PDA

View Full Version : CMOS future of DSLR?


mxwphoto
10th of October 2006 (Tue), 02:25
Till recently, Canon has been the sole user of CMOS sensors for their DSLR. Now Nikon's D2X(s) use CMOS also (made by Sony). Will CMOS completely take over CCD's market share in the future of DSLR? Why is Sony's a-100 not a CMOS SLR when they supply Nikon with CMOS sensors? Are we on the verge of a standardized sensor revolution? What does everyone make of this? :D

COKE CAN
10th of October 2006 (Tue), 02:28
I am unaware of what you are speaking of. What is Nikon? You have a doorstop and paper weight with the NIKON name too? lol jk

CMOS may become standard over time, but CANON > *

cdifoto
10th of October 2006 (Tue), 02:31
I have no idea and don't really care so long as my camera produces what I want it to produce.

gjl711
10th of October 2006 (Tue), 11:05
The CMOS/CCD debate continues to perplex me. It is completely counter intuitive. Being in the electronics design industry most know that for image quality, dynamic range, and noise, and light sensitivity, CCD is king, but for cheap, easy to manufacture, low power, lets put a camera on a disposable phone, CMOS rules. But the best camera for noise out there is a CMOS chip camera. Obviously software plays a big part, and if so, imagine a Canon camera based on CCD technology with Canon noise reduction software. Am I missing something?

cdifoto
10th of October 2006 (Tue), 11:09
The CMOS/CCD debate continues to perplex me. It is completely counter intuitive. Being in the electronics design industry most know that for image quality, dynamic range, and noise, and light sensitivity, CCD is king, but for cheap, easy to manufacture, low power, lets put a camera on a disposable phone, CMOS rules. But the best camera for noise out there is a CMOS chip camera. Obviously software plays a big part, and if so, imagine a Canon camera based on CCD technology with Canon noise reduction software. Am I missing something?

Must be cuz the 1D original is CCD and noisy, as are all of Canon's CCD based P&S cams (which is all of them I think)...not to mention other brand's CCD cams.

gjl711
10th of October 2006 (Tue), 11:47
I have no idea and don't really care so long as my camera produces what I want it to produce.
This is my attitude exactly and after exhaustive analysis, for I am an engineer and can not make a decision unless I have at least 3 spreadsheets and a mountains of data sheets in front of me, my son cam to visit with his drebel and the noise at high ISO sold me.

Must be cuz the 1D original is CCD and noisy, as are all of Canon's CCD based P&S cams (which is all of them I think)...not to mention other brand's CCD cams.
It still doesn’t make sense though. My sons drebel at 1600 was cleaner than my old Nikon 8700 at 400 that was plain to see. Even when pixel peeping it was plain to see. But the engineer in me still what’s to know why? I have been all over the web and have not found any decent link explaining other than Canon is using a very high quality CMOS chip technology.

Bottom line though is that I am extreamly pleased with the result regardless of the reason and am a Canon convert.

cdifoto
10th of October 2006 (Tue), 12:01
It still doesn’t make sense though. My sons drebel at 1600 was cleaner than my old Nikon 8700 at 400 that was plain to see. Even when pixel peeping it was plain to see. But the engineer in me still what’s to know why? I have been all over the web and have not found any decent link explaining other than Canon is using a very high quality CMOS chip technology.

Bottom line though is that I am extreamly pleased with the result regardless of the reason and am a Canon convert.

I agree it doesn't make much sense but I guess we can just look at the little Japanese guys over in the Canon R&D Department as kids see Mickey Mouse in Fantasia. It's Magic. :)

mxwphoto
10th of October 2006 (Tue), 12:22
It still doesn’t make sense though. My sons drebel at 1600 was cleaner than my old Nikon 8700 at 400 that was plain to see. Even when pixel peeping it was plain to see. But the engineer in me still what’s to know why? I have been all over the web and have not found any decent link explaining other than Canon is using a very high quality CMOS chip technology.

Doesn't the 8700 have a smaller sensor size than the DRebel? Smaller sensor = higher noise.

gjl711
10th of October 2006 (Tue), 13:18
Doesn't the 8700 have a smaller sensor size than the DRebel? Smaller sensor = higher noise.

