1. When doing comparison shots, have a mouse over load the second image and a mouseoff show the first. Or simply have one image and a list of links that dynamical and instantly change that image so comparisons can be made in detail. This is what any self respecting site that has to so with image comparisons would do. This screams "early 90s"
2. Though one is usually important, its not here for the even worse reason that the originals were all far better than the filtered images! Why bother uploading this rubbish? Did you even bother to compair or is your eye that attracted to the early digital noise reduction effect of smearing?
2a. Look at the red detain in the water of the beach in the unaltered image. Then look at them in the "improved" version
2b. Look at the boat image, look at the sail lines in the unaltered then the altered.
HORRBLE and clear examples. Im not sure about this program. Im not sure what it can really do, but for someone to upload this crap as examples shows that I shouldent really rely on them for a review of the product. If i were trying to sell this product id have a word or two with you about taking this whole thing down because at 70$ this isnt aimed at grandma. This is aimed at the kind of people reading this! I wouldn't buy this if you gave me 70$ to screw up my images judging by this review.
If any of you want a real program to actual improve digital imagery, here is one made by a actual professional photographer with many years experience and a sharp eye for detail;
I was a bit wary at first in investing even a bit of money in software noise filters. Most smell of snake oil. But the current crop of noise filters, especially Noise Ninja do a very good job in the right hands. The examples shown lack detail where it counts: grass, waves, leaves, hair trees, sand and texture in general. Worst still the images look like a bad job with photoshop. Show us what a pro can do with that software. reducing noise is a trade-off. I don't mind film-like grain that hides noise. Inspecting one of the images(800-coming-lrg), the histogram is compressed and the highlights are blown. For this you could have considered a tri-pod and longer exposure.
I stopped reading after seeing the Sigma images. The processed ones were smeared and flat compared to the original ones. I did not even look at the full size images.
Just look at the sea in front of the lonely boat in one of the pictures. Also the trees/bushes on the right side hill above the houses. I mean how can you be satisfied with such a loss of detail?
These heavy noise processing programs are only viable if your subject had large flat colored surfaces with little fine detail. Otherwise you end up with blocks of pixels sharing the color of grass but lacking any kind of detail that it actualy IS grass (same with stucco walls etc.).
NR has its place, but even in these web-res samples you can see an utter devastation of detail in the water shots.
NR simply replaces random noise with deterministic noise. Where possible you want it to be a subtle as possible. Most professional restoration or NR efforts are very hands on and manual and for good reason: differentiating between detail and noise is very non trivial, and even more-so with still images.
IMHO in these samples it's about as subtle as a brick to the head.
Perhaps these samples are not indicative of what the software can really do?
you have got to be kidding me. those Noiseware-processed Sigma images are TERRIBLE. it totally obliterates fine contrast and kills the resolution of the images. at these web image sizes, i clearly prefer the originals. i work closely with image editors at a professional stock photography site and believe me, those processed photos would NEVER be admitted into our library. the original "noisy" ones might be acceptable though (if the subject matter were more compelling, but that's a different matter).
we all appreciate the effort, and we know the ad dollars help the site which in turn helps us, but please - leave the digital photography reviews to the pros, and stick to topics your staff is qualified to comment on.
I didn't see a rebuttal here, but for those that compared the file sizes differ drastically which is why the quality difference that is not the case. JPEG has an incredibly hard time compressing noise, which is why when the noise is removed the file size drops by huge amounts.
Take for example the two images below. I wandered outside and took an ISO 100 TIFF with my old Olympus C5050Z. I used Photoshop to chop out a 1000x1000 section of cloud. I saved the new image as a JPEG with very good details settings. I then did a Gaussian Blur of 1.5 pixels and saved the same image (simulating a noise reduction, I usually use Neat Image but recently went through a reinstall) and saved the image with the same JPEG settings. Now each image that is 1000x1000 is either 296505 bytes, or 119464 bytes, less than half the size.
Keep in mind you can use this trick to blur things in images to make your content smaller. If you have a web cam you can put things just a shade out of focus to increase your frame rates, and also if you can find a video noise filter program for your web cam to seriously improve your frame rates with lower bandwidth.
well the JPEG compression will also create more loss of detail on the noise-reduced images, because it can compress them more aggressively. so that's a double edged sword.
Many people have been complaining about your using 'Nvidia SLI' box for the high-ISO/noise-level test, yet you don't seem to care.
I mean, what kind of reviwer uses 'glossy paper box' for noise tests?
Please use something that has details.
From time to time, you bashed internet reviews that are more like benchmarks and promised to bring field reviews. You not only failed to bring such reviews but also failed to do a simple benchmark one.
I agree that the box is a horrible test subject. dpreview's (for one example) shots of a grey patch, and then multiple crops of a high detail image are excellent for showing the level of noise AND detail that exist at each ISO setting. The box only gives you the level of noise, which is only half the story.
Was the photog shooting through the windows inside a grounded helicopter?
I've been shooting with a D3 since February, and I've been getting images that were just "not doable" below ISO 6400 without a flash - and a flash would've ruined the moment. Yes, David Black could've used a slower shutter speed and lower ISO to get the same exposure, but would he have been able to freeze the hockey action at 1/125s and ISO 800? I seriously doubt it.
I surely would like more manufacturers to implement some of Foveon's X3 layered tecnology in their future sensors, but not at the cost of reduced light sensitivity. Noiseware, Noise Ninja, Neat Image... they all do a fine job of cleaning up noise, but they do not do miracles.
The selling point of Sigma cameras are their sharp/full-of-detail images yet the reviwer thinks it's better to smear all of those in favor of low noise. That's just stupid.
If he buys a Ferrari, he'd cripple the engine for the sake of low noise, then talk about how great it is to have low noise on Ferrari.
1. That D3 ISO 6400 sample image has shutter speed of 1/1000. Even Nikon D70 or Canon A350 that are known for high noise would only show minimal noise when it's shot in the bright area.
Don't get me wrong, I'm not saying D3 low-noise/high-ISO sucks. D3, without a doubt, shows the least noise out of all the cameras in market. However, showing ISO 6400 image with shutter speed of 1/1000 and saying 'look at this low noise in ISO 6400' is simply wrong. Hell, I can even show you low-noise ISO 1600 images with D70 so you can talk about how D70 shows very little noise.
2. "However, many incorrectly criticize Sony, for instance, on their "heavy" image processing in the A350. Canon and Nikon also heavily process images in-camera; they just make slightly different choices in their processing algorithms."
That's because, as you know, Sony kills noise in favor of detail and they do it way too much. What good is 14.2MP when it's crippled; in other word, you don't get to see the advantage of 14.2MP when the ISO goes up as the noise processing algorithm kills most of the details. Again, Canon and Nikon doesn't smear images as Sony does.
