosed to sacrifice megapixel size for a better signal to noise ratio.
(note that the camera has less MP than the 7D)
The decision of course fits with the design of providing a camera that performs well at high iso, instead of one that produces a very large size but useable only until ISO 800 or so.

-


PostPosted: Tue Oct 20, 2009 3:22 pm    Post subject: Reply with quote
Nice pro camera.

Nice to see the MP war truly is over and that the high ISO war is in full swing between the big guys.


PostPosted: Tue Oct 20, 2009 3:30 pm    Post subject: Reply with quote
A little correction : i misread and i thought it was about the Nikon D3s !!
SO all my post above is about it Rolling Eyes


PostPosted: Tue Oct 20, 2009 4:18 pm    Post subject: Reply with quote
Sounds like an amazing camera. At that price, though, I won't be buying one anytime soon. My daughter needs a clean used car.

Honestly, I don't get the whole APS-H thing anyway. I'm kinda surprised Canon's still supporting it. I can see having FF and APS-C cameras in a product line, but I don't get a line that has that mid-size one. Me, I wouldn't buy one of these cameras, even if I could afford it, because I'd rather spend a bit more (or less for the 5DII) and get an FF. So what am I missing? What's the attraction to APS-H?


PostPosted: Tue Oct 20, 2009 4:27 pm    Post subject: Reply with quote
This camera is marketed to the event shooters.
Sport, politics, public gatherings etc........
It's made for speed and for shooting far from the action.
The 1.33 crop gives a few advantages.
The crop factor adds effective focal length to lenses.
More importantly it increases the DOF at wide open (as well as stopped) for a given FL.
This helps quite dramatically in regards to getting things in focus.

Consider that for example. A 35mm lens now has the field of view of a 50 on this camera but, with the DOF of the 35.
Hitting accurate focus on emerging events is much easier.

I've been using a 1.33 crop camera for just under a year now.
I really do find that in many cases the crop factor is a blessing.

Hope that helps. Maybe others will add more.
Cheers


PostPosted: Tue Oct 20, 2009 4:32 pm    Post subject: Reply with quote
Okay I get all that, but then let me ask, what is it about the 1.33 crop factor that's more advantageous than the 1.6 crop factor? Because all that you mention in favor of the 1.33 can be said about the 1.6. Except, I suppose, for the fact that a 1.6 cf camera tends to turn WA lenses into normals whereas with a 1.33, you still have some WA left with 28mm and shorter?


PostPosted: Tue Oct 20, 2009 4:35 pm    Post subject: Reply with quote
cooltouch wrote:
So what am I missing? What's the attraction to APS-H?


An APS-H camera at 16 MP is going to have a similar or perhaps even better signal/noise ratio than the 5D Mark II.
The smaller format, on equal image quality, has 2 significant advantages for sport shooters:

- easier handling of tele lenses (the crop allowing to use smaller lenses to obtain the same enlargement of larger lenses. This might not mean a thing for the birds shooters, but for people who shoots live events, and maybe has to move around, it is really important to have a lighter bag and smaller size lenses

- less image data means faster writing on the card which means higher frequency of continuous shooting: a crucial feature for pro sport photographers who rely on dense sequences of continuous shots to pick the best one for a magazine

At the same time, since most sport shooting ends up, at the largest, on magazine covers, and often, on low resolution newspapers, the sport photographers don't need the high resolution of a 20 MP camera.


PostPosted: Tue Oct 20, 2009 4:37 pm    Post subject: Reply with quote
cooltouch wrote:
Okay I get all that, but then let me ask, what is it about the 1.33 crop factor that's more advantageous than the 1.6 crop factor? Because all that you mention in favor of the 1.33 can be said about the 1.6. Except, I suppose, for the fact that a 1.6 cf camera tends to turn WA lenses into normals whereas with a 1.33, you still have some WA left with 28mm and shorter?


Yes you have it.

With the 1.33 you simply step up roughly one FL.

15=21
21=28
28=35 etc...........

As to 1.6 or even x2 (4/3) the DOF advantage is even greater although as you said you lose your wides almost completely.


