Discussion in 'News - Discussion' started by CSBC, Sep 17, 2008.
Yeah I don't get that either.
I think that that's the point. It is just a camera, it's not an essential, people have to want to buy it. Reading the spec do you want to buy it? It's a personal decision and at the mo I don't want to buy it.
It's not the end of the world. It's just a camera and something to talk about.
Well, I've now downloaded both samples from the Canon website. They both open to >60MB in Photoshop so the stock photographers should be pleased. I know Canon are offering their 24-105 zoom with it, but I wish they'd used a prime lens for their shot of the glacier. (Check the lower LHS.) Interestingly, Breezebrowser displays the portrait correctly, but Photoshop displays it in landscape mode, as did Firefox. The portrait was taken with the EF 50 f1.2L at f2, and the depth of field is quite slim.
As a long time 5D owner and user, am I impressed? Not sure, as the increase in pixel space isn't that enormous.
It is becoming clear that pixels alone do not a super camera make. I suspect that for many generalist photographers the lower pixel count of the D3/D700 will be more suitable. There will be specialists who take the opposite view and want all the pixels they can get and will use flash when the light falls. Current technology seems to favour lower pixel counts (bigger pixels) for sensitivity and low noise so the new Canon is likely to be noisier earlier, not necessarily a problem.
Not that I have any intention of buying one.
I definitely want one! And am glad I waited rather than plump for a 5D, I will wait until the price comes down a fair bit or some stonking cashback offers materialise! Perhaps Canon will give me one for a couple of months to try out? I will need a new laptop too with the much larger files as opposed to my 30D files!
So you don't agree with what Damien wrote the other week?
I can't remember what DD wrote But technology is advancing at a great pace in this area, so why can't Canon develop a low noise 20+ Mega pixies sensor comparable or better than that of the D3/D700/D300? Lets wait and see, as many have said the proof of the pudding........ I for one will be watching out for the reviews.
What he wrote was that after using the 1Ds Mk III for some time, he had great difficulty adjusting back to the "low" resolutions of other cameras.
That camera isn't exactly bad for noise, but I suspect this one will be significantly better, as it appears Canon have on one level improved physical aspects of the sensor, and also have a new processor. We'll see, of course. I'm not sure I really do want to know how well it does after all - for now, I'm more than happy with the 5D.
There's three components to this.
1. Quantum noise - the number of photons detected by each photosite has to be a whole number, the smaller the photosite is the smaller the number, with 5 micron photosites the number is in the tens so the "random" (Poisson process) arrival of photons has a significant effect. This gets worse with smaller photosites and is an inevitable consequence of physical laws, not amenable to engineering except by making the photosites bigger - which means less of them, or a larger sensor.
2. Electronic noise - there is some scope for improvement here but the manufacturers are tending to go the other way, if you read the sensor more slowly you can get a big decrease in electronic readout noise. But they read the sensor faster than is desireable so as to get higher frame rates.
3. Image processing - yes, with "improvements" in image processing the apparent noise can be reduced, but always at the expense of hiding real detail in the image and with the introduction of artifacts from the processing.
Sensor technology hasn't changed much in the last 5 years. The processing has got more "sophisticated", but really there isn't much point in capturing more data than there is detail in the raw image captured by the sensor (the extra data has to be artifact) nor than can be resolved by the lens (and this is getting to be a very real issue).
The people who really know about low noise imaging are the astronomers. The top grade sensors sold for astronomical imaging have a photosite pitch of around 20 microns i.e. 1800x1200 or "2 megapixels" in full-frame terms (astronomical sensors are actually square format). This should be evidence enough that 20 megapixels on a full-frame camera is far too many.
Interesting. The review in Outdoor Photography some months back said something along the lines of there being visible noise as low as ISO400 in certain out-of-focus backgrounds (I'm in the library now, so don't have the mag with me). I must admit I am leaning towards Michael Reichmann's comment in his first article on the Nikon D3, in which he suggests that for the vast bulk of people a camera of approximately 12 megapixel is perfectly sufficient, as one can print up to A3 without the need to ress up, which, he claimed, was as big as most people would need.
Larger pixels do mean less noise, especially when combined with state-of-the-art processing, and I must admit that for more than that magic 12mp I'd tend to lean towards a medium format solution. These are still hideously expensive, although Hasselblad's current promotion of their 31 megapixel H3DII does open a narrow window (until end of the year I believe) for people with a bit of spare cash (bad timing for now, sadly, unless you're a dodgy hedge fund trader! /forums/images/graemlins/tongue.gif). Still, a mere 31 pixies should more than suffice for Joe/Jo Amateur. The 60-odd now coming into view is just mind boggling, and may well need a further lens revision to cope with it.
My local astronomer friend [Peter Birtwhistle, with over 100 discoveries to his name] tells me that he cools his sensor down to -40C.
The sensor is 1024 by 1024 pixels, but this is more resolution than he needs, so his processing algorithm looks at neighboring 2x2 pixel groups and takes the resolution down to a 'clean' 512 by 512 (256k pixels)
Does that assume printing at 300dpi? is there a visible difference in an A3 print done at 300dpi and one done at 200dpi?
Are the differences in resolution between 5/10/15/20/25mp cameras only going to be visible when viewed at 100%, is there going to be a visible difference at print sizes of A4 - A3. Surely of you print at anything smaller than the native image size/300dpi you're throwing pixels away, printing larger requires you to interpolate or reduce the dpi at which you wish to print. Is anything larger than 12mp only an actual benefit if.
a. You print larger than A3
b. You need to crop radically and still print at A3
Does a 20mp image contain more detail than a 10mp image if both are printed at A4?
Yes, cooling helps reduce electronic noise - but does nothing about quantum noise.
2x2 binning is commonly done but does not eliminate noise - it just reduces it.
Canon made a videoabout it, explaining some stuff as well.
From memory (check the D3 feature at LL for confirmation or otherwise) it was at 180dpi.
A Nikon D3 file will print A3 at 260dpi
thanks, I'll have a look.
Thanks Fen, was going to check it in Photoshop but couldn't remember the dimensions for A3.
You have a D3 don't you?
I sure do
A3 is 297mm x 420mm, or 11.6" x 16.5"
I'm not even slightly jealous /forums/images/graemlins/tongue.gif (ok, maybe a little bit )
Separate names with a comma.