3. Either your expectation is too low or my expectation is too high, but Imageware creates really smeared images just like Sonys'. Well, if you are a fan of watercolor-like or plastic-like pictures that shows the least detail, I guess Imageware is all good. But....really....pictures processed with Imageware really look terrible.
Given that, I hope you don't waste your money getting quality lenses. You'd be fine with lenses made with window glasses.
I suppose we will all have to chide Nikon for their obvious stupidity. The Hockey image was supplied by Nikon as an example ISO 6400 image in their D3 launch kit.
It completely ruins all the detail in all the photos that include bodies of water. Seriously, in the Noiseware processed one you can't see any of the ripples/waves/other details in the water anymore. Terrible!
I have been curious for some time now why you guys work with JPEGs. Hell, I would not even consider myself either a professional photographer or image re-toucher, but I can do rather well with image manipulation given enough time. Either realistic correction, restoration, or artistic.
Ideally most everyone I know that re-touches images works with RAW, or TIFF. Now, I am not trying to tell you what to do, but I suspect if you guys used one of the above formats, and THEN posted your results afterwards in JPEG . . .
Also I have to agree with the poster above me, heavy handed filtering is not good, and layers exist in Photoshop for a reason ; )
ALL SD14 images were shot in RAW and processed in Sigma Photo Pro and saved as highest quality JPEGs. We did not shoot any SD14 images using in-camera JPEG.
As pointed out in the article Noiseware does not work on RAW files - the standalone works on JPEG, TIFF, BMP, and PNG files. Even Photoshop plug-ins do not work on RAW since Camera RAW is a plug-in itself.
We have had endless discussion in-house on how to present files in photo articles, as no one, including the most repected sites, does a very good job of presenting photo files. If you read through at some well-respected sites you sometimes wonder if you can possibly be looking at the same file you are viewing. They can be calling it "sharp" or noisy" and it looks completely different on your monitor.
We are very open to suggestions on how you would like photo files presented. We try to publish the original image as well as crops, normally with EXIF data as well. This often lands us in hot water which may be why some of the large, well-respected sites don't do this with their test images. We would still prefer to disclose as much imfo as possible even if it invites stoning.
If you use the RAW converter Bibble (Pro or Lite, an excellent and VERY fast RAW converter from Bibble Labs), it comes with a lite (but free) version of Noise Ninja.
If you have a registered copy of Noise Ninja, it will unlock the full featured Noise Ninja built into Bibble. The lite and registered versions provide noise reduction on the RAW file (in a nondestructive manner like most RAW converters) prior to it being converted to a JPEG or TIFF, this allows the noise reduction to be processed on the highest quality version of the image without worrying that jpeg compression artifacts or file conversion degradation will effect the process.
And an another plus, is the registered version of Noise Ninja in Bibble does unsharp mask and performs the sharpening alongside the noise reduction properly (instead of before or after with a seperate tool) so as to retain as much detail as possible.
The only possible negative I have to say about using a dedicated noise reduction program, or one like Noise Ninja, is that they can give you almost too many controls to fine tune with. But with that ability you can also get amazing results once you understand how they work. Luckily there is an auto profile and typical default settings beginners can work with.
Thank you for your suggestion. I had forgotten that Bibble processes most RAW images and includes a lite version of Ninja Pro. Bibble certainly sounds like a useful tool for Pentax K20D and Sony A350/A700 users. Unfortunately Bibble does not support Sigma RAW so it would not be as useful for the small group of SD14 Foveon users.
I certainly intend to give Bibble another look in the near future.
Well I have to say that aguilpa1 had a point that you should have tried to remove croma noise when imported into CS3 or whatever application you used. HOWEVER, since this is a software noise removal review, that would not be true to what I think you guys are trying to show us. i.e. 'we' want to see how this filter/program does on its own. That said, it would be hard to say how this should have been done from every single readers view point.
The best way I can think of doing this (right now) would be first to:
1) Import RAW images into CS3 making sure all import settings remain the same.
2) Check Image-> mode bits/channel is set to 8bits/channel. Save an original at this point for showing the original image.
3) Go to Image -> Mode, and set bits/channel to 16bits/channel.
4) make multiple layer copies of the original, applying a few different levels of noise removal on each layer. Some kind of consistency will be needed if comparing filters of course.
5) Mark all layers but the original un-touched layers as hidden, find a good noisy area, and select it.
6) Set each individual layer as visible(one at a time,and hiding the others), and copy paste into a new image within photoshop. Making sure that the new->file bit/channel is of course 16bit/channel. This should also be done with one or more of the re-touches to be shown as a full sized image, after deselecting your selection of course.
7) Go to Image-> mode once more for each image and change bit/channel to 8bit/channel before saving to JPEG.
So now you could have the original to show, a full re-touched image, and several crops of a noisy area with different filter strength applied. Using 16bit/channel is something I have become used to while editing images to help ensure that colors do not get squashed too much while editing images, even if saving out to JPEG(which I usually do). In some cases it may be an un-needed step, but I never know if I will ever make prints of my work or not, AND this is something that really does not take much work.
Anyways, some food for thought, and maybe you already do something similar to this already ?
This is kind of a double edged sword, but if I went purely by the automatic settings of this filter/application, I would have to say that this filter is junk. In reality though this just means that the automatic settings for this application are way too harsh, but the filter/application *could* be very good when applied correctly. Now if something like i kind of laid out above was done, there would be more information for someone like me to make a better decision.
I agree with the folks that say the use of noiseware with the Sigma pictures is horrible. It represents either the default or very heavy handed use of the filter. I use noiseware as a plugin with Photoshop CS3 but not always and carefully. If you don't use it carefully it will ruin your picture and turn it into a watercolor like what you have in those Sigma pictures. If you guys had taken your original SD14 pics in RAW and done some chroma noise reduction to remove the more offensive color blotchiness first than run your noiseware afterwards in PS without being so aggressive on your settings, preferably on a duplicate layer so you could (erase) the noise filtering on your more detailed subjects you would have gotten better results.
As stated in the article all Noiseware processing used the AUTO mode, and should be considered the worst case or baseline of Noiseware use. As we pointed out there are many more controls available for custom tweaking and adjusting to parameters important to the user.
The idea of this article was to expose people to the idea of noise reduction post processing. It may surprise some of the "experts" commenting on this article that most of our readers who use digital cameras do not even know programs like Noiseware exist or exactly what they do. This article was targeted at that audience.
The quality of digital photography analysis on this site is not up to par with the rest of the site. Serious readers choose sites like DPReview for their photo information. I wouldn't go to DPReview for info on the latest chipsets because they have the good sense to stay away from topics they aren't experts in. I wish Anandtech would do the same.