Sure, the 8700 being a p/s has a pixel density of about 141000 pixels per square mm while my XT has a density of 25000 pixels per square mm so that makes sense. Six times the density will result in more noise, but I also compared the XT against a Nikon D70 and it took the D70 down. The D70 has even a lower pixel density at about 17000 pixels per square mm. So like I said, counter-intuitive.

Like cdi-ink says, "It's magic"

mxwphoto
10th of October 2006 (Tue), 13:25
I read somewhere that because CMOS requires less "electrical charge (which causes noise)", less noise is introduced compared to CCD equivalents. I'm no engineer tho, so I can only regurgitate what I've read. :)

cdifoto
10th of October 2006 (Tue), 13:27
And it was nicely spewed too mxw ;)

FlyingPete
10th of October 2006 (Tue), 21:14
I read somewhere that because CMOS requires less "electrical charge (which causes noise)", less noise is introduced compared to CCD equivalents. I'm no engineer tho, so I can only regurgitate what I've read. :)

It was something like that, other things I have read have stated that Nikon had to go CMOS to make larger sensors, 6-8MP seemed to be the limit of large pixel (as in 1.5FOV sensor, not compacts) sensors, CMOS was required to go bigger than that.

lostdoggy
10th of October 2006 (Tue), 21:27
It was something like that, other things I have read have stated that Nikon had to go CMOS to make larger sensors, 6-8MP seemed to be the limit of large pixel (as in 1.5FOV sensor, not compacts) sensors, CMOS was required to go bigger than that.

Isn't the A100 larger then 6-8MP???

FlyingPete
10th of October 2006 (Tue), 21:31
Isn't the A100 larger then 6-8MP???

Yes same as the Nikon D200. They must have overcome what ever limitation they hit.

lostdoggy
10th of October 2006 (Tue), 21:34
It is completely counter intuitive. Being in the electronics design industry most know that for image quality, dynamic range, and noise, and light sensitivity, CCD is king, but for cheap, easy to manufacture, low power, lets put a camera on a disposable phone, CMOS rules...Am I missing something?

That is scary!!! Wonder why they are out sourcing??? Need to think outside of the bottle.

I believe that Canon was able to break the com bottle neck of CMOS by including essential circuits within the Die and chose CMOS because it draws less current then CCD. Low Noise was just a residual of the additional on die circuitry.

lostdoggy
10th of October 2006 (Tue), 21:36
Kodak currently is in joint venture w/ IBM and a Asian Chip company on R&D for new CMOS sensors so the future for CMOS sensor is looking brighter. Kodak makes sensors for Medium format camera currently so it would not surprise me if they release a CMOS version soon.

ron chappel
11th of October 2006 (Wed), 00:43
Not to appear rude or anything but there is way too much thinking out loud here and far too few facts.
The simple truth is that the CMOS *is* noisier than CCD's but canon have found ways to dramatically reduce the noise output.
Canon publish quite alot of info about how they achieve this ,see here-
http://www.robgalbraith.com/public_files/Canon_Full-Frame_CMOS_White_Paper.pdf

Some techniques are very obvious and basic (on chip amplifiers and fixed pattern noise reduction) but others less so.Thier technique for reducing random noise is extremely neat.

Overall the CCD verses CMOS thing is definitely not decided yet ! Some of the CCD's drawbacks are turning out to be nowhere near as bad as claimed by canon.Notice the supposedly more expensive CCD cameras are easly flogging canon on price !
Other issues like battery life and high ISO noise don't seem to be issues for most.
In fact high ISO noise seems to be the only remaining 'problem' for CCD's .

