Canon EOS 60D Pictures?

[Rocky]

I was hoping Canon might do the same with the 60D. After all, 2x18MP APS-C cameras are already available from Canon, therefore diversifying its commercial offer could have proven a successful move.

I do think that having the highest pixel-count in the APS-C market actually is a very successful move (from an economic perspective).

By the way, lower pixel count does not mean lower noise.
Most people compare noise levels at 100% view which does not reflect reality - it's the whole image that counts.
After all, the amount of light that hits the sensor is only depending on total sensor surface area (excluding microlens gaps of course) and not on pixel size.

 
The amount of light that hits each pixel will only depends on the F stop and the shutter speed.It does not depends on the total sensor size. (The same analogy as for given the same scenery, both the 6X6 and 35mm camera will have the same setting, while then 6X6 are having 4 times the area of a 35mm frame). As for inherit noise per pixel, the smaller the pixel, the more the noise, due to less amount of photon that hits the pixel, while the electronic noise is the same. So it sill have a lower sinal to noise ratio. That means more noise per pixel. The auther also mentions in a later posting that down size a 18MP picture to 10 MP. The resulting picture is as good as a picture taken with a 10 Mp pixel camera. I have a serious doubt about that. Anytime you transfrorm a picture, there are alway a small amount of uncertainty during the transformation, and result will be deteriated. Even the transformed picture is as good as tyhe 10 Mp picture without trnsformation, why bother with the transformation, why not just use a 10 Mp camera to start with.

[dilbert]

I was hoping Canon might do the same with the 60D. After all, 2x18MP APS-C cameras are already available from Canon, therefore diversifying its commercial offer could have proven a successful move.

I do think that having the highest pixel-count in the APS-C market actually is a very successful move (from an economic perspective).

By the way, lower pixel count does not mean lower noise.
Most people compare noise levels at 100% view which does not reflect reality - it's the whole image that counts.
After all, the amount of light that hits the sensor is only depending on total sensor surface area (excluding microlens gaps of course) and not on pixel size.

Wrong. This idea doesn't take into account the inherent noise in the technology.

  http://www.clarkvision.com/imagedetail/does.pixel.size.matter/#Depth_of_Field_Myth

People like to point to photos that show the 7D doing just fine, BUT that is in well lit scenes.  When the light begins to drop -- when some of us photographers like to shoot or are forced to shoot -- the 7D's small pixels are a problem. There is no substitute for larger pixels, period, if you are looking for better dynamic range and reduced noise.  In the link above he explores the ideal pixel size (too few hurts images and too small hurts).  The 7D and 50D crossed the line for the APS-C.  I think a 13-14 megapixel APS-C is the sweet spot.  Nikon, if the rumors are true, is being dragged into this pixel war that is not helping us folks.

The only good scientific analysis I've seen on this issue is from DXO-Labs, here:


http://www.dxomark.com/index.php/en/Our-publications/DxOMark-reviews/DxOMark-review-for-advanced-compacts

The one-line summary is this:
The Canon S90, with a small sensor, gets better per pixel sensor performance than the full-frame Nikon D3X.

DXOLabs are professionals in producing hardware and software to measure digital camera sensor performance. I think I trust their conclusions and results more than the other link you've included above.

[Michal]

Finally, something that is real and believable.  I’d rather see a post like this once every 3 months than pure fantasies and fanboy speculations every week.

[Deeksie]

"The Canon S90, with a small sensor, gets better per pixel sensor performance than the full-frame Nikon D3X."

Ah, you quoted it wrong. They said the S10 had better performance per unit of sensor AREA than the D3X, not per pixel. That is a huge difference!

Maybe the better technology in the smaller sensors is much harder to replicate on a larger chip?!

I used "bang out mate" before...re-reading I realise how offensive that could sound, sorry if you read it and felt offended!

[Gcon]

Is that a Direct Print button I see down below the thumb?   *awaits direct print jokes*

[dilbert]

"The Canon S90, with a small sensor, gets better per pixel sensor performance than the full-frame Nikon D3X."

Ah, you quoted it wrong. They said the S10 had better performance per unit of sensor AREA than the D3X, not per pixel. That is a huge difference!