Didn't you already have a whining post on the first page? Not all of find DPRReview all that useful. I went to that site before I bought my Canon and to be honest, it was no help at all. I ended up borrowing a friends camera, liked it, and then bought the same one. Here is the section on noise reduction at DPR Reviews web site:
Is it just me, or is the whole point of that DPreview article just a way to try and get people to purchase the training course off of the linked website?
I also have to wonder at the number of "new" users commenting on the photography articles. I recognize some of the names, but the vast majority appear to be people that have only registered to complain about this one article. Anyone want to bet that several of these "anonymous" users happen to write for other websites? "Bitch bitch bitch, moan moan moan, you guys aren't like DPreview, which is just awesome! [link]"
-----
FYI, this is more a comment on the original poster and is not meant as an attack on kmmatney or yyrkoon, who are at least regulars here as I see posts from them all the time. It appears half of the commenters are relative unkowns. As for me, I tend to lurk more than post, but if you look around at previous articles, I've been here for about a year now and appreciate the variety of articles. It would be interesting to see a comparative review of several of the top noise removal programs, including customized results rather than just blanket "remove noise everywhere" auto settings.
It would be nice if people like you would stop telling others what they should / should not do. If you do not like their 'photography' article(which I would probably consider this more of a retouching article), then do not read them.
the big sensor helps the D3 considerably at high ISO but there is a great deal of luminance noise in the picture that is characteristic of Nikon cameras that focus mainly on chroma noise reduction. This makes the picture grainy but not blotchy colored (chroma noise)which is more acceptable but its not a smooth toned image. Most professional digital SLR's can achieve similar results although with luminance noise reduction applied to RAW in post processing and not in camera as Nikon does.
To make a clear example of the noise processing, you should show a cutout of the main picture at full resolution. At the image size shown, it is not possible to see the noise effect. Also, as mharris points out, the JPEG compression kills all in the comparison.
I also think you can check other Noise Reduction tools, like Noise Ninja.
1:1 crops from a larger image, with and without Noiseware processing, are shown on pages 7, 8, and 9. Clicking on any smaller image brings up the full image, with and without processing.
The full ISO range was included for comparison, although we would never recommend using software noise reduction on lower ISO images. As stated in the article all noise reduction is a balancing act between image detail and softening/smearing. If you don't like the Noiseware auto processing you caa fortunately dial in other presets with different NR parameters or create your own based on parameters that are most important to you as a photographer.
I found Noiseware to be an extremely useful and flexible tool, but there are many other noise reduction programs and plug-ins available.
We did try Noise Ninja on the SD14 images. The results were very poor. This article was not conceived as a comparison of all the available software noise reduction programs or plug-ins. The idea was to show that post-processing can sometimes salvage high-noise images.
Noiseware also worked best on the images that needed the most help. Used sparingly software noise reduction like Noiseware can be a useful tool. This is particularly true with images from cameras like the SD14, K20D, and A350. At the other end of the spectrum is a camera like the Nikon D3, with exceptionally low noise at high ISO.
I just tried running Noise Ninja on your original with the Sigma SD14 profile from picturecodes homepage, and i must say that it gives a much better result than what you show here.
I don't understand why you say that it gives poor results. With default settings it gives much more detail, and if i want the "plastic look" that you show in your examples it's just a matter of cranking up filter strength and smoothness...
I would have to agree with others here in that the reduced noise images look terrible in a lot of cases - especially the shots with water in them. However, some people here may fancy themselves professional photographers, but its obvious they know very little about image retouching. Point #1, if you care about an image you NEVER use a filters auto settings UNLESS they just so happen to work the best(very rare). Point #2 - a professional image retoucher will hardly ever use a single layer when 'correcting' an image. You keep the original layer, apply separate filters to copies of the original, and blend as appropriate.
I would however have to disagree that noise ninja would do a worse job here. Noise ninja is known as one of the best noise filters out there, and from personal experience I'd have to agree. It however is *not* the end all be all of image retouching, and does require some finessing. Either way, running noise filters on a JPEG image is NOT the way to go . . .
Oh, and I just noticed that a lot of that 'smearing' people were talking about was actually done in camera. Take another look if you do not believe me.
If you look at the photos, you'll notice that the unfiltered photos are nearly 10x bigger than the filtered photos. So the loss of details is due to a ridiculously low JPEG compression, not the noise filter.
"If you look at the photos, you'll notice that the unfiltered photos are nearly 10x bigger than the filtered photos. So the loss of details is due to a ridiculously low JPEG compression, not the noise filter."
Not it's not. Would have thought this was obvious... the file size is lower because of the lower detail in the noise reduction filtered images. This is because of how JPEG compression works. It's the same with any image after noise reduction has been used, it nearly always produces smaller files sizes 'cause of the resulting lower detail.
Are you actually serious? You can't be, look at the amount of detail lost in the photos done by Noiseware. You lose all the detail and the photos look very, very smeared.
And compare the 3D to a 1DMK3 or 5D, this comparison is useless.
There is no doubt the Nikon D3 is the low-noise champion, but we weren't comparing it to other PRO cameras like the ID Mark III. The comments were comparing relative photosire size across the spectrum of digital SLR sensors.
We do agree the ID Mk III at 10 megaixel with a 1.3X (APS-H) crop factor is more directly comparable in photosite size to the D3. The 5D at full-frame 12.8 megapixels is certainly comparable in photosite size if not speed or high ISO performance.
The 5D is only enabled on-camera to ISO 3200, but plenty of users use effectively higher ISOs by deliberately underexposing then pushing the exposure in post-process with decent results.
Are you guys blind? some of thos vacation shots are terrible after the noise reduction. I'd rather stick with the noise in some of them. The beach shot looks ok in the foreground, but as soon as you hit the water its all water paiting.
the sail boat on the water is one of the worst, in the original shot you see waves breaking, and caps. The after processing shot destroy's the feel of the water.
IMHO, alot of the pictures looks better just shrunk (which filter the nose by itself), then they did after noise removal.
Same here, there noise reduction seriously destroys the detail present in the images. A lot of the areas of high contrast is destroyed after the noise reduction resulting in smearing.
I'm neither a professional or even a prosumer, but it was quite noticeable to me that the details were significantly reduced with the noise reduction where I definitely would not consider this software.
I also could not tell the difference in the low ISO shots for the noise reduction although I am on a 6-bit TN LCD panel so that could probably be the reason. Did anyone else see a difference in the low-ISO shots?
Noiseware on images from any of the superzooms? I often find myself to be a little tentative of taking higher ISO setting shots on my superzoom because of the pronounced noise levels. I would be interested to see some results from some of the different brands' superzoom models.
