You can't have live view on a CCD sensor. And an EVIL camera requires live view which in turn requires a CMOS sensor.
Jerry_R wrote:
There may be kind of CCD live view, with slow refresh rate. It was discussed on Leica forum, something like this is in use in middle format people wrote.
And whoever said that is obviously full of beans too. I can point to 4 or 5 cameras with live-view which are CCD cameras. The Minolta A2 comes to mind right off the bat. It's live view is fairly close to my own GH1 even.
Ah, but Bifurcator, the A2 was a 2/3" camera! Of course you can pull an image from that small a sensor relatively fast, regardless of sensor technology and without overheating. We're talking about large sensor CCD cameras here, not small sensors. You'll find that there is not a single ASP-C or larger CCD camera that has live view in the normal sense (i.e with a reasonable refresh rate).
There are a couple of medium format cameras that support tethered "live view" with a refresh rate in the region of 1-2 fps, but that's it.
Assuming CCD development doesn't stop, the main issues with using CCD's for live view in a FF sensor will be solved, which are heat/noise, and read speed/refresh rate.
The live view in Olympus' CCD based 4/3's cameras were useable, just not as good as CMOS,
so I guess I'll rephrase my statement.
There is no reason CCD can't be used if they solve the heat and read speed issues.
I do agree with you denoir, that as it stands, CCD is unlikely to be used for a live view camera.
And the ratio of R&D $ going into CCD and CMOS is heavy on the CMOS side, so innovations to CCD's will be slow.
So by the time we actually see a FF EVIL these issues may be solved....
The live view in the IQ180 is painfully slow. The truth of the matter is that CMOS has leaped ahead of CCD in development R&D over the last number of years, so cost would really be the only reason Leica wouldn't use a Sony CMOS sensor.
Anyways, being someone who's been following Sony pretty closely for 5+ years, I think these pics look as real as any other Sony leak pics. I'm not thrilled with the ZA E 24mm size, but the NEX-7 looks pretty good.
douglasf13 wrote:
The truth of the matter is that CMOS has leaped ahead of CCD in development R&D over the last number of years, so cost would really be the only reason Leica wouldn't use a Sony CMOS sensor.
Image quality rather than cost is the primary reason - CCDs are more expensive than CMOS sensors. The M9 uses a CCD for the same reason digital medium format backs use it - very high low ISO image quality superior to what you can get from CMOS today.
You are right however that CMOS sensors unlike CCDs have undergone a massive improvement over the past few years. Originally CMOS was just a cheaper alternative but now they are superior in several ways - particularly when it comes to high ISO performance. I don't think CCDs have much of a future. AFIK the M9 is the only modern small format camera that has one.
In a way it's a pity that it will die out as it definitely has IQ advantages at low ISO, but I think it is inevitable.
Douglas, from the size of the mount I would say the 7 is not bigger than the original 3, and if so, the size difference should be very small. As for the Zeiss, it has a filter size starting with 4 so in the worst case, it would be 49mm, but probably smaller. I don't think it is bigger than the kit lens.
denoir wrote:
Ah, but Bifurcator, the A2 was a 2/3" camera! Of course you can pull an image from that small a sensor relatively fast, regardless of sensor technology and without overheating. We're talking about large sensor CCD cameras here, not small sensors. You'll find that there is not a single ASP-C or larger CCD camera that has live view in the normal sense (i.e with a reasonable refresh rate).
There are a couple of medium format cameras that support tethered "live view" with a refresh rate in the region of 1-2 fps, but that's it.
Ah, well in that case... All your beans belonged to me!
denoir wrote:
Ah, but Bifurcator, the A2 was a 2/3" camera! Of course you can pull an image from that small a sensor relatively fast, regardless of sensor technology and without overheating. We're talking about large sensor CCD cameras here, not small sensors. You'll find that there is not a single ASP-C or larger CCD camera that has live view in the normal sense (i.e with a reasonable refresh rate).
There are a couple of medium format cameras that support tethered "live view" with a refresh rate in the region of 1-2 fps, but that's it.
It is not how big the sensor is what matters for LiveView. It is the pixel count and what kind of sub sampling matters. There are not that many cameras out there which display full resolution from the sensor in LV mode. . There is something called skipping / binning in which one out of every 4/8 rows/columns is read out. It is not really refresh rate concerns that matter. An 8 MP camera can easily be refreshed at 30 fps but it will eat into battery power, and very high resolution images are not really required on 3" screens.
For example, I would not be very surprised if the lcd displayed image from a 5D2 is about 4-8 MP ( also the reason why critical focus needs higher magnification)
All this can be done with CCD sensors but the array architecture makes it harder if not undoable.
denoir wrote: douglasf13 wroteThe truth of the matter is that CMOS has leaped ahead of CCD in development R&D over the last number of years, so cost would really be the only reason Leica wouldn't use a Sony CMOS sensor.Image quality rather than cost is the primary reason - CCDs are more expensive than CMOS sensors. The M9 uses a CCD for the same reason digital medium format backs use it - very high low ISO image quality superior to what you can get from CMOS today.
