John
Interesting! While I knew the smaller chips were nosier at higher ISO I didn’t expect that the resolution would also suffer. Of course this is just one images comparison so I best keep an open mind.
Grant
Well, I’ve been stalking the Minolta A2(same chip as the Sony) and I have to say I’ve been pretty dismayed at the images that have been posted from that camera in the various forums, although the people who have the camera seem quite happy.
FWIW, there have been a few threads to the effect that for the Minolta at least, the photoshop CS RAW plug-in fares miserably in comparison to the camera software.
I suppose the real comparison would be where the rubber hits the road Β an 8 x 10" print, say.
If my math is correct, images from each of these cameras would offer the following maximum dpi to print an 8 x 10 (the Rebel and 1Ds images would be cropped because of their 2:3 format instead of 3:4):
1Ds (4064 x 2704): 338 dpi
Pro 1 (3264 x 2448): 306 dpi
Rebel (3072 x 2048): 256 dpi
G5 (2592 x 1944): 243 dpi
G3 (2272 x 1704): 213 dpi
Then consider the pixels per square mm. The higher the number, the smaller they must be, and the more potential for noise. In general, a LOWER pixel per square millimeter would be desireable.
1Ds (11.0mp, 35.8 x 23.8 sensor): 12,910
Pro 1 (08.0mp, 08.8 x 06.6 sensor): 137,741
Rebel (06.0mp, 22.7 x 15.1 sensor): 17,504
G5 (05.0mp, 07.2 x 05.3 sensor): 131,027
G3 (03.9mp, 07.2 x 05.3 sensor): 102,201
I don’t think there is any ‘straight’ relationship between these two (total pixels and pixels-per-sensor-size). The final kicker is the printing device. Interpolation and other technologies will certainly have a bearing on the finished print.
Ultimately, of course, quality is somewhat subjective. I am pleased with 8 x 12" prints from my Digital Rebel. My 2mp S100 images top out at 5 x 7, and those look a bit grainy. I have never seen an 8 x 10 from a 1Ds, Pro1, G5 or G3, but it would be interesting to have a comparison print of the same scene from all of these.
Everyone wants the best quality. But the 1Ds is about $8000 US. I think I’ll stick with my Rebel for now… π
In those head-to-head comparisons on Steve’s, it appeared to me that the Nikon 8700 had an edge; wonder what they’re doing differently?
Chuck
What can I say? Oh, how about NIKON π
Grant
Grant, I can’t argue with you…this time!
π
I have purchased the canon powershot pro 1. There is little doubt that Canon are happy to mislead their customers in respect of this camera. Noise is a real problem – they say, in their blurb – "can make A3" but noise is unacceptable at that size of print. I seriously wonder if I have a case against Canon – but I should think trying that would only make their solicitors rich(er).
What is the opinion out there?
I think everyone knew that noise would be a problem with the small sensor high megapixel camera.
—
Have A Nice Day, π
James Hutchinson
http://www.pbase.com/myeyesview http://www.myeyesviewstudio.com/ wrote in message
I have purchased the canon powershot pro 1. There is little doubt that
Canon are happy to mislead their customers in respect of this camera. Noise is a real problem – they say, in their blurb – "can make A3" but noise is unacceptable at that size of print. I seriously wonder if I have a case against Canon – but I should think trying that would only make their solicitors rich(er).
What is the opinion out there?
Maurice
If you are not happy send it back.
Grant
BTW, since this has come up, what does everyone recommend for noise reduction? (Personally I’m interested in mac solutions, but I’m sure others–Maurice, maybe?–would like to know for pc. Incidentally, I know about Grain Surgery–just wondered if any of the less expensive products were worthwhile, also.)
What is the opinion out there?
About suing Canon?
Go for it, if you can find a mouthpiece that will take it on commission.
What are your damages? The cost of the camera itself, tops.
Good luck.
Mac
Barbara,
Regarding Noise reduction. I’ve been awaiting the release of Neat Image for Mac, which is currently under developement. I’ve heard good things about it but I don’t know yet ‘when’.
Barbara
Is the A1 that noisy? All the images I have seen fro this camera were taken at low ISO and the DiMAGE seemed to behave very well.
Grant
Grant, no, it’s not very noisy at all at lower ISOs, at least not in print. Onscreen, though, the area behind the focal plane is not exactly noisy, but it’s not as smooth as it was with the canon, either. It’s more of a frosted glass effect, I guess. I think of the difference between the two cameras as travel mag v. photojournalistic style, if that makes any sense.