All in all, I've been pretty disappointed in the Anandtech articles about cameras and image manipulation, the articles gives just a rough idea what noise reduction software does, but does not go into settings, original noise levels, or RAW vs JPEG noise levels (lens, apeture, shutter speed, post processing software).
If you're a beginner photographer, or just care about software please read. But if you are seriously interested in photography please try a different site or get your hands on a real book.
I think I'm just disappointed about having an amateur write a review as opposed to a pro.
agreed. i criticized some earlier reviews, but was holding out for the quality of writing and analysis to improve. that is clearly not happening. even the article synopsis in my anandtech RSS feed was completely wrong:
"Not all JPEG processing is created equal, but Noiseware can often fix what camera JPEG processing leaves undone...."
i was truly puzzled by this. noise reduction has NOTHING to do with JPEG compression - that's like saying, "not all gasoline is created equal... let's see how this motor oil performs."
but i guess if it brings in the ad dollars, it's mission accomplished for the site... even though ultimately it's doing the readers a disservice.
You are technically correct, but it appears you are straining to make your point. Certainly you, and everyone else, understands that noise reduction is one of the processes that usually happens during in-camera "JPEG" processing or during software conversion from RAW to a finished image format which is normally JPEG, but could be TIFF or even another format.
In the interest of being more precise in the description I have changed it to:
"Not all image processing is created equal, but Noiseware can often reduce noise that in-camera or software processing leaves behind."
Eh, I'd say that your quoted caption is an accurate description of the article. There's on-chip noise reduction that is performed before the RAW file is created, then there's software noise reduction that is performed as part of the image processing engine in-camera. The output of that image processing engine is usually JPEG or TIFF, and the RAW preview image.
So, if a camera's processing engine isn't running noise reduction before writing to JPEG, then this software would finish the noise reduction stage of post-processing.
As for the software itself, I don't know about the images at full magnification, but the small previews shown indicate fairly heavy smearing. Lightroom does less smearing, and seems to be extremely capable in my experience at chroma noise reduction, and good enough at luminance.
As for writing quality, I can't do better, so I won't criticize.
it is NOT an accurate description. image processing has NOTHING to do with the final output format of the file. you said it yourself - the output could be TIFF, so what does it have to do with "JPEG processing"? call it a nit if you will, but anandtech is a technical publication, so such fundamental inaccuracies should be considered anathema.
and saying you can't criticize because you can't do any better yourself seems symptomatic of the sort of mediocrity malaise that results in sub-standard institutions like the US Postal Service. ;) if we expected everything in the world to be no better than what we ourselves can accomplish, the world would be a very sad place...
Personally I think there's too much smearing going on, to the point that I like the original picture better. The first picture for example has quite a loss of detail on the windows of the buildings and on the slope of the mountain. In the second picture all the ripples are gone and the boat seems to float above instead of in the water. All non-macro low-iso shots show an incredible amount of smearing (stones on the stairway, vegetation on every background mountain, ripples in water). I don't know how anyone would consider those filtered low-ISO shots an improvement.
I've long been impressed w/ Canon's low noise compared to other makes. Up 'til now, I've always heard/assumed that it was the sensor. So is it possible to determine to what exactly the noise reduction is attributable in many cameras? Is it due to an algorithm or is it really the sensor? I guess we'll never know, since the images are already 'processed' in RAW, will we?
Canon states that the only noise reduction done to their RAW images is dark noise subtraction. They have plenty of whitepapers on their technology floating around.
Thanks for the article, noise is certainly always a problem in digital photography. I've messed around with different noise-reduction methods via Photoshop itself, as well as a plugin called Noise Ninja which usually worked pretty well for myself. All in all it takes some patience and tinkering to ever get it perfect. :) Perhaps I shall look into the mentioned program as well.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
61 Comments
Back to Article
aeternitas - Saturday, August 2, 2008 - link
http://www.fredmiranda.com/IS/">http://www.fredmiranda.com/IS/aeternitas - Saturday, August 2, 2008 - link
1. When doing comparison shots, have a mouse over load the second image and a mouseoff show the first. Or simply have one image and a list of links that dynamical and instantly change that image so comparisons can be made in detail. This is what any self respecting site that has to so with image comparisons would do. This screams "early 90s"2. Though one is usually important, its not here for the even worse reason that the originals were all far better than the filtered images! Why bother uploading this rubbish? Did you even bother to compair or is your eye that attracted to the early digital noise reduction effect of smearing?
2a. Look at the red detain in the water of the beach in the unaltered image. Then look at them in the "improved" version
2b. Look at the boat image, look at the sail lines in the unaltered then the altered.
HORRBLE and clear examples. Im not sure about this program. Im not sure what it can really do, but for someone to upload this crap as examples shows that I shouldent really rely on them for a review of the product. If i were trying to sell this product id have a word or two with you about taking this whole thing down because at 70$ this isnt aimed at grandma. This is aimed at the kind of people reading this! I wouldn't buy this if you gave me 70$ to screw up my images judging by this review.
If any of you want a real program to actual improve digital imagery, here is one made by a actual professional photographer with many years experience and a sharp eye for detail;
http://www.fredmiranda.com/software/">http://www.fredmiranda.com/software/
elmerFudge - Wednesday, July 30, 2008 - link
I was a bit wary at first in investing even a bit of money in software noise filters. Most smell of snake oil. But the current crop of noise filters, especially Noise Ninja do a very good job in the right hands. The examples shown lack detail where it counts: grass, waves, leaves, hair trees, sand and texture in general. Worst still the images look like a bad job with photoshop. Show us what a pro can do with that software. reducing noise is a trade-off. I don't mind film-like grain that hides noise. Inspecting one of the images(800-coming-lrg), the histogram is compressed and the highlights are blown. For this you could have considered a tri-pod and longer exposure.haplo602 - Tuesday, July 29, 2008 - link
I stopped reading after seeing the Sigma images. The processed ones were smeared and flat compared to the original ones. I did not even look at the full size images.Just look at the sea in front of the lonely boat in one of the pictures. Also the trees/bushes on the right side hill above the houses. I mean how can you be satisfied with such a loss of detail?
These heavy noise processing programs are only viable if your subject had large flat colored surfaces with little fine detail. Otherwise you end up with blocks of pixels sharing the color of grass but lacking any kind of detail that it actualy IS grass (same with stucco walls etc.).
royalcrown - Tuesday, July 29, 2008 - link
I realize the pace of exciting hardware has slowed since the 90's...but all these camera, Ipod, Iphone..yada yada, E3 reviews are BOOOOORING !! Did I mention boring yet........zzzzzzzzzzzzzzzzzzzzz ?