Maybe the better technology in the smaller sensors is much harder to replicate on a larger chip?!

I used "bang out mate" before...re-reading I realise how offensive that could sound, sorry if you read it and felt offended!

A full frame sensor at S90 pixel density would be almost 200MP.

There are other limits to digital camera engineering besides the sensor. For example the CPU, buffers, etc.
How big is a 200MP RAW file going to be? At 1MB/MP, that's 200MB per picture or 5 pics/GB, 32GB = 150 RAW files.
How quickly can the CPU process it? (current cameras are maxing out below 150MP/sec.)

There's also the manufacturing of the sensor and getting the error rates per pixel down low enough so that they don't throw away too many sensors due to flaws.

[Rocky]

Wrong. This idea doesn't take into account the inherent noise in the technology.

Right, not wrong.

I think both my physics degree and 10 years of experience in semiconductor production do help me having a clear understanding what goes on in a solid-state imaging sensor and signal processing in general.

Resizing a 18MP image to 10 MP using a decent algorithm is not very much different to using a 10MP sensor with larger pixels in terms of signal processing. After all it's always about integrating photon impact over time and position - it doesn't matter so much if it's done in the hard or the software.

Still Wrong.
With a physic degree, you should have a good understand of clasiccal optics and wave optics. Then you should know that the 18MP is diffraction limited at F6.7. Any opening smaller than that will cause unsharpness of the picture. Your previous comment says the amount of light that hit the pixel  is only depends on the overall size of the sensor not the size of pixel is also wrong. Please go back to Optics 101. There is a big difference of "integrating photon overtime and position" between hardwared and software. If you integrating by soft ware, you are integrating electronics noise from the circuitry of eachy pixel. If you have a larger pixel, you will have more poton and the elcectrical noise from only ONE pixel. Therefore larger pixel will have less noise. Semiconductor production does not make you a noise expert.
You do not need to show your credential to convince people that you are right, if you have a valid reason.

[unfocused]

When Nikon releases their 16MP D95 and Canon releases the 18MP 60D, would it be possible to stop all the insanity over megapixels?

Getting real tired of the debate.

Alternatively, maybe we could create a special Megapixel War thread and let the two sides fight to the death.

I shot film for most of my life. If you used Tri-X you got some grain. If you pushed it to 1600 you got more grain. If it bothered you, you went to a larger format. It seems like people want to push their digital cameras to 6400 and think they shouldn't get noise. I'm just so incredibly amazed at the quality of the shots that come out of my 7D that I can't figure out what the issues are.

[Fadhillah]

THEY ARE FAKE PHOTOS!

The photos are very, very, obvious photoshops, and I cant believe this thread has gone on this long with no one noticing the obvious signs (or maybe I missed the post were someone noticed?)

1st, photo #1, notice the right side of the camera, look at the top of his hand and follow the side of the camera down, and you will see very easily that the lines dont match.

2nd, notice the spacing of the buttons on the right of the screen in the first photo, now look at the second photo, and you will notice the spacing is way off again between photos

I could list a bunch of other minute differences, but those two should be enough to convince anyone the photos are fake

of course the spacing is different because it is taken from a different angle.

[roadracer]

of course the spacing is different because it is taken from a different angle.

ah, never mind, I was thinking it was a obvious photoshop, but I dont feel like debating everyone about it  got better things to do

[Grummbeerbauer]

Seems like Canon wants to stick with the main power switch under the mode dial on APS-C cameras like introduced with the 7D. To be honest, I think that's a totally stupid place. Every single digital Canon camera I owned so far could be conveniently switched on and off either with the thumb or the index finger of the right hand, while holding the camera in the same way while shooting. With the power switch on the "new position" you definitely need your left hand for switching on and off. To me, that's a step-down in ergonomics.

I have to agree on this one. I had (well, still have...) a 450D before I got the 7D, and the 7D forces you to use two hands much more often. When I use the 450D with a handstrap, I can do almost everything (aside from entering the menu), including mode dial changes, with one hand only. The 7D requires two hands for on/off as you mentioned and for changing the shooting mode.

I overall still like the 7D better, though. 