I look forward to the future when ways are found to add on-chip noise reduction technologies to CCD's that are currently only feasable with CMOS.If that happens it'll be the end of the road for CMOS which are allready near the end of their development potential.

mxwphoto
11th of October 2006 (Wed), 01:34
The simple truth is that the CMOS *is* noisier than CCD's but canon have found ways to dramatically reduce the noise output.
Canon publish quite alot of info about how they achieve this ,see here-
http://www.robgalbraith.com/public_files/Canon_Full-Frame_CMOS_White_Paper.pdf

Thanks for the corrections! It's surprising to find out that CCD in unmodified terms has better noise characteristics than CMOS. So in the end, I guess it all depends on the various methods used by camera companies to artificially fix each's respective shortcomings. After reading this, it seems that it doesn't matter as much which sensor is chosen, but rather what each company can do with them.:)

ron chappel
11th of October 2006 (Wed), 06:56
Yep that's very much it :)

...i probably should rephrase my asertions that CMOS are near the end of their development life.I don't know this of course -i'm just guessing by the way canon have had no major breakthroughs since the D60*,just minor improvements.

* I'm not sure of the exact timing -the random noise reduction system may have come later than the D60

Tom W
11th of October 2006 (Wed), 07:14
Yes same as the Nikon D200. They must have overcome what ever limitation they hit.


I think that the only limitation Nikon hit was that Sony didn't make one for them yet.

fi20100
11th of October 2006 (Wed), 10:13
So is there any new emerging censor technology that will leave both CCD and CMOS in the dust?

gjl711
11th of October 2006 (Wed), 14:26
Yup, the front runner seems to be the Foveon X3 technology image sensor. Though it is still relatively new and has not reached its potential, and at its core is based on CMOS, there are several architectural advantages that will in theory will supersede both CMOS and CCD. Top on that list is that each and every pixel is capable of capturing all colors and their pixels are stacked, not just RG or B as both CMOS and CCD are limited to today. Foveon just announced a 14.1 m-pixel and Sigma has build a couple of cameras based on this technology. It still has issues but think back where CCD or CMOS was 3 years after it was introduced and where these guys are today. You can take a peek at the future here (http://www.foveon.com/article.php?a=67).

FlyingPete
11th of October 2006 (Wed), 15:05
Yup, the front runner seems to be the Foveon X3 technology image sensor. Though it is still relatively new and has not reached its potential, and at its core is based on CMOS, there are several architectural advantages that will in theory will supersede both CMOS and CCD. Top on that list is that each and every pixel is capable of capturing all colors and their pixels are stacked, not just RG or B as both CMOS and CCD are limited to today. Foveon just announced a 14.1 m-pixel and Sigma has build a couple of cameras based on this technology. It still has issues but think back where CCD or CMOS was 3 years after it was introduced and where these guys are today. You can take a peek at the future here (http://www.foveon.com/article.php?a=67).


The killer for Foveon is that it is a licenced technology only available from one place.

Another interesting development has been the Fuji sensors with their two pixel types to increase dynamic range - which is one thing that could be improved on in just about everyones technologies!

rhys
11th of October 2006 (Wed), 15:13
Does it really matter? As long as the photos come out OK I don't particularly care whether they use CMOS, CCD or pickled hedgehogs as the sensor!

gjl711
11th of October 2006 (Wed), 15:24
And then there are the wacky single pixel guys…

Here (http://www.usatoday.com/tech/science/columnist/2006-09-30-single-pixel_x.htm)and here (http://www.dsp.ece.rice.edu/cs/cscamera/)

Citizensmith
12th of October 2006 (Thu), 00:04
Another interesting development has been the Fuji sensors with their two pixel types to increase dynamic range - which is one thing that could be improved on in just about everyones technologies!

Well thats kid of how our eyes work, two different receptors (rods and cones) one for detail and one for light sensitivity. Probably a good thing to try and copy.

FlyingPete
12th of October 2006 (Thu), 02:17
And then there are the wacky single pixel guys…

Here (http://www.usatoday.com/tech/science/columnist/2006-09-30-single-pixel_x.htm)and here (http://www.dsp.ece.rice.edu/cs/cscamera/)

Now if they combine that technology with several pixels all of different sensitivity, then they will really be onto something!

mxwphoto
12th of October 2006 (Thu), 02:36
The problem with single pixels it seems is that it still takes quite long relative to the conventional method. I don't think it will be of any real photographic use for at least a few more years.

ron chappel
12th of October 2006 (Thu), 08:45
So is there any new emerging censor technology that will leave both CCD and CMOS in the dust?