I only come here once a week at most because all this gadget crap is a snoozefest. No offense because I suppose it's slim pickings, just my 2 cents !!
ianken - Tuesday, July 29, 2008 - link
NR has its place, but even in these web-res samples you can see an utter devastation of detail in the water shots.NR simply replaces random noise with deterministic noise. Where possible you want it to be a subtle as possible. Most professional restoration or NR efforts are very hands on and manual and for good reason: differentiating between detail and noise is very non trivial, and even more-so with still images.
IMHO in these samples it's about as subtle as a brick to the head.
Perhaps these samples are not indicative of what the software can really do?
n4bby - Monday, July 28, 2008 - link
you have got to be kidding me. those Noiseware-processed Sigma images are TERRIBLE. it totally obliterates fine contrast and kills the resolution of the images. at these web image sizes, i clearly prefer the originals. i work closely with image editors at a professional stock photography site and believe me, those processed photos would NEVER be admitted into our library. the original "noisy" ones might be acceptable though (if the subject matter were more compelling, but that's a different matter).we all appreciate the effort, and we know the ad dollars help the site which in turn helps us, but please - leave the digital photography reviews to the pros, and stick to topics your staff is qualified to comment on.
Traciatim - Monday, July 28, 2008 - link
I didn't see a rebuttal here, but for those that compared the file sizes differ drastically which is why the quality difference that is not the case. JPEG has an incredibly hard time compressing noise, which is why when the noise is removed the file size drops by huge amounts.Take for example the two images below. I wandered outside and took an ISO 100 TIFF with my old Olympus C5050Z. I used Photoshop to chop out a 1000x1000 section of cloud. I saved the new image as a JPEG with very good details settings. I then did a Gaussian Blur of 1.5 pixels and saved the same image (simulating a noise reduction, I usually use Neat Image but recently went through a reinstall) and saved the image with the same JPEG settings. Now each image that is 1000x1000 is either 296505 bytes, or 119464 bytes, less than half the size.
Keep in mind you can use this trick to blur things in images to make your content smaller. If you have a web cam you can put things just a shade out of focus to increase your frame rates, and also if you can find a video noise filter program for your web cam to seriously improve your frame rates with lower bandwidth.
Image 1 (Bo Blur):
http://i201.photobucket.com/albums/aa229/Traciatim...">http://i201.photobucket.com/albums/aa229/Traciatim...
Image 2 (Blur):
http://i201.photobucket.com/albums/aa229/Traciatim...">http://i201.photobucket.com/albums/aa229/Traciatim...
haplo602 - Tuesday, July 29, 2008 - link
well the JPEG compression will also create more loss of detail on the noise-reduced images, because it can compress them more aggressively. so that's a double edged sword.Deadtrees - Monday, July 28, 2008 - link
Many people have been complaining about your using 'Nvidia SLI' box for the high-ISO/noise-level test, yet you don't seem to care.I mean, what kind of reviwer uses 'glossy paper box' for noise tests?
Please use something that has details.
From time to time, you bashed internet reviews that are more like benchmarks and promised to bring field reviews. You not only failed to bring such reviews but also failed to do a simple benchmark one.
Please...Anandtech deserves better....
GoSharks - Thursday, July 31, 2008 - link
I agree that the box is a horrible test subject. dpreview's (for one example) shots of a grey patch, and then multiple crops of a high detail image are excellent for showing the level of noise AND detail that exist at each ISO setting. The box only gives you the level of noise, which is only half the story.marokero - Monday, July 28, 2008 - link
Was the photog shooting through the windows inside a grounded helicopter?I've been shooting with a D3 since February, and I've been getting images that were just "not doable" below ISO 6400 without a flash - and a flash would've ruined the moment. Yes, David Black could've used a slower shutter speed and lower ISO to get the same exposure, but would he have been able to freeze the hockey action at 1/125s and ISO 800? I seriously doubt it.
I surely would like more manufacturers to implement some of Foveon's X3 layered tecnology in their future sensors, but not at the cost of reduced light sensitivity. Noiseware, Noise Ninja, Neat Image... they all do a fine job of cleaning up noise, but they do not do miracles.
peroni - Monday, July 28, 2008 - link
Sorry Wesley but this article should be withdrawned, the quality is not on par with the rest of this site articles.p.s.
In nearly all the pictures you have posted the originals look better.
Deadtrees - Monday, July 28, 2008 - link
The selling point of Sigma cameras are their sharp/full-of-detail images yet the reviwer thinks it's better to smear all of those in favor of low noise. That's just stupid.If he buys a Ferrari, he'd cripple the engine for the sake of low noise, then talk about how great it is to have low noise on Ferrari.
pepsimax2k - Monday, July 28, 2008 - link
is that dido on the main page pic?woohoo *wins award for least geeky anandtech post in, ohh, 5 minutes?*
Deadtrees - Monday, July 28, 2008 - link
1. That D3 ISO 6400 sample image has shutter speed of 1/1000. Even Nikon D70 or Canon A350 that are known for high noise would only show minimal noise when it's shot in the bright area.Don't get me wrong, I'm not saying D3 low-noise/high-ISO sucks. D3, without a doubt, shows the least noise out of all the cameras in market. However, showing ISO 6400 image with shutter speed of 1/1000 and saying 'look at this low noise in ISO 6400' is simply wrong. Hell, I can even show you low-noise ISO 1600 images with D70 so you can talk about how D70 shows very little noise.
2. "However, many incorrectly criticize Sony, for instance, on their "heavy" image processing in the A350. Canon and Nikon also heavily process images in-camera; they just make slightly different choices in their processing algorithms."
That's because, as you know, Sony kills noise in favor of detail and they do it way too much. What good is 14.2MP when it's crippled; in other word, you don't get to see the advantage of 14.2MP when the ISO goes up as the noise processing algorithm kills most of the details. Again, Canon and Nikon doesn't smear images as Sony does.
3. Either your expectation is too low or my expectation is too high, but Imageware creates really smeared images just like Sonys'. Well, if you are a fan of watercolor-like or plastic-like pictures that shows the least detail, I guess Imageware is all good. But....really....pictures processed with Imageware really look terrible.
Given that, I hope you don't waste your money getting quality lenses. You'd be fine with lenses made with window glasses.
