DPReview um, reviews the Canon PowerShot SX1!
- Thread starter Flipped Gazelle
- Start date
The differences between that camera and the SX10 are only somewhat apparent at 100% viewing. They recognize this in their review but it seems like they were expecting miracles out of the image sensor technology.
Kinda reminds me of their "just" recommended review of the Canon 50D. All the features and many new additions to the 40D but because the image is slightly noisier(only really noticeable at 100% viewing with RAW images, it gets a "Just" recommended review.
Kinda reminds me of their "just" recommended review of the Canon 50D. All the features and many new additions to the 40D but because the image is slightly noisier(only really noticeable at 100% viewing with RAW images, it gets a "Just" recommended review.
I don't know if DPR was expecting miracles, but certainly, given Canon's history with CMOS, I'm sure they were expecting better IQ.
And, given that the SX1 is over $200 more expensive than the SX10, it's pretty easy - and justifiable - to be disappointed, IMHO.
perhaps...but lets not forget that your getting more than just a CMOS sensor. There are benefits that come with that change other than IQ.
People are talking like CMOS is some superior technology that is supposed to bring better IQ. There are trade-offs between CCD and CMOS technology, and while CMOS enables some features like video recording, better IQ is not one of them. In fact, CCD's have inherently better technical IQ, although it really depends on many factors.
What does sensor type have to do with video recording? There have been both CCD and CMOS video cameras for years now...
Also, if CCD has inherently better technical IQ, why did every major DSLR company switch to CMOS?
I used to have a Nikon D200, which was among the last CCD equipped DSLRs that Nikon made. At ISO 800, the D200 was producing more noise than my EOS 40D at ISO 1600, despite having a slightly larger sensor (comparing the RAW files from each). The 40D also has a stop more dynamic range than the D200, and experiences much less of the drop-off in dynamic range as ISO is raised.
So I'm curious as to what you mean when you say CCDs have "inherently better technical IQ"
CMOS can make it easier/cheaper/more practical to implement video in a portable device, but that doesn't mean it's impossible to do it with CCD.
The 40D was released a year and a half following the D200, of course there's a lot of technological advances in HW and improved firmware algorithms that give it better IQ. The resulting IQ depends on a lot more than just the sensor itself. However, it's inherently more difficult to reach a certain level of dynamic range, low noise and uniformity of amplification in a CMOS than a comparable CCD sensor because the light-sensitive pixels are relatively smaller, due to having additional processing logic and amplifiers on the sensor itself.
Also, I'm pretty sure Nikon used CCD's on a few more camera's after the D200. The Nikon D40, for example, also used CCD's.
- Mar 4, 2000
- 27,370
- 239
- 106
The primary reason for using CMOS over CCD is cost. High res CCD production is much more expensive than CMOS, and market competition dictates the rest. CCDs would therefore be used only on high cost, top drawer models where the cost can be recovered.
yup, as mentioned CCDs are reckoned to have better IQ but you can integrate a lot of onchip processing on a CMOS sensor which makes the camera as a whole cheaper to produce.
The evidence is every where; Google is your friend.
It's true that CMOS is supposed to have lesser IQ and more noise than CCD. However, things have changed over the years especially since when Canon matured CMOS technology showing how CMOS can have good IQ.
If we're strictly talking about CMOS as it is, I'd agree with you, but since CMOS in design is coupled with other supporting devices enabling it to have superioir IQ, I don't agree with you.
Also, CMOS manufacturing is supposed to be cheaper than CCD, but in reality, the cost is very identical. In fact, if you consider costs of designing CMOS, CMOS costs more than CCD.
Link: http://www.dalsa.com/corp/markets/CCD_vs_CMOS.aspx
Read PDF files there.
It's true that CMOS is supposed to have lesser IQ and more noise than CCD. However, things have changed over the years especially since when Canon matured CMOS technology showing how CMOS can have good IQ.
If we're strictly talking about CMOS as it is, I'd agree with you, but since CMOS in design is coupled with other supporting devices enabling it to have superioir IQ, I don't agree with you.
Also, CMOS manufacturing is supposed to be cheaper than CCD, but in reality, the cost is very identical. In fact, if you consider costs of designing CMOS, CMOS costs more than CCD.
Link: http://www.dalsa.com/corp/markets/CCD_vs_CMOS.aspx
Read PDF files there.