At the moment -emphatically NO :(
Ignore the talk from fovean.The newest model is just over 4.7Mp (NOT the 14Mp quoted above).Yes it inherently does a better job per pixel than bayer type sensors but only about 30% better (i forget the exact number) -it's certainly not the three times multiple they claim!
In case you doubt this there's no need to wonder -just check out the online review of any fovean equiped camera.The results are inescapable.
To be honest i really wonder how this kind of silliness continues.People seem to assume that a bayer filter array throws away pixels by interpolation - utter dumbness:rolleyes: . The bayer sensor interpolates COLOUR only .No pixels are lost!
The reason fovean type sensors are slightly more efficient is because colour imformation helps us see things more clearly.Fovean sensors don't have to interpolate that colour info ,which means they are slightly sharper per pixel.
What fovean really need is s sensor of about 6Mp+ ,giving a 'bayer comparable' resolution of about 9Mp
Make no mistake -the new four-point-something megapixel fovean will be a lovely camera.It's just that it's NO better than current competitors,and in some ways it's worse

The special high dynamic range sensor Fuji are offering is a seriously interesting technology.Unfortunately that's the only real benifit it offers...and it costs a fair bit .

rhys
12th of October 2006 (Thu), 09:38
There was an utterly stupid argument on FM some while ago about Bayer vs Foveon sensors.

While it's true that a Foveon image can be enlarged to a greater degree than a Bayer before artifacts become a problem, the fact is that because it's a smaller sensor it can't be enlarged to the same size as current Bayer sensor images - they're a few generations behind the times in image size.

Where Foveon really scores is at the pixel-peeping level. There is no flaw at the pixellular level - Foveon sensor images look sharp at 100%, which Bayer images do not.

Having said that, by the time a Bayer image has been through a printer (which automatically adds dithering) it will look no different from a Foveon.

Where Bayer really scores is two areas:
1. High ISO - Foveon really sucks at this.
2. Image size - Foveon really cannot compete with 4.7mp against 16mp Bayer however much they want to claim they can.

Essentially, around about the time FM started getting silly arguments like "Foveon is better than Bayer" (which I pointed out were silly) was about the time I left FM to their nonsense.

Tom W
12th of October 2006 (Thu), 09:44
I remember that thread, Rhys. Long and drawn out, and the OP sounded more like a salesman than anything else at times. I gave up on it when the proof of his assertions failed to materialize.

rhys
12th of October 2006 (Thu), 11:05
I remember that thread, Rhys. Long and drawn out, and the OP sounded more like a salesman than anything else at times. I gave up on it when the proof of his assertions failed to materialize.

I entered into the debate a couple of times but everything I said got dissed and poo-poohed so I left them at it.

Oddly enough, about a week later I tried to get back into FM and found I'd been banned. Heaven knows why and it's a testiment to the worth of the forum that I never bothered to find out why I'd been banned and never even tried to contest it. Shortly after I didn't get banned from another forum (very closely linked to FM and formed by FM members) but they didn't "ban" me but instead put a block on my IP address (sure, I can get around it easily enough) but the same attitude seems to prevail in both places so I figured it wasn't worth the effort of getting around bans and blocks.

It just struck me that when the debate drops to the level of the Foveon v Bayer a-la-FM that the forum has ceased to host meaningful debate. i'm very glad to see that POTN is not following the FM route. The other forum did a very similar thing over aviation security and pointing out the holes in the system seems to be what earned my ban. I reckon the guy running it must have been TSA or something and didn't like the holes in his system being pointed out - lol.

Debates on all forums tend to get very bland at times. I always try to inject reality and good humour whenever possible.

I do like the idea of the Foveon sensor but when Jeff reviewed it at dcresource, the lenses supplied were less than stellar. Indeed, one of the lenses (as far as I know, only Sigma make lenses for the SD series) had a soft spot on the bottom right third which meant for a very blurry section of a cornfield. This was confirmed by viewing an architectural shot.