Wesley Fink - Monday, July 28, 2008 - link
I suppose we will all have to chide Nikon for their obvious stupidity. The Hockey image was supplied by Nikon as an example ISO 6400 image in their D3 launch kit.Deadtrees - Monday, July 28, 2008 - link
No, the problem is you saying 'Look at this image! it's ISO 6400 and shows so little noise!'Some1ne - Monday, July 28, 2008 - link
It completely ruins all the detail in all the photos that include bodies of water. Seriously, in the Noiseware processed one you can't see any of the ripples/waves/other details in the water anymore. Terrible!Ratinator - Monday, July 28, 2008 - link
Has anyone heard anything about when Kodak is going to release their new filter that was discussed about a year ago mentioned here http://www.dpreview.com/news/0706/07061401kodakhig...">http://www.dpreview.com/news/0706/07061401kodakhig...yyrkoon - Monday, July 28, 2008 - link
I have been curious for some time now why you guys work with JPEGs. Hell, I would not even consider myself either a professional photographer or image re-toucher, but I can do rather well with image manipulation given enough time. Either realistic correction, restoration, or artistic.Ideally most everyone I know that re-touches images works with RAW, or TIFF. Now, I am not trying to tell you what to do, but I suspect if you guys used one of the above formats, and THEN posted your results afterwards in JPEG . . .
Also I have to agree with the poster above me, heavy handed filtering is not good, and layers exist in Photoshop for a reason ; )
Wesley Fink - Monday, July 28, 2008 - link
ALL SD14 images were shot in RAW and processed in Sigma Photo Pro and saved as highest quality JPEGs. We did not shoot any SD14 images using in-camera JPEG.As pointed out in the article Noiseware does not work on RAW files - the standalone works on JPEG, TIFF, BMP, and PNG files. Even Photoshop plug-ins do not work on RAW since Camera RAW is a plug-in itself.
We have had endless discussion in-house on how to present files in photo articles, as no one, including the most repected sites, does a very good job of presenting photo files. If you read through at some well-respected sites you sometimes wonder if you can possibly be looking at the same file you are viewing. They can be calling it "sharp" or noisy" and it looks completely different on your monitor.
We are very open to suggestions on how you would like photo files presented. We try to publish the original image as well as crops, normally with EXIF data as well. This often lands us in hot water which may be why some of the large, well-respected sites don't do this with their test images. We would still prefer to disclose as much imfo as possible even if it invites stoning.
VirtualMirage - Tuesday, July 29, 2008 - link
If you use the RAW converter Bibble (Pro or Lite, an excellent and VERY fast RAW converter from Bibble Labs), it comes with a lite (but free) version of Noise Ninja.If you have a registered copy of Noise Ninja, it will unlock the full featured Noise Ninja built into Bibble. The lite and registered versions provide noise reduction on the RAW file (in a nondestructive manner like most RAW converters) prior to it being converted to a JPEG or TIFF, this allows the noise reduction to be processed on the highest quality version of the image without worrying that jpeg compression artifacts or file conversion degradation will effect the process.
And an another plus, is the registered version of Noise Ninja in Bibble does unsharp mask and performs the sharpening alongside the noise reduction properly (instead of before or after with a seperate tool) so as to retain as much detail as possible.
The only possible negative I have to say about using a dedicated noise reduction program, or one like Noise Ninja, is that they can give you almost too many controls to fine tune with. But with that ability you can also get amazing results once you understand how they work. Luckily there is an auto profile and typical default settings beginners can work with.
Paul
Wesley Fink - Tuesday, July 29, 2008 - link
Paul -Thank you for your suggestion. I had forgotten that Bibble processes most RAW images and includes a lite version of Ninja Pro. Bibble certainly sounds like a useful tool for Pentax K20D and Sony A350/A700 users. Unfortunately Bibble does not support Sigma RAW so it would not be as useful for the small group of SD14 Foveon users.
I certainly intend to give Bibble another look in the near future.
yyrkoon - Monday, July 28, 2008 - link
Well I have to say that aguilpa1 had a point that you should have tried to remove croma noise when imported into CS3 or whatever application you used. HOWEVER, since this is a software noise removal review, that would not be true to what I think you guys are trying to show us. i.e. 'we' want to see how this filter/program does on its own. That said, it would be hard to say how this should have been done from every single readers view point.The best way I can think of doing this (right now) would be first to:
1) Import RAW images into CS3 making sure all import settings remain the same.
2) Check Image-> mode bits/channel is set to 8bits/channel. Save an original at this point for showing the original image.
3) Go to Image -> Mode, and set bits/channel to 16bits/channel.
4) make multiple layer copies of the original, applying a few different levels of noise removal on each layer. Some kind of consistency will be needed if comparing filters of course.
5) Mark all layers but the original un-touched layers as hidden, find a good noisy area, and select it.
6) Set each individual layer as visible(one at a time,and hiding the others), and copy paste into a new image within photoshop. Making sure that the new->file bit/channel is of course 16bit/channel. This should also be done with one or more of the re-touches to be shown as a full sized image, after deselecting your selection of course.
7) Go to Image-> mode once more for each image and change bit/channel to 8bit/channel before saving to JPEG.
So now you could have the original to show, a full re-touched image, and several crops of a noisy area with different filter strength applied. Using 16bit/channel is something I have become used to while editing images to help ensure that colors do not get squashed too much while editing images, even if saving out to JPEG(which I usually do). In some cases it may be an un-needed step, but I never know if I will ever make prints of my work or not, AND this is something that really does not take much work.
Anyways, some food for thought, and maybe you already do something similar to this already ?
This is kind of a double edged sword, but if I went purely by the automatic settings of this filter/application, I would have to say that this filter is junk. In reality though this just means that the automatic settings for this application are way too harsh, but the filter/application *could* be very good when applied correctly. Now if something like i kind of laid out above was done, there would be more information for someone like me to make a better decision.
Maybe someone else has a better idea ?
aguilpa1 - Monday, July 28, 2008 - link
I agree with the folks that say the use of noiseware with the Sigma pictures is horrible. It represents either the default or very heavy handed use of the filter. I use noiseware as a plugin with Photoshop CS3 but not always and carefully. If you don't use it carefully it will ruin your picture and turn it into a watercolor like what you have in those Sigma pictures. If you guys had taken your original SD14 pics in RAW and done some chroma noise reduction to remove the more offensive color blotchiness first than run your noiseware afterwards in PS without being so aggressive on your settings, preferably on a duplicate layer so you could (erase) the noise filtering on your more detailed subjects you would have gotten better results.Wesley Fink - Monday, July 28, 2008 - link
As stated in the article all Noiseware processing used the AUTO mode, and should be considered the worst case or baseline of Noiseware use. As we pointed out there are many more controls available for custom tweaking and adjusting to parameters important to the user.The idea of this article was to expose people to the idea of noise reduction post processing. It may surprise some of the "experts" commenting on this article that most of our readers who use digital cameras do not even know programs like Noiseware exist or exactly what they do. This article was targeted at that audience.