I'm well familiar with the PDF files on that site. While they state technological advances have allowed CMOS sensors to rival CCD's in IQ, nowhere does it say CMOS is superior in IQ than CCD's, unlike what Canon marketing would like to have you believe. In fact, Canon is not the holy grail of IQ - there are digital medium format camera systems which offer substantially superior IQ, and those use CCD sensors.
I've never said or implied that CMOS has better IQ than CCD. I was pointing out how CCD does not have better IQ than CMOS due to the fact CMOS evolved over the years.
Sensors for digital medium format are very different issue. They are desinged to fit into certain market thus optimizing certain characteristics of the sensor while abandoning others.
Canon, of course not, is the holy grail of IQ. When I talked about Canon, I wasn't talking about how Canon is the king of IQ and that they use CMOS.
I was talking about how they decided to go for CMOS when others were saying that they're crazy to so. However, not only they have made CMOS sensors that really worked well, their series of CMOS sensors, had no problem competing others' CCD tech in terms of IQ.
You see, technically, it's true that CMOS is supposed to show a lot more noise than CCD. However, if someone comes out saying that CMOS does show more noise than CCD, he's just being ignorant becasue he's disregarding how things are different in the real world. Again, it's true CMOS is supposed to have weaker IQ than CCD but as I've pointed out, things have changed over the years.
Sensors for digital medium format are very different issue. They are desinged to fit into certain market thus optimizing certain characteristics of the sensor while abandoning others.
Canon, of course not, is the holy grail of IQ. When I talked about Canon, I wasn't talking about how Canon is the king of IQ and that they use CMOS.
I was talking about how they decided to go for CMOS when others were saying that they're crazy to so. However, not only they have made CMOS sensors that really worked well, their series of CMOS sensors, had no problem competing others' CCD tech in terms of IQ.
You see, technically, it's true that CMOS is supposed to show a lot more noise than CCD. However, if someone comes out saying that CMOS does show more noise than CCD, he's just being ignorant becasue he's disregarding how things are different in the real world. Again, it's true CMOS is supposed to have weaker IQ than CCD but as I've pointed out, things have changed over the years.
I've never seen anything from Canon saying that the '1 has a cmos chip for better quality. I'm almost positive they came out and said it was only for features not for picture quality. That said I don't see why anyone would buy the '1 over the '10....The cost is just simply so high for the tradeoffs.
And I don't know if I would say the digital medium format camera systems are "better" than say..a 5Dmk2 or a D3x...they are just a completely different tool for different purposes. They deliver a bigger, higher resolution file, sure...And in some cases with no anti-aliasing filter robbing of detail, etc. However, they aren't necessarily better, they are just a completely different tool. If you need extreme resolution they are what you'd want... But try doing anything where you need to focus quickly, or anything where you need a high iso...not even a high iso really, hell, iso 800 even. Have you seen high iso on a digital MF camera? For real high isos..well, it's moot anyway--you can't use iso 3200 or 6400 on a digital mf camera. If you just want obscene high detailed resolution there are also large format scanning digital backs, though I've never seen one in person...
Just saying, a medium format dslr is not "better" than a regular dslr from nikon or canon in the same way that a dslr is better than a point and shoot...it's not going to open up new picture possibilities or allow you to take shots you couldn't before, or make your prints look any better at all unless they are huge, etc. It's only an advantage if you really need one of the things the mf offers. Lets be honest regular digital slrs are pretty safely, quality wise, in the territory that medium format film used to be.
Just look at Canon's web page on the SX1: Text
I can see how that might lead some people to believe that CMOS sensor is supposed to be an advantage in terms of IQ, somehow possibly even rivaling that of EOS DSLR's. But then again, considering the continuous bragging about its movie mode, never mind the fact that its rolling shutter severely limits the usefulness of such a gimmick, and it's easy to see how the hype surrounding the camera might create unrealistic expectations.
As far as medium format cameras go, they may not rival or beat DSLR's in every scenario like shooting sports, but in a studio setting if you want the ultimate IQ, they can capture detail which DSLR's miss. My point was they ALL use CCD sensors, despite the fact that most DSLR's are switching to CMOS.
MF cameras have much better image quality mainly because the physical size of their sensors are much larger...it's not an apples to apples comparison.