But doing so much macro stuff, I am upsampling pretty tight crops quite a substantial amount and there I do start to see some color noise that I would like to have a go at smoothing a little. Never in the focus area–that’s always clean as a whistle (or horribly out of focus, of course, but that’s not the camera’s fault.)
Really, I can’t say enough good things about this camera–I like more every time I use it.
I’ve been getting around it when necessary by using the lens blur filter in CS on the background, but I would like to see how large I can go for prints and I’m pretty sure if I go up to a A3+ I’m going to see noise on the subject, too.
Barbara
Thank you that does make complete sense to me. I can spot grain from a mile away but unlike most of the digital reviewers I don’t see this as necessarily a bad thing. There is good grain and bad grain. While often I strive for less grain sometime I want more grain.
Grant
Hi, Grant. I agree with you completely. A total absence of grain all the time in all situations is also a total absence of character a lot of the time.
Here’s <
http://members.aol.com/bkbrun/fiddlehead3.jpg> an example of what I was talking about. Not the greatest pic, by any means, but the noisiest I have.
Juergen,
Well, *I’m* impressed!
Did you mask out the subject, or let the filter run on the whole image?
Byron
Byron,
No, you just run the whole image thru as is.
Juergen
Juergen, that did a great job! Will have to look into Neat Image…..thanks!
Juergen, which of the four versions of Neat Image did you settle on?
Chuck
Thanks, Juergen, that was pretty impressive. Now if the OS X version would just become availableΒ
Barbara, I really hate to buy a plug-in that costs as much or more than the program….but I may make an exception on this one.
Chuck
Chuck, there are several free alternatives for windows. Have you tried any of them?
Barbara, I suppose I ought to do that and compare the results with the trial version of Neat Image, huh?
Chuck,
I have the free demo version. On their website they say: "The Demo edition of Neat Image is distributed as freeware for non-commercial use."
Juergen
Juergen, thanks. I’ll start there. The souped-up Pro version is a plug-in with 16-bit and batch capability – but it’s $75…
Chuck
Juergen, it took 3 seconds to download; I’m going to try it on some troublesome images….
Sorry, Barbara, I forgot that you are running OS X. π
Juergen
No problem, Juergen. If it works that well on the pc hopefully the mac version will be equally good when it comes out.
Chuck,
Good! You may want to try Barbara’s original (see above) and compare…
Juergen
Juergen, you convinced me with your rendition. I’m going to find one of my grainier shots and put Neat Image through its paces…
While waiting for neat image on the mac to finish beta testing I’ve been using the Elements adaptation of Digital Denoise deluxe (from the snapactions web page). If you have full photoshop the original, rather more versatile actions are at
<
http://www.2morrow.dk/75ppi/coolpix/actions/default.htm> They are free…
Once I received Barbara’s base image and Juergen’s redo I decided to have some fun
I copied Juergen’s redo image onto a layer over Barbara’s. Then I added six copies of Barbara’s image between them. I then proceeded to apply separate noise filters on three of layers, a Median, a Despeckle and, a Dust and Scratch all at their lowest levels. Finally I created a very fast feathered mask and used it on three more layers and applied a stronger level of Median, Despeckle, and Dust and Scratch. Then by clicking all they eyes off I could select any two to compare and by toggling the top level on and off I could easily see the difference. While the commercial filter was very good it was easily beaten but two of the three masked layers. The Dust and Scratch was only marginally better than the commercial filter.
It was a good exercise and I recommend every one try these things.
Grant
Susan – good set of actions – thanks!
Chuck
Grant: What is a ‘fast feathered mask’?
Chuck
Chuck
Use the lasso tool and set a feather on it (the output and experience determined degree of feather). For this image feather of 3 to 6 will work I used 3. Then a very fast rough capture of the image. Some where between 5 and 10 seconds for the fiddlehead. Because of the feathering the selection doesn’t have to be too accurate as the transition occurs over a broad range.
Grant
Grant, thanks! I’ve tended to go with the selection brush lately, although it sometimes lets me down. If you inadvertently use a soft brush with it, it tends to not do a complete job of selecting (semi-transparent).
Chuck
With images like this one where only the background has noise, then I agree with Grant that masking off the subject and applying the denoise filters to the background is just as good, and much much faster. Neat image and the actions I linked to can be really slow. Where Neat Image comes into its own is in images where the subject is also noisy (ISO 400 images on my G3, and ISO 100 and 200 where they were underexposed and I’ve had to bring up the shadow detail). If you set it up right – and you can set it up very precisely for the particular camera, ISO setting and even the particular image – then it can do much better in these circumstances than actions or hand masking.