If Foveon got their act together and produced a full-frame 10mp sensor with excellent high-ISO characteristics then it could well be a winner and could well be used as a replacement for film as originally conceived (drop in cartridge/back for 35mm SLRs. Unfortunately I don't see this as likely to happen.

Fuji's approach seems interesting with the sensor's ability to compensate for highlights so that the shadows have detail and the highlights aren't blown out.

Regarding the argument over whether we should have CCD, Foveon, CMOS or pickled hedgehogs as sensors, I can't really say I particularly care. All I'm interested in is a good picture and the technology takes a backseat.

Wilt
12th of October 2006 (Thu), 11:11
Oddly enough, about a week later I tried to get back into FM and found I'd been banned. Heaven knows why and it's a testiment to the worth of the forum that I never bothered to find out why I'd been banned and never even tried to contest it. Shortly after I didn't get banned from another forum (very closely linked to FM and formed by FM members) but they didn't "ban" me but instead put a block on my IP address (sure, I can get around it easily enough) but the same attitude seems to prevail in both places so I figured it wasn't worth the effort of getting around bans and blocks.

Just a comment that a ban is NOT necessarily intentional! I have been participating on another forum for over a year, to recently discover I had been 'banned'. No warning, no notice, so I inquired to the administrator. Simple error caused the 'ban' warning to pop up and it is now resolved. Sometimes one should not simply 'roll over'.

FlyingPete
12th of October 2006 (Thu), 14:47
Just a comment that a ban is NOT necessarily intentional! I have been participating on another forum for over a year, to recently discover I had been 'banned'. No warning, no notice, so I inquired to the administrator. Simple error caused the 'ban' warning to pop up and it is now resolved. Sometimes one should not simply 'roll over'.

The above mentioned forum does seem to have a bit of a reputation in the past for being a bit jumpy.

Wilt
12th of October 2006 (Thu), 14:50
The above mentioned forum does seem to have a bit of a reputation in the past for being a bit jumpy.

Nevertheless, one should not simply 'roll over'!

rhys
12th of October 2006 (Thu), 15:20
Just a comment that a ban is NOT necessarily intentional! I have been participating on another forum for over a year, to recently discover I had been 'banned'. No warning, no notice, so I inquired to the administrator. Simple error caused the 'ban' warning to pop up and it is now resolved. Sometimes one should not simply 'roll over'.

When I got banned, I'd already got a little tired of the backbiting. I can't say that fighting a ban on a forum that had already plummeted in my opinion seems at all worthwhile. I shed no tears about not being able to access a forum that I'd decided the previous week was mostly a waste of time. I noticed also that all the messages I'd posted in threads that were still current (having not posted in a week) had been deleted so I assume they *intended* to remove all traces of me. I noticed that the forums were getting slower and that they'd been banning people willy-nilly. Never thought I'd get banned but what the heck - I'm not that bothered. It was a free forum and I had no investment in it. I never paid for photo hosting as I have my very own website. As far as I'm concerned, bans are water off a duck's back - by the time they did I didn't have any interest in the forum any more. I never posted anything objectionable as that's just not my style. I figure they must have either been clearing out non-paying members in order to speed up the servers or they were overtaken by some form of madness (i'm being very charitable and presuming the latter).

Wilt
12th of October 2006 (Thu), 15:49
I have to admit there are several photography forums where I, too, would not bother fighting the ban, even if it were undeserved!

Few sites have the quality of interaction as POTN, which is why about a month ago I promoted the idea that members provide a contribution rather than POTN relying only upon purchases of POTN items. 'Insurance' to keep POTN alive and able to expand their servers as membership and volume of messages continually go up, is important IMHO.

ron chappel
12th of October 2006 (Thu), 20:04
Interesting points guys :)
I can fully understand Rhys not bothering to fight that ban- some places are simply not worth the grief;)
I am a little surprised that the FM forum can get like that though.I've never had much to do with the site but it looks like a high quality place,and i assumed that because it's run by an enthusiast it would be run well.
Is the FM forum usually abit narky or is the above example abit of a one off ?

rhys
12th of October 2006 (Thu), 20:47
Is the FM forum usually abit narky or is the above example abit of a one off ?