PostPosted: Tue Oct 20, 2009 4:52 pm    Post subject: Reply with quote
cooltouch wrote:
SMe, I wouldn't buy one of these cameras, even if I could afford it, because I'd rather spend a bit more (or less for the 5DII) and get an FF. So what am I missing? What's the attraction to APS-H?


I don't get the APS-H either… I'd rather have a full frame and crop or downsample if necessary. Even the lower megapixel count for supposedly better performance is something of a marketing trick; if they'd kept the 20+ megapixel full frame sensor I'm fairly certain the images could be downsampled to 16 megapixels for even better high ISO performance, but of course it would cost more to make such a sensor. So they borrow a trick from Nikon's D700 marketing and say that a lower pixel count is somehow better by itself. Personally I'd prefer to have the option of downsampling for high ISO performance and larger images for low ISO. And full frame, of course.

(Too bad major reviewers like dpreview will buy up this crap and undoubtedly compare high ISO performance by showing side-by-side 100% crops from cameras with different megapixel counts, like they always do, instead of resizing them for a fair and sensible comparison that would actually have practical significance…)

Umm, sorry, rant over. =)


PostPosted: Tue Oct 20, 2009 4:59 pm    Post subject: Reply with quote
Arkku wrote:
cooltouch wrote:
SMe, I wouldn't buy one of these cameras, even if I could afford it, because I'd rather spend a bit more (or less for the 5DII) and get an FF. So what am I missing? What's the attraction to APS-H?


I don't get the APS-H either… I'd rather have a full frame and crop or downsample if necessary. Even the lower megapixel count for supposedly better performance is something of a marketing trick; if they'd kept the 20+ megapixel full frame sensor I'm fairly certain the images could be downsampled to 16 megapixels for even better high ISO performance, but of course it would cost more to make such a sensor. So they borrow a trick from Nikon's D700 marketing and say that a lower pixel count is somehow better by itself. Personally I'd prefer to have the option of downsampling for high ISO performance and larger images for low ISO. And full frame, of course.

(Too bad major reviewers like dpreview will buy up this crap and undoubtedly compare high ISO performance by showing side-by-side 100% crops from cameras with different megapixel counts, like they always do, instead of resizing them for a fair and sensible comparison that would actually have practical significance…)

Umm, sorry, rant over. =)


Down sampling to improve High ISO performance is a farce. It just does not work that way.
If you disagree. Please explain


PostPosted: Tue Oct 20, 2009 5:58 pm    Post subject: Reply with quote
Looks to me as though Canon has done a lot of work to protect its leadership in this very lucrative pro field. Sportphotogs will probably go for it big time. Those huge tele lenses will be even more tele and the shooting rate will enable them to capture basketballs in mid-air!


patrickh


PostPosted: Wed Oct 21, 2009 2:41 am    Post subject: Reply with quote
Viewed a movie done at iso 6400 with this camera by Vincent Laforet
Sorry it is no longer available but, likely will be shortly.
My guess is that Canon pushed a password on him.
Check back after the official announcement.
It's amazing what this camera is capable of. Shocked

http://vincentlaforet.smugmug.com/Laforet-Videos/Nocturne-Canon-1DMKIV-Video/10024122_sqhwE#686345820_EeDCa-XL-LB


PostPosted: Wed Oct 21, 2009 6:05 am    Post subject: Reply with quote
Out of my price range so not an option for me.

I was hoping they would leave the video mode off. Not sure I want a camera that records video.


PostPosted: Wed Oct 21, 2009 9:40 am    Post subject: Reply with quote
First higher MP issue, now who's got the best ISO ... its a real joke, who will actually use the 100 000 anyway ? Besides for this price they could have put a 24x36 sensor and get rid of the video Mad


PostPosted: Wed Oct 21, 2009 10:47 am    Post subject: Reply with quote
F16SUNSHINE wrote:
Down sampling to improve High ISO performance is a farce. It just does not work that way.
If you disagree. Please explain


When downsampling with any intelligent algorithm, the resulting new pixels (i.e. the pixels of the smaller, downsampled image) are formed by combining values from multiple original pixels. This means that if each one of these original pixels has some amount of random noise, the randomness will "average out" in the downsampling process.