Barbara – I’m curious as to what ISO setting you used and how close a crop that image is. The subject detail is fine, but the background noise is distracting.
Hi, Susan. ISO 100, and here’s <
http://members.aol.com/bkbrun/fernsmall.jpg> a link to a reduced version of the original so you can see the crop.
What I find a bit strange about this camera is that at 100% the onscreen display is as you see, but the prints don’t reflect that, fortunately. Not sure why that is.
EDIT BTW, are you beta testing NI or do you use it on a pc?
Barbara. Hmm Interesting. I’m glad the prints don’t reflect the noise – but it does reinforce the message for me that the next camera I buy is a DSLR. All the higher Mpixel prosumer cameras seem to have more noise than I can deal with, and I find the 4Mpixel G3 iritating enough. – Although it’s odd, as the zoom on my SLR is slow I’ve always tended to shoot with ISO 400 film which is just (if not more so) as noisy – but I find something about the quality of film grain less irritating than digital camera noise.
Yes I’m beta testing NI. No I’m not going to say any more about it than that! Susan S
No I’m not going to say any more about it than that!
Tell them to hurry up! π
A DSLR would be nice, but I think it will be a long time before I can get one with that kind of lens for that kind of price.
Grant, I’ve been playing around with your methods, and while they work great on the background, I would have to say that Juergen got better results on the out-of-focus areas of the stem than I was able to get with the existing PE filters. I couldn’t get rid of the noise there without sacrificing more detail than NI took when Juergen did it. Just an observation.
It’s all about trade-offs isn’t it? – Given I already have the G3 it doesn’t seem worth moving up to the new top level prosumer non-DSLR camera such as the A2 or the Pro1 – the increase in quality of images, better zoom range, better macro, more megapixels which would give better cropped images, would be there – at the expense of (possibly) slightly more noise, particulaly at higher ISOs. Given that the really major issues I have with the G3 are the focusing speed and poor manual focus and the noise levels, for me the considerable extra cost is not worth it and I’m saving for a DSLR. On the other hand I know I’ll really miss the flip out screen and all-in-one portability of the G3 if and when I go back to using an SLR. (I don’t like the really small digital cameras – I’ve got big hands for a female and i find it hard to hold them steady and find the buttons)
Hi, Susan. Yes, there are definitely things I would prefer about an SLR myself. The EVF on the A1 is a big step up from an LCD, but I still find it tricky for fine-tuning the focus. (Of course, it would have helped that poor fern if I had remembered to turn off the anti-shake, but I’m slowly getting better at thinking of that.)
I wish there were more prosumers with lower megapixel counts–the A1 is generally a lot better noise-wise than that photo indicates (I looked for my worst one), but it’s not as smooth as my s400, for sure. I definitely don’t think the A2 is worth $400 more for the improved EVF and faster focus, though.
Here’s <
http://photohobby.net/webboard/show.php?Category=alldata& ;No=666> an interesting comparison of the Pro 1 and the A2.
Interesting comparison, Barabara. I didn’t like the noise in the A2 and I thought the purple fringing in the Pro1 was unpleasant. But you really do have to blow the images up to 100 per cent to notice them. I still think if I didn’t already have a digicam I would pay slightly less than half the price of a Pro1 (at Australian street prices) and get a G5 (given the G3 is not available in Australia now). The Pro1 is actually 300$A more than a 300D with kit lens here, (high street prices) which makes it difficult to justify!
Barbara
The NI filter does a wonderful job in remove the pattern in the background. There are two things that I don’t like about NI filter. On is the cost but then I tend to be too frugal at times. The other is that the filter also removes information from where I want it to be. The highly polished stem up of the fern does loose some detail with the NI filter. I guess it is all a matter of what detail you are not concerned with. I thought about making my test file available but it ended up to be 16 Meg.
Grant
That’s more what I would hope for – it looks much as my G3 does at 50 ISO with an out of focus background in macro shots – some noise but not distracting (and at normal print sizes the noise doesn’t show either). The benefit of the better lens and resolving power of your camera shows up too. I find with my G3 at higher ISOs with an in focus image and correct exposure the noise can be managed – it’s a really fine line to tread though, particularly with white or pale subjects and contrasty scenes, where you don’t want to overexpose and blow out the highlights, but underexposing shows a lot of noise in the rest of the image if you try to bring out shadow details.
Hi, Susan. Just wanted to let you know that I have switched to using Dalibor Jelinek’s plug-in for .mrw and the noise situation is hugely improved from what I was getting with the CS raw plug-in. No comparison at all.