[Rocky]

When Nikon releases their 16MP D95 and Canon releases the 18MP 60D, would it be possible to stop all the insanity over megapixels?

Getting real tired of the debate.

Alternatively, maybe we could create a special Megapixel War thread and let the two sides fight to the death.

I shot film for most of my life. If you used Tri-X you got some grain. If you pushed it to 1600 you got more grain. If it bothered you, you went to a larger format. It seems like people want to push their digital cameras to 6400 and think they shouldn't get noise. I'm just so incredibly amazed at the quality of the shots that come out of my 7D that I can't figure out what the issues are.

My issues on the 18MP APS-C camera are: 1. The sensor has out resolved the bset of the prime lens.  2.Diffraction limited at f6.7.  3.Noise at higher ISO (400 nad  UP)is even worst than 40D.  4.Built in noise reduction further reduced the sharpness og the picture.  5. Dynamic range is 1 stop less than the 40D.
If you are happy with your 7D. That is great. I have been waiting for a replacement for my 20D for a long time. When the 50D come out, It is disappointing. Then comes the 7D. It is better than the 50D but It is 18MP. So I gave up and brought a used 40D with only 7K shutter activation.

[afrank99]

Wrong. This idea doesn't take into account the inherent noise in the technology.

Right, not wrong.

I think both my physics degree and 10 years of experience in semiconductor production do help me having a clear understanding what goes on in a solid-state imaging sensor and signal processing in general.

Resizing a 18MP image to 10 MP using a decent algorithm is not very much different to using a 10MP sensor with larger pixels in terms of signal processing. After all it's always about integrating photon impact over time and position - it doesn't matter so much if it's done in the hard or the software.

Well you have me beat in credentials. But I can't understand how this can be true. Would a 40mp APS-C then not be any worse for noise and even better if down sampled?  I think we would both agree not.

Why not?
Doubling the pixel size (while keeping sensor area equal) means that every pixel captures the signal from two former smaller pixels (averaging the two). Just like pixel binning does. And just like downsampling does.

There may be effects in reality that limit this procedure, but in the end, there won't be much difference between a low pixel count sensor and a downsampled image.

Imagine taking the same photograph 100 times @ ISO 100. Then use the images to average each pixel.
What do you get? A super-clean image like it was taken with ISO 5 or something (don't know what the exact number would be).
Why?
Because the signal always stays the same, but the noise is random, and different for each image, so it will average to zero. A simple way to increase SNR.

Downsampling an image is the same using the area as averaging variable instead of time.

[wuschba]

A finally...

Did someone say "touchscreen"? That would be awesome. Not that it helps much, but it's just great to show off ;-)

[afrank99]

With a physic degree, you should have a good understand of clasiccal optics and wave optics. Then you should know that the 18MP is diffraction limited at F6.7. Any opening smaller than that will cause unsharpness of the picture.

Yes, of course, but so what?
Of course at some point more sensor resolution is not usable for details anymore, but this is not what we are talking about right now.
The image the optics throw at the sensor is only depending on the optics, and not the sensor.

Your previous comment says the amount of light that hit the pixel  is only depends on the overall size of the sensor not the size of pixel is also wrong. Please go back to Optics 101. There is a big difference of "integrating photon overtime and position" between hardwared and software. If you integrating by soft ware, you are integrating electronics noise from the circuitry of eachy pixel. If you have a larger pixel, you will have more poton and the elcectrical noise from only ONE pixel. Therefore larger pixel will have less noise.

You need to understand the difference between SIGNAL and NOISE.
Integrating noise means eliminating it.
Noise is random. Integrating lots of lots of noise makes it disappear.

Let me give you an example:
Let's say we have 1 Million 50% gray pixels. Noise introduces some pixels that are brighter, and some that are darker than 50% gray. The too-bright pixels will be roughly the same number than the too-dark pixels.
Downsampling the image means building an average - and the average of this noise is ZERO (well, almost).

Semiconductor production does not make you a noise expert.
You do not need to show your credential to convince people that you are right, if you have a valid reason.

There are so many people around not knowing what they are talking about, I just tried to separate from them.
Sorry for this.