You are right however that CMOS sensors unlike CCDs have undergone a massive improvement over the past few years. Originally CMOS was just a cheaper alternative but now they are superior in several ways - particularly when it comes to high ISO performance. I don't think CCDs have much of a future. AFIK the M9 is the only modern small format camera that has one.
In a way it's a pity that it will die out as it definitely has IQ advantages at low ISO, but I think it is inevitable. ...Show more →
From what I understand, the cost of CMOS is lower only in large production quantities, because current CMOS development costs are so much higher. Given Leica's relatively tiny sensor needs, I would not be surprised that a 135 Sony CMOS would cost them more per unit. Heck, even Sony still used CCD in their lower end DSLRs until a year or so ago.
I've owned plenty of both CCD and CMOS cameras, and, while I think a low ISO case could have been made for CCD 4 or 5 years ago, I don't think there's much of anything from CCD that can currently touch Sony's most recent CMOS...given that we're comparing the same sensor size, sensor toppings, lack of AA, etc. Even Irwin Puts would probably agree with that.
edwardkaraa wrote:
Douglas, from the size of the mount I would say the 7 is not bigger than the original 3, and if so, the size difference should be very small. As for the Zeiss, it has a filter size starting with 4 so in the worst case, it would be 49mm, but probably smaller. I don't think it is bigger than the kit lens.
I agree that the NEX-7 size isn't bad. It's the lens that I'm not thrilled with, but I also thought the kit zoom was too big, and I didn't buy that, either.
Ajay, regarding subsampling, that's correct - but it depends on the sensor in question. The 5DII for instance can pull 1080p video through the HDMI connection. The new Red EPIC for instance on the other hand can pull full 4000k footage at 120 fps.
Now, there are two big differences when it comes to live view and CCD vs CMOS (correct me if I'm wrong on this).
First a CMOS sensor can read pixel by pixel while the CCD has to offload the whole image at once (i.e rolling shutter vs global shutter) The second difference is that a CCD sensor uses much more energy and heats up much more. The problem with live view there is the same as for long exposures (something CCDs are terrible at) - even if you have the energy to spend, the sensor overheats and the noise levels shoot up.
douglasf13 wrote:
I've owned plenty of both CCD and CMOS cameras, and, while I think a low ISO case could have been made for CCD 4 or 5 years ago, I don't think there's much of anything from CCD that can currently touch Sony's most recent CMOS...given that we're comparing the same sensor size, sensor toppings, lack of AA, etc. Even Irwin Puts would probably agree with that.
Oh, I seriously doubt he would agree to any such thing. I doubt that anybody using a modern high end CCD would give it up lightly - me included. And I'd really love to have live view on the M10.
To be honest I have not seen anything particularly impressive from Sony CMOS sensors be it in my Leica X1 or my recent NEX-C3. On the contrary, I'm more partial to the output of Canon's CMOS sensors that now a few years old. But perhaps it's different for the higher end FF Sony sensors.
And on the other hand, nothing in small format comes even close to the output of the M9 in several respects. Digital medium format backs are of course as a rule superior.
Now as for an AA filter, there must be a good reason why CCD sensors so often come without it and CMOS sensors have it. This is pure conjecture and anecdotal evidence but I've used a 5DII with its AA filter removed and first of all the aliasing problems were staggering and second the image wasn't all that impressively crisp. So I'm guessing that there is something with a CMOS sensor that makes the use of an AA filter a necessity.
I think sensor wise the king of the hill is still the Nikon D3X (Sony sensor?), both in actual resolution, dynamic range and noise performance. Even though the M9 has a higher per pixel quality and produces more crisp images with high micro contrast, the D3X sensor is without a doubt superior. That doesn't automatically mean that it's superior in every way. If the M10 would have an equivalent sensor quite a bit would be gained but some would be lost as well.
Of course, I'm not comparing smaller format sensor to the M9. My simple point is that there are too many factors in the equation to single out any superiority of CCD over Sony CMOS. Sensor size? AA filter? CFA? Tuned for which native ISO? Joakim has certainly mentioned the relatively pedestrian qualities of the M9 sensor base.
I've owned digital backs, CCD aps-c cameras, CMOS 135 cameras, CMOS aps-c cameras...and I certainly can't claim that it's just the sensor architecture making the difference between them. You're actually one of the first people that I've ever seen mention that they prefer Canon CMOS output to Sony (taking only into account those that own both.) Maybe you just have weird taste? I'm kidding.