As a simple theoretical example, consider a very boring image of nothing but a grey card, exposed so that the image should ideally be a perfect mid-grey, i.e. every single pixel should have the same value. Let's say this image is 1000×1000 pixels, i.e. 1 megapixel in size. If every color component of every pixel has random noise such that it deviates from its "true" value (i.e. mid-grey) by, say, up to 5%, a 100% crop of this image will look very noisy. However, if we downsample this image to 1×1 size by the trivial method of taking the average of all the original one million pixels, that one pixel will almost certainly* be nearly perfect mid-grey, because all that random noise is averages out. And yes, this is an extreme example, but as a mathematically-oriented person (computer scientist by profession) I find that this kind of example intuitively shows how it works regardless of scale, as long as the downsampling method is such that it combines values from multiple pixels (i.e. basically any common algorithm except "nearest neighbour").

If you don't believe this example (or that it applies in a smaller scale to less extreme amounts of downsampling), you can easily try this in Photoshop: create a new image of some large dimensions (i.e. a megapixel or more), fill it with your chosen colour (e.g. 50% grey), and add gaussian noise to it in the desired amount (about 5% already looks very noisy in such a detail-free image). Downsample the image by varying amounts using some sensible method (like bicubic), and you'll find that the more you downsample, the less noisy your image will appear. Why? Because even though all that noise goes into the downsampled image, you are improving the signal to noise ratio of each "new" pixel by putting more signal (multiple original pixels) into it.


This is not really different from why having larger pixels in the sensor given less apparent noise per pixel; more photons go into that larger area, i.e. you have more of the "signal" per pixel even if every pixel by itself has some given amount of random noise. And that's also why you can get similar improvements by downsampling a larger image; you are effectively making your pixels cover a larger area.

(This is also similar to how stacking multiple exposures of the same subject works, e.g. in astrophotography, only you are not taking the additional signal from corresponding pixels in entirely new images but from adjacent pixels that end up representing the same area in the smaller, downsampled image.)

Of course individual sensors will have different "per pixel" noise characteristics regardless of how many pixels there are total, and it's probably easier to make a high-performance sensor with less pixels… But, it makes absolutely no sense to compare noise performance at 100% crops; the comparison must be made at equal size (e.g. either with the images resized to the same pixel count, or printed at the same size—the latter seems the obvious way of comparing things on film, so I find it difficult to understand why people think digital should be compared at unequal sizes).


Another the great thing about downsampling with colour-filter array sensors is that it also helps get rid of demosaicking artefacts (effectively getting benefits similar to a Foveon sensor). More pixels also give more options for further digital filtering. Note that I'm not saying that the megapixel wars should continue, we are pretty near the maximum useful pixel counts for these sensor sizes—the sensor should grow next. But there's no reason for the customer to specifically want less pixels if they are available at the given price point (e.g. check out the price of a Sony A850 with 24.6 Mpix full frame, based on the same Sony-made sensor as Nikon D3X).


* I say "almost certainly" above because in theory it's possible that the randomness for a particular instance "conspires against us" and happens to lead the result in a particular direction. (Similar to how it's possible in theory to flip a coin a million times and get heads every time with a fair coin, but if it happened to you you'd probably assume that the coin was rigged somehow.) This is not something that you'd need to worry about in practice with a digital camera… =)


PostPosted: Wed Oct 21, 2009 4:48 pm    Post subject: Reply with quote
Quote:
* I say "almost certainly" above because in theory it's possible that the randomness for a particular instance "conspires against us" and happens to lead the result in a particular direction. (Similar to how it's possible in theory to flip a coin a million times and get heads every time with a fair coin, but if it happened to you you'd probably assume that the coin was rigged somehow.) This is not something that you'd need to worry about in practice with a digital camera… =)


Thank you Arkku for the response. I can see that you have a point however, I believe it to be highly theoretical.

In the real world as your final paragraph states more or less. We don't actually gain ISO by downsampling.