studio1 - Monday, July 28, 2008 - link
The quality of digital photography analysis on this site is not up to par with the rest of the site. Serious readers choose sites like DPReview for their photo information. I wouldn't go to DPReview for info on the latest chipsets because they have the good sense to stay away from topics they aren't experts in. I wish Anandtech would do the same.kmmatney - Monday, July 28, 2008 - link
Didn't you already have a whining post on the first page? Not all of find DPRReview all that useful. I went to that site before I bought my Canon and to be honest, it was no help at all. I ended up borrowing a friends camera, liked it, and then bought the same one. Here is the section on noise reduction at DPR Reviews web site:http://www.dpreview.com/learn/?/Glossary/Digital_I...">http://www.dpreview.com/learn/?/Glossary/Digital_I...
strikeback03 - Monday, July 28, 2008 - link
The glossary section of DPR is not great and out-of-date for the most part. Reading the reviews or the forums will give you much better information.whatthehey - Monday, July 28, 2008 - link
Is it just me, or is the whole point of that DPreview article just a way to try and get people to purchase the training course off of the linked website?I also have to wonder at the number of "new" users commenting on the photography articles. I recognize some of the names, but the vast majority appear to be people that have only registered to complain about this one article. Anyone want to bet that several of these "anonymous" users happen to write for other websites? "Bitch bitch bitch, moan moan moan, you guys aren't like DPreview, which is just awesome! [link]"
-----
FYI, this is more a comment on the original poster and is not meant as an attack on kmmatney or yyrkoon, who are at least regulars here as I see posts from them all the time. It appears half of the commenters are relative unkowns. As for me, I tend to lurk more than post, but if you look around at previous articles, I've been here for about a year now and appreciate the variety of articles. It would be interesting to see a comparative review of several of the top noise removal programs, including customized results rather than just blanket "remove noise everywhere" auto settings.
yyrkoon - Monday, July 28, 2008 - link
It would be nice if people like you would stop telling others what they should / should not do. If you do not like their 'photography' article(which I would probably consider this more of a retouching article), then do not read them.aguilpa1 - Monday, July 28, 2008 - link
the big sensor helps the D3 considerably at high ISO but there is a great deal of luminance noise in the picture that is characteristic of Nikon cameras that focus mainly on chroma noise reduction. This makes the picture grainy but not blotchy colored (chroma noise)which is more acceptable but its not a smooth toned image. Most professional digital SLR's can achieve similar results although with luminance noise reduction applied to RAW in post processing and not in camera as Nikon does.gmarcus - Monday, July 28, 2008 - link
To make a clear example of the noise processing, you should show a cutout of the main picture at full resolution. At the image size shown, it is not possible to see the noise effect. Also, as mharris points out, the JPEG compression kills all in the comparison.I also think you can check other Noise Reduction tools, like Noise Ninja.
Wesley Fink - Monday, July 28, 2008 - link
1:1 crops from a larger image, with and without Noiseware processing, are shown on pages 7, 8, and 9. Clicking on any smaller image brings up the full image, with and without processing.The full ISO range was included for comparison, although we would never recommend using software noise reduction on lower ISO images. As stated in the article all noise reduction is a balancing act between image detail and softening/smearing. If you don't like the Noiseware auto processing you caa fortunately dial in other presets with different NR parameters or create your own based on parameters that are most important to you as a photographer.
I found Noiseware to be an extremely useful and flexible tool, but there are many other noise reduction programs and plug-ins available.
Wesley Fink - Monday, July 28, 2008 - link
We did try Noise Ninja on the SD14 images. The results were very poor. This article was not conceived as a comparison of all the available software noise reduction programs or plug-ins. The idea was to show that post-processing can sometimes salvage high-noise images.Noiseware also worked best on the images that needed the most help. Used sparingly software noise reduction like Noiseware can be a useful tool. This is particularly true with images from cameras like the SD14, K20D, and A350. At the other end of the spectrum is a camera like the Nikon D3, with exceptionally low noise at high ISO.
sejer - Monday, July 28, 2008 - link
I just tried running Noise Ninja on your original with the Sigma SD14 profile from picturecodes homepage, and i must say that it gives a much better result than what you show here.I don't understand why you say that it gives poor results. With default settings it gives much more detail, and if i want the "plastic look" that you show in your examples it's just a matter of cranking up filter strength and smoothness...
yyrkoon - Monday, July 28, 2008 - link
I would have to agree with others here in that the reduced noise images look terrible in a lot of cases - especially the shots with water in them. However, some people here may fancy themselves professional photographers, but its obvious they know very little about image retouching. Point #1, if you care about an image you NEVER use a filters auto settings UNLESS they just so happen to work the best(very rare). Point #2 - a professional image retoucher will hardly ever use a single layer when 'correcting' an image. You keep the original layer, apply separate filters to copies of the original, and blend as appropriate.I would however have to disagree that noise ninja would do a worse job here. Noise ninja is known as one of the best noise filters out there, and from personal experience I'd have to agree. It however is *not* the end all be all of image retouching, and does require some finessing. Either way, running noise filters on a JPEG image is NOT the way to go . . .
yyrkoon - Monday, July 28, 2008 - link
Oh, and I just noticed that a lot of that 'smearing' people were talking about was actually done in camera. Take another look if you do not believe me.strikeback03 - Monday, July 28, 2008 - link
I bought Neat Image primarily because when I was shopping they were the only one to offer a useful trial. I'm happy with the results.mharris - Monday, July 28, 2008 - link
If you look at the photos, you'll notice that the unfiltered photos are nearly 10x bigger than the filtered photos. So the loss of details is due to a ridiculously low JPEG compression, not the noise filter.B3an - Friday, August 1, 2008 - link
"If you look at the photos, you'll notice that the unfiltered photos are nearly 10x bigger than the filtered photos. So the loss of details is due to a ridiculously low JPEG compression, not the noise filter."Not it's not. Would have thought this was obvious... the file size is lower because of the lower detail in the noise reduction filtered images. This is because of how JPEG compression works. It's the same with any image after noise reduction has been used, it nearly always produces smaller files sizes 'cause of the resulting lower detail.
Baviaan - Monday, July 28, 2008 - link
Are you actually serious? You can't be, look at the amount of detail lost in the photos done by Noiseware. You lose all the detail and the photos look very, very smeared.And compare the 3D to a 1DMK3 or 5D, this comparison is useless.