Comments on another forum revealed that a lot of the usual posters had been banned from FM - not argumentative people but people who usually posted interesting and worthwhile, well-thought-out articles.

My experience was that it was fairly easy-going until several things hit one after another.

First the servers started to slow. They sorted that out but at the same time something strange happened to some of the links.
Then they started to get spammers posting and never seemed to be too good at stopping them.
Then the servers started to slow again and wholesale bans began to appear - people you'd never have considered would have been banned.
Then the debates became a bit Tom and Jerry.
Then I got banned.

I really don't know what has happened to FM other than it went downhill about 6 months after I joined. Having said that they were a bit snooty to me when I joined. I assumed it was a case of somebody new joining the club.

I can't really say other than this is what my experience was.

Wilt
12th of October 2006 (Thu), 21:21
It is still rather slow!

Tom W
12th of October 2006 (Thu), 22:43
It is a bit slow on FM, and not what it used to be. I noticed a couple of good members get banned or otherwise chased away as well.

I think it got overrun with DPR-style posters over the last several months. You know, the bickering, the baiting, and all that.

I had a little discussion recently with the "buy and sell" moderator over a minor matter.

rhys
13th of October 2006 (Fri), 10:10
It is a bit slow on FM, and not what it used to be. I noticed a couple of good members get banned or otherwise chased away as well.

I think it got overrun with DPR-style posters over the last several months. You know, the bickering, the baiting, and all that.

I had a little discussion recently with the "buy and sell" moderator over a minor matter.

Well at least somebody else has noticed that and it's not just me saying what happened to me and reporting what others have said. That makes me feel a lot better. I was beginning to feel like I was the only person here that had had problems on FM and that I was the only person known to have been banned from it. I was beginning to feel that people were looking at me and saying "ooh. That's the pot calling the kettle black".

I don't know what a DPR poster is. I assume DPR is another group? Possibly Digital Photography Resource?

Myself, I also use DCResource and Columbia Camera Club. DCResource is quite good. I've never used DPR. Of course, Columbia Camera Club is good but with less than 10 active posters it's a little spartan.

Tom W
13th of October 2006 (Fri), 12:30
Sorry - DPR is the forum at "DPReview". There's plenty of flame wars over there in the forum, with minimal moderation.
Don't confuse that forum with the great camera reviews that the site owner, Phil Askey provides.

ron chappel
14th of October 2006 (Sat), 08:01
it's odd huh? DPreview has arguably THE best camera reviews in the world yet allows the idiocy that goes on in thew forums

FlyingPete
14th of October 2006 (Sat), 14:10
it's odd huh? DPreview has arguably THE best camera reviews in the world yet allows the idiocy that goes on in thew forums

Now there is a forum I keep well clear from :shock:

It is amazing the different cultures that develop in the forums, I think we do so well here as we are all enthusiasts who use the same brand of gear, so there is little N vs C bickering etc (I know the arguments are more than just that there).

Oh and it has to be about the worst forum engine I have ever seen!

Tom W
14th of October 2006 (Sat), 14:23
Now there is a forum I keep well clear from :shock:

It is amazing the different cultures that develop in the forums, I think we do so well here as we are all enthusiasts who use the same brand of gear, so there is little N vs C bickering etc (I know the arguments are more than just that there).

Oh and it has to be about the worst forum engine I have ever seen!

It's a strange forum, for certain. There are some very knowledgeable posters there, but there's just so much trash talk that it takes time to find the good stuff. And then you have to weed out 1/2 the replies.

Things are much better behaved here. :)

rhys
14th of October 2006 (Sat), 17:24
I am always amazed by Nikon versus Canon bickering.

I'm equally amazed at the constant film v digital bickering too.

Now I'll shock everbody - I love my Nikon 35mm SLRs and yet I use a Canon dSLR.

I went over to Canon not because Nikon was bad but for three reasons:
1. I heard that Canon's image quality was slightly better.
2. Nikon claimed to support my AIS lenses but in fact does no such thing due to the utter lack of an aperture window in the viewfinder and the complete lack of coupled metering in many of their dSLRs.
3. Canon already had a full frame sensor in the pipeline.