If I have a noisy 1600iso sample at 24mp. Dumping it down to 16MP is not really going to help compared to having a larger pixel doing the work in the first place.

At a Pixel peeping level yes, maybe we see a difference. In a print.......?????
I'll take the larger pixelled cleaner sensor.

Cheers Mate


PostPosted: Wed Oct 21, 2009 5:39 pm    Post subject: Reply with quote
F16SUNSHINE wrote:

In the real world as your final paragraph states more or less. We don't actually gain ISO by downsampling.


I never said we gain ISO (sensitivity) by downsampling, but we gain better noise performance at a given ISO by downsampling, i.e. a downsampled image has less apparent noise than the original larger image when both are viewed at 100%, and the noise becomes less and less apparent the more the image is downsampled. This is equivalent to having larger pixels to begin with, because downsampling (using any sensible method) effectively merges the smaller pixels into larger ones.

Note that if you view both images at equal size (i.e. not at the same number of pixels per inch), such as is the case with printing to a given paper size, you are essentially doing the same thing as resampling one image to match the size of the other, and my point all along has been that this is the proper way to compare noise!


F16SUNSHINE wrote:

At a Pixel peeping level yes, maybe we see a difference. In a print.......?????
I'll take the larger pixelled cleaner sensor.


Ah, that's exactly right—the noise you see in the 100% crop comparisons is nothing but pixel peeping, and it has little to do with performance in print. To compare the actual noise performance one either has to make both images equal size, either by printing or by resizing digitally. The end result is the same. You cannot say that the sensor with fewer (large) pixels is cleaner in print at a given size by looking at a 100% crops side by side; the opposite may well be true if you print both images and make a fair comparison. So again: if you wish to compare images digitally, you need to make them of equal size!

In other words, comparing 100% crops of images with different pixel counts is like comparing prints of different size. Very few people would consider that a fair comparison, but for some reason many of the same people seem to think it's fair to assess the noise performance of a sensors at 100% which is comparing different sizes… Certain camera manufacturers like to encourage this (and unfortunately certain review sites like dpreview play along), which leads to the myth that larger pixels automatically give better noise performance at a given size, when all they are saying that they give better noise performance in an unfair comparison at different sizes.

Last edited by Arkku on Wed Oct 21, 2009 8:52 pm; edited 1 time in total


PostPosted: Wed Oct 21, 2009 7:37 pm    Post subject: Reply with quote
F16, I think you're wrong about DoF. It can't be true that taking a photo with a certain lens and then chopping the sides off (which is what a crop sensor does) changes the DoF in that picture.

I can accept it creates the illusion of doing that, but only because you have decided (or been forced) to use a wider lens and so end up with a smaller image.

You could have shot the same image with a full frame camera without filling the frame and got the same DoF.

What will change DoF is the pixel density if you view two images shot with different cameras and then view them at a fixed number of pixels per inch, then the one with denser pixels will produce a bigger final image and a smaller DoF from the same lens.

It follows that the increased number of pixels in this camera, compared with its predecessors, will reduce the perceived advantage in DoF.

Of course, this is all a bit arcane and what matters is what people find works best for them.


PostPosted: Thu Oct 22, 2009 4:57 pm    Post subject: Reply with quote
The nice thing about APS-H is that it gives you a little extra reach on the telephoto end with out totally screwing you on the wide-angle end. At the wide end, the 16-35 zoom gives a similar a angle of view to a 20 or 21mm on full frame. And some primes still work nice at 1.3x, while 1.6x is just too much change in FOV. (Still, I'd love to have a full frame camera.)

Another answer to this question can be found in one of Canon's white papers. It seems that a full-frame sensor is too big to burn on the silicon wafer all at once. It takes three precisely aligned exposures to create the sensor. The biggest sensor they can burn in one step is....yes, APS-H. So in theory it's easier and cheaper to make an APS-H camera rather than fullframe. Looking at the price of the Mark IV, I wonder how much this is really helping the consumer.


--Geoff
Post new topic   Reply to topic
Home   Official News and Announcements
Page 1 of 1