If you compare CCD vs CMOS ONLY in the world of DSLRs (where sensor sizes are similar), then the overall package that CMOS can deliver is better than that from CCD.
What's the best CCD DSLR in terms of image quality? Probably the Nikon D200. So since you say it's not fair for me to compare it to the EOS 40D, let's compare it to the EOS 30D, a camera which was from the same era as the D200 and priced lower than the D200.
Even when compared to a CMOS sensor equipped Canon EOS 30D from the same era, the D200 cannot compare in high ISO performance.
Also, if you look at dynamic range, the D200 also cannot compare. At ISO 1600, the 30D still has 8.0 stops of dynamic range, not far off from the 8.4 stops it had at ISO 100. The D200, on the other hand, has dropped from 8.2 stops at ISO 100 to 6.9 at ISO 1600. Even at ISO 800, the D200 is showing pronounced DR drops, falling from 8.2 stops to 7.2. At ISO 800 on the Canon 30D, DR is still a strong 8.2 stops (which is what the D200 has at ISO 100, by the way).
Both of these cameras are from the same era, so you can't say that one has a major technological advantage over the other. Now, as far as the argument of whether CCD has inherently better IQ by itself, that's quite irrelevant to me. Who cares if a CCD is inherently better by itself? When a sensor is used in a camera, it is used in conjunction with many other parts, and it is the sum of these parts that results in the camera's final IQ. In this sense, it is clear that Nikon wasn't able to get as much out of the D200's CCD as Canon was able to get out of the 30D's slightly smaller CMOS sensor.
Umm...at first, you were saying how it's the people who are talking CMOS is better than CCD and it's not true. I and the others pointed out that they're very identical and CMOS might even be better nowadays. The trade-offs that existed long time ago is not so valid as CMOS tech. has grown.
Given Canon's sucess with CMOS over the years, I don't think Canon is fooling people. At best, you could've said what Flipped Gazelle said: "given Canon's history with CMOS, I'm sure they were expecting better IQ." In other words, what Flipped Gazelle said does make sense but saying CMOS just isn't as good as CCD in terms of IQ due to it being CMOS is just plain ignorant.
You made an example out of MF sensors being CCD and said it's becasue CCD has better IQ than CMOS. As I've pointed out before, it's not a vaild example because MF sensors are desinged to be used in very specific settings. In that boundary, MF CCD does bring superior quality but out of it, it becomes terrible, much worse than that of much smaller CMOS sensors. That's called 'trade-off,' the trade-off you've mentioned is just not valid nowadays. In addition, As 996GT2 has pointed out, you failed to consider the physical size being different. If I follow your route, I could even argue that CMOS is just better if you look at IQ of Sony A900 and other cropped sensor cameras from the same brand. Hell, I could even say CMOS is just better making an example out of Canon 5D and all the other cameras that use cropped CCD sensor.
Again and Again, things have changed and tech. has evolved. It's not like several years ago when people were saying Canon's venture into CMOS would fail because of inherit limitation of CMOS. Following your logic, one could even argue those cameras that use CMOS will show more noise due to the fact CMOS is noiser inherently. You know it isn't so. Same goes for this IQ issue.
Where are you getting those numbers from? And how does a single number quantify the amount of detail captured in a scene at a given ISO?
So you don't think Canon is fooling people by mentioning "CMOS" and "superior IQ" in the same sentence when referring to a camera that obviously has inferior IQ when compared to its CCD counterpart? Moreover, when it is not even in the same league as the DSLR's from "Canon's lauded EOS line?"
At first, what you were saying was based on your 'PEOPLE think CMOS is better than CCD in terms of IQ but CCD is better' talk. I and others pointed out how it's not true anymore.
After that, you, all of the sudden, came up with that Canon's marketing talk trying to directthe disscussion into somethine else.
"The 10.0-megapixel SX1 IS is the very first PowerShot equipped with a CMOS sensor, a hallmark of Canon's lauded EOS line.
The first PowerShot camera to feature a 10.0 Megapixel Canon CMOS sensor for superb image quality."
What's wrong with that? It's is true that SX1 IS is the first powershot utilizing CMOS sensor that's been first introduced in EOS line. If Canon implied that the new Powershot now produces images like EOS cameras due to its CMOS sensor, I'd say it's wrong but nowhere Canon said that. The second line is just plain and simple line that's been used over and over by Canon and other makers. 'Someting something megapixel whatever for superb image quality, performance, and whatever' has been used so many times that it's become its own cliche. Even back that when Canon was using Sony's CCD, same slogan was used. CMOS was noted because it's the first time Canon used CMOS sensor in that line. Had Canon said 'Thanks to the new CMOS sensor, its IQ is superior,' then it's be fooling people thus wrong.