To be honest, I have yet to see anyhting particularly impressive from the M9 in the color department. In the contrary, it seems to require a lot of PP and the colors out of the camera are way off. I agree though about the D3X, with Sony sensor, and if I fault Sony for anything, it is putting their excellent lenses, especially the Zeiss, at disadvantage, by having the AA filter further away from the sensor. This makes the small dust particles less visible, but in the same time, destroys the detail at pixel level.
douglasf13 wrote:
I've owned digital backs, CCD aps-c cameras, CMOS 135 cameras, CMOS aps-c cameras...and I certainly can't claim that it's just the sensor architecture making the difference between them. You're actually one of the first people that I've ever seen mention that they prefer Canon CMOS output to Sony (taking only into account those that own both.) Maybe you just have weird taste? I'm kidding.
It's quite possible
From what I understand, the cost of CMOS is lower only in large production quantities, because current CMOS development costs are so much higher. Given Leica's relatively tiny sensor needs, I would not be surprised that a 135 Sony CMOS would cost them more per unit. Heck, even Sony still used CCD in their lower end DSLRs until a year or so ago.
Having Kodak develop a custom made sensor for them must have been far more expensive than using an existing production CMOS model. It would certainly be cheaper for them to buy a mass produced one - as they by the way did for the X1. It has a Sony ASP-C CMOS sensor.
There's more to it though. I'm sure the M9 has a CCD because the M8 had a CCD as they did not have to change much of the electronics and software. And the M8 has a CCD because when it was released there was a definite advantage in image quality. So although the Kodak CCD for the M9 is almost certainly more expensive than a mass produced Sony CMOS, changing the whole camera architecture to work with a CMOS might have been more expensive.
edwardkaraa wrote:
To be honest, I have yet to see anyhting particularly impressive from the M9 in the color department. In the contrary, it seems to require a lot of PP and the colors out of the camera are way off. I agree though about the D3X, with Sony sensor, and if I fault Sony for anything, it is putting their excellent lenses, especially the Zeiss, at disadvantage, by having the AA filter further away from the sensor. This makes the small dust particles less visible, but in the same time, destroys the detail at pixel level.
The M9 outputs very flat RAW files so you usually need to do a bit of contrast tweaking. As for color accuracy, the first profile that they released was quite a bit off but it has improved since. Otherwise it's a mixed bag. It can produce some marvelous deep blue colors - something I've only seen from Velvia slide film. At the same time the color separation of greens is so-so. I think the 5DII does a better job there. What I've seen from the C3 in the color department so far is substantially inferior, but I'm pretty sure the Sony RAW converter is to blame and I'll wait for ACR before making any definitive statements there.
These are just general impressions, I have not made any proper side by side comparisons.. Color tests between cameras tend to become impossibly complex as you need to take into consideration different RAW developers.
I am not happy with the colors of my girlfriend's Nex 3 either, but let's face it, Nex 3 is the lowest end and the C3 is even lower. I love the colors from my A900 and as mentioned above, my only gripe is the lack of biting sharpness at pixel level caused by the AA filter placed further away from the sensor. I would think the Nex 7 should be a more suited candidate to judge the color abilities of the Sony.
edwardkaraa wrote:
I am not happy with the colors of my girlfriend's Nex 3 either, but let's face it, Nex 3 is the lowest end and the C3 is even lower. I love the colors from my A900 and as mentioned above, my only gripe is the lack of biting sharpness at pixel level caused by the AA filter placed further away from the sensor. I would think the Nex 7 should be a more suited candidate to judge the color abilities of the Sony.
According to DxO, from a metamarism standpoint, all of the NEX cameras are ranked near the top, just below the A900. Of course, Joakim can chime in on whether that is relevant or not.
Speaking of DxO - they claim that the sensor in the C3 is the same as in the Sony A580, Pentax K5 & Nikon D7000 but that it has been tuned differently.
The problem with DxO is that I've found their results often to be at complete odds with reality. They've made some pretty ludicrous claims - especially when it comes to lenses - such as that the Canon 85/1.8 is superior to the Zeiss 100/2 Makro Planar. So I don't really consider them to be a reliable source.
denoir wrote:
Speaking of DxO - they claim that the sensor in the C3 is the same as in the Sony A580, Pentax K5 & Nikon D7000 but that it has been tuned differently.
The problem with DxO is that I've found their results often to be at complete odds with reality. They've made some pretty ludicrous claims - especially when it comes to lenses - such as that the Canon 85/1.8 is superior to the Zeiss 100/2 Makro Planar. So I don't really consider them to be a reliable source.
I think most would agree that their sensor data is pretty reliable. It is deciphering the data that can hang us up. As for lens testing, like with most lens testing sites, there are too many variable for me to hold much stock in those.
I would imagine that the C3 should be tuned differently, considering the power requirements of such a small camera. Maybe similar to how Intel's mobile chips work?