Wesley Fink - Monday, July 28, 2008 - link
There is no doubt the Nikon D3 is the low-noise champion, but we weren't comparing it to other PRO cameras like the ID Mark III. The comments were comparing relative photosire size across the spectrum of digital SLR sensors.We do agree the ID Mk III at 10 megaixel with a 1.3X (APS-H) crop factor is more directly comparable in photosite size to the D3. The 5D at full-frame 12.8 megapixels is certainly comparable in photosite size if not speed or high ISO performance.
strikeback03 - Monday, July 28, 2008 - link
The 5D is only enabled on-camera to ISO 3200, but plenty of users use effectively higher ISOs by deliberately underexposing then pushing the exposure in post-process with decent results.michal1980 - Monday, July 28, 2008 - link
I agree with others.Are you guys blind? some of thos vacation shots are terrible after the noise reduction. I'd rather stick with the noise in some of them. The beach shot looks ok in the foreground, but as soon as you hit the water its all water paiting.
the sail boat on the water is one of the worst, in the original shot you see waves breaking, and caps. The after processing shot destroy's the feel of the water.
IMHO, alot of the pictures looks better just shrunk (which filter the nose by itself), then they did after noise removal.
Jedi2155 - Monday, July 28, 2008 - link
Same here, there noise reduction seriously destroys the detail present in the images. A lot of the areas of high contrast is destroyed after the noise reduction resulting in smearing.I'm neither a professional or even a prosumer, but it was quite noticeable to me that the details were significantly reduced with the noise reduction where I definitely would not consider this software.
I also could not tell the difference in the low ISO shots for the noise reduction although I am on a 6-bit TN LCD panel so that could probably be the reason. Did anyone else see a difference in the low-ISO shots?
B3an - Monday, July 28, 2008 - link
From the images it's hardly any better than Photoshops noise reduction filter. Messing around with the settings i can get very nearly as good results.And like all noise reduction filters it makes the image lose detail, messes up some colours, and sometimes over sharpens the edges.
I dont think this is good software or worth the money.
eetnoyer - Monday, July 28, 2008 - link
Noiseware on images from any of the superzooms? I often find myself to be a little tentative of taking higher ISO setting shots on my superzoom because of the pronounced noise levels. I would be interested to see some results from some of the different brands' superzoom models.Thanks
guitargeek27 - Monday, July 28, 2008 - link
All in all, I've been pretty disappointed in the Anandtech articles about cameras and image manipulation, the articles gives just a rough idea what noise reduction software does, but does not go into settings, original noise levels, or RAW vs JPEG noise levels (lens, apeture, shutter speed, post processing software).If you're a beginner photographer, or just care about software please read. But if you are seriously interested in photography please try a different site or get your hands on a real book.
I think I'm just disappointed about having an amateur write a review as opposed to a pro.
n4bby - Monday, July 28, 2008 - link
agreed. i criticized some earlier reviews, but was holding out for the quality of writing and analysis to improve. that is clearly not happening. even the article synopsis in my anandtech RSS feed was completely wrong:"Not all JPEG processing is created equal, but Noiseware can often fix what camera JPEG processing leaves undone...."
i was truly puzzled by this. noise reduction has NOTHING to do with JPEG compression - that's like saying, "not all gasoline is created equal... let's see how this motor oil performs."
but i guess if it brings in the ad dollars, it's mission accomplished for the site... even though ultimately it's doing the readers a disservice.
Wesley Fink - Tuesday, July 29, 2008 - link
You are technically correct, but it appears you are straining to make your point. Certainly you, and everyone else, understands that noise reduction is one of the processes that usually happens during in-camera "JPEG" processing or during software conversion from RAW to a finished image format which is normally JPEG, but could be TIFF or even another format.In the interest of being more precise in the description I have changed it to:
"Not all image processing is created equal, but Noiseware can often reduce noise that in-camera or software processing leaves behind."
n4bby - Tuesday, July 29, 2008 - link
thank you for fixing that."Not all image processing is created equal, but Noiseware can often reduce noise that in-camera or software processing leaves behind."
so does gaussian blur! ;)
soydios - Monday, July 28, 2008 - link
Eh, I'd say that your quoted caption is an accurate description of the article. There's on-chip noise reduction that is performed before the RAW file is created, then there's software noise reduction that is performed as part of the image processing engine in-camera. The output of that image processing engine is usually JPEG or TIFF, and the RAW preview image.So, if a camera's processing engine isn't running noise reduction before writing to JPEG, then this software would finish the noise reduction stage of post-processing.
As for the software itself, I don't know about the images at full magnification, but the small previews shown indicate fairly heavy smearing. Lightroom does less smearing, and seems to be extremely capable in my experience at chroma noise reduction, and good enough at luminance.
As for writing quality, I can't do better, so I won't criticize.
n4bby - Monday, July 28, 2008 - link
it is NOT an accurate description. image processing has NOTHING to do with the final output format of the file. you said it yourself - the output could be TIFF, so what does it have to do with "JPEG processing"? call it a nit if you will, but anandtech is a technical publication, so such fundamental inaccuracies should be considered anathema.and saying you can't criticize because you can't do any better yourself seems symptomatic of the sort of mediocrity malaise that results in sub-standard institutions like the US Postal Service. ;) if we expected everything in the world to be no better than what we ourselves can accomplish, the world would be a very sad place...
guitargeek27 - Monday, July 28, 2008 - link
It's kinda unfair to compare the noise of a Nikon D3 to a Canon XSI, as the Canon has a smaller sensor (as well as a few thousand dollars less)If you wanted to be fair, how about comparing the D3 to a Canon 5D (full frame and less expensive than the d3) or the EOS 1D Mark II or III?
Spoelie - Monday, July 28, 2008 - link
Personally I think there's too much smearing going on, to the point that I like the original picture better. The first picture for example has quite a loss of detail on the windows of the buildings and on the slope of the mountain. In the second picture all the ripples are gone and the boat seems to float above instead of in the water. All non-macro low-iso shots show an incredible amount of smearing (stones on the stairway, vegetation on every background mountain, ripples in water). I don't know how anyone would consider those filtered low-ISO shots an improvement.cparka23 - Monday, July 28, 2008 - link
I've long been impressed w/ Canon's low noise compared to other makes. Up 'til now, I've always heard/assumed that it was the sensor. So is it possible to determine to what exactly the noise reduction is attributable in many cameras? Is it due to an algorithm or is it really the sensor? I guess we'll never know, since the images are already 'processed' in RAW, will we?cparka23 - Monday, July 28, 2008 - link
hmm.. after rereading, that first question didn't come across as I had entirely hoped. regardless, the article helped me understand this a bit.Thanks, Wes.
strikeback03 - Monday, July 28, 2008 - link
Canon states that the only noise reduction done to their RAW images is dark noise subtraction. They have plenty of whitepapers on their technology floating around.Degloriath - Monday, July 28, 2008 - link
Thanks for the article, noise is certainly always a problem in digital photography. I've messed around with different noise-reduction methods via Photoshop itself, as well as a plugin called Noise Ninja which usually worked pretty well for myself. All in all it takes some patience and tinkering to ever get it perfect. :) Perhaps I shall look into the mentioned program as well.