However, on this issue, I do understand that someone may read it in your way and that'd make a case that you can argue. Nonetheless, it's a different 2nd issue.
Now, back to the original disscussion that you ignore to answer, do you still think CCD is better than CMOS in terms of IQ? Based on what you've stated, you still seem to think CCD is just better even at this stage. Your back-up example is MF cameras still using CCD over CMOS. As pointed out by me and others, such argument is not valid. You see, you shouldn't make an example out of very a few narrow cases. CMOS using SX1 being worse than its CCD counterpart does NOT mean CMOS is inferior than CCD. I agree SX1 is worse than SX10 but it's just that. Just read 996GT's posts and yours. 996GT is coming out with specific numbers and cases backing up his points. You, on the other hand, doesn't seem to have anything better to say 'MF cameras use CCD over CMOS so that means CCD is better.'
After that, you, all of the sudden, came up with that Canon's marketing talk trying to directthe disscussion into somethine else.
"The 10.0-megapixel SX1 IS is the very first PowerShot equipped with a CMOS sensor, a hallmark of Canon's lauded EOS line.
The first PowerShot camera to feature a 10.0 Megapixel Canon CMOS sensor for superb image quality."
What's wrong with that? It's is true that SX1 IS is the first powershot utilizing CMOS sensor that's been first introduced in EOS line. If Canon implied that the new Powershot now produces images like EOS cameras due to its CMOS sensor, I'd say it's wrong but nowhere Canon said that. The second line is just plain and simple line that's been used over and over by Canon and other makers. 'Someting something megapixel whatever for superb image quality, performance, and whatever' has been used so many times that it's become its own cliche. Even back that when Canon was using Sony's CCD, same slogan was used. CMOS was noted because it's the first time Canon used CMOS sensor in that line. Had Canon said 'Thanks to the new CMOS sensor, its IQ is superior,' then it's be fooling people thus wrong.
However, on this issue, I do understand that someone may read it in your way and that'd make a case that you can argue. Nonetheless, it's a different 2nd issue.
Now, back to the original disscussion that you ignore to answer, do you still think CCD is better than CMOS in terms of IQ? Based on what you've stated, you still seem to think CCD is just better even at this stage. Your back-up example is MF cameras still using CCD over CMOS. As pointed out by me and others, such argument is not valid. You see, you shouldn't make an example out of very a few narrow cases. CMOS using SX1 being worse than its CCD counterpart does NOT mean CMOS is inferior than CCD. I agree SX1 is worse than SX10 but it's just that. Just read 996GT's posts and yours. 996GT is coming out with specific numbers and cases backing up his points. You, on the other hand, doesn't seem to have anything better to say 'MF cameras use CCD over CMOS so that means CCD is better.'
Go look at dpreview's tests of each camera.
I quote their testing methods for dynamic range:
The 30D has much more (over a stop) dynamic range than the D200 at higher ISOs (above 400), meaning its images possess much more highlight and shadow detail. This is a very important trait (even more so if you post process and adjust exposure in Camera RAW), and is one reason that I believe CMOS to be better than CCD.
I also quote dpreview's ISO noise testing results:
Keep in mind that this is comparing the 3 year old Canon EOS 30D to the last CCD camera Nikon made. Current CMOS cameras are even better in terms of noise and DR performance. Even with slightly better NR algorithms than Canon (at the time), Nikon's D200 still cannot come out on top in terms of high ISO performance.
By that logic, I could argue the 30D is just better than the D200, and that's that. You have an example where CMOS offers better IQ, and with the SX1, I have an example that shows the opposite. So you can't say CMOS is just plain better across the board either.
As for MF cameras, you can try to disqualify their merit because they only show better IQ in low-iso shots, and in huge prints, but the fact remains that under those conditions, they still have better IQ. My point is they don't automatically get better IQ just because they use CCD, but in a market with users demanding the highest possible IQ for their needs, and supported by 5-figure prices, ALL MF cameras are using CCD sensors, not CMOS.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 21K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter