Usefulness
According to a dictonary usefulness means the quality of being of practical use. This can be so much true in photography too. Some cameras seem perfect on paper, but produces horrible photographs. Some cameras look perfect, but the images can be awkward. Are less pixels better, or is everything determed by pixel density? Theory versus practicality. Or are there other aspects that might influence the image quality?
I have no technical background in photography. Everything is self taught, from the photo making process until the post processing, despite the fact that my grandfather was a professional photographer. I am not interested in facts and figures of a camera. I don’t care so much about sensor dimensions, or pixel density. I want to make photographs and a camera should not hinder this. So a camera needs to fit like a glove. The grip should be ergonomically shaped, adjusting the ISO value, aperture, or applying an EV compensation must be done as quickly as possible. And preferably, this should be done without diving to deep into the menu of the camera. The main reason why I so much like the Ricoh cameras.
Secondly I don’t need a perfect, flawless, and noise free, super clean image file. I like character. I don’t mind noise. As said before on my blog I like imperfection and limitations. You have to think and push harder. I like to achieve results with a camera that surprises me, and hopefully other viewers too. I prefer B&W photography and I have found a process method that creates results I envision when I make a photograph. I know how I can push a RAW file and work on my style. And that is the only way I judge a different camera.
Momentarily I use a Fujifilm Finepix E900 Zoom as a loan camera. The E900 is a camera from 2005 and was considered a cheaper alternative for the Canon PowerShot G6. The camera has a comparable lens zoom range, and RAW capability. I don’t like to pixel peep, but the jpegs show some distinctive smearing of the details. The noise reduction is quite aggressive in jpeg mode. To avoid this you can better shoot RAW.
The last view years there is an intense debate about the ever increasing amount of pixels on sensors that don’t (or hardly) increase in size. This resulting in higher pixel densities. According to many people on the internet in forums this increase in pixels provokes more noise at higher ISO’s. And many also assume it results in less dynamic range. So based on their experiences modern day cameras should have more noisy images than a few years ago. Now I don’t pixel peep, but occasionally I have to zoom in 100% to judge whether my post processing results in too much blodging. And I have noticed that the RAW files of the Fujifilm E900 do have more noise at ISO 100 than the Ricoh GX100 RAW files. Secondly I noticed that I can’t push the RAW files of the E900 as hard as the Ricoh ones. Ugly contrast borders can be detected and especially the blue channel contains more noise. I personally think the RAW files of the GX100 have a greater dynamic range than those of the E900. The GX100 files can be recovered pretty well. Maybe this is more the result of better electronics and camera/sensor software in the last view years.
Below you see an example of a Fuji E900 image at 100% and a Ricoh GX100 image at 100%. Both files are at ISO 100. It becomes obvious that the GX100 draws sharper and more detailed. These images originate from RAW files, processed at auto setting in RAW Therapee with no sharpening applied.
100% crop unprocessed examples of the Fujifilm E900 and Ricoh GX100
After that I processed my photographs in Adobe Lightroom with the B&W conversion I described before on my blog. It becomes obvious that both images get a bit noisier, but the GX100 file holds up well. I can push it harder and can add more contrast. The Fuji file on the other hand displays more flaws with serious noise and ugly edges around borders (like the trees). But remember, the sensor of the Fuji camera is slightly larger and has 1 million less pixels. So it seems, that despite the lower pixel density of the Fuji, the signal to noise ratio is worser. That could partly be the result of the lens too. I think the Ricoh lens has the edge on sharpness and contrast.
100% crop processed examples of the Fujifilm E900 and Ricoh GX100
Although we don not necesarely need the amount of pixels, increasing the amount of pixels does not necesarely supply worser images. Really good glass in front of the camera could make a bigger difference in the final image quality than the size and pixels count of the sensor. A combination of carefully selecting a good sensor and designing a fast lens, and high quality lens increases the usefulness of the camera and its captured images. This could be a major point for Panasonic with their new Lumix DMC-LX3 incorperating a fast F2.0-F2.8 24-60mm lens. This will be a serious camera to consider when you want to own a serious compact camera.
In the meantime I suggest you just keep making photographs (or at least try too), and lets stop whining about pixelcounts (unless you spent large amount of times on the dpreview talks discussing cameras instand of photographing).
Filed under: 2008, Photography | 11 Comments
Tags: image quality, lens, pixel density, pixels, processing, raw, sensor, signal noise ratio, usefulness
-
-
-
Recommended at Amazon
-
RSS - Posts
I wonder how much of the difference may be down to the noise reduction algorithms used by each manufacturer?
Both examples originate from RAW files. Although they should be “unprocessed” it is a fact that still many manufacturers apply some sort of processing to the RAW files as well. Many believe the RAW file of the Panasonic LX2 gets some sort of noise reduction. I seriously don’t know if Fuji or Ricoh apply noise reduction to their RAW files in-camera.
Cheers,
Wouter
This is quite a difference between the two cameras. Would have thought the Fuji will hold up better with the Super CCD.
Funny to have been litening nothing but praise about Fuji´s Super CCD and it´s suposed edge in the signal to noise ratio & clean files to come and stumble upon a “real life” experience. You need to be carefull with internet and the reviews and opinions you find there. An opinion is like an ass: everybody has one… Nice thoughts and a good empiric demonstration that the most important side of photography is actually USING the camera and not speculating about it..
Damn, the LX3 has become a dangerous thing, hasn´t it..?
Erik.
I was pretty suprised too. RAW can show the flaws in camera design.
Lies, damned lies, and statistics!
Just wondering from the Fuji colour photo if the edge problem was actually Chromatic Aberration (CA) rather than a processing artifact?
Already we have seen with the few sample photos for the LX3 a huge variation in quality which can really only be down to the operator as the many reviewers at the launch event were all using the same camera model & firmware. Never under-estimate the human component!
I thought Chromatic Aberration is mostly a lens problem. I did not mention that the edge problem was a processing artifact. It is a problem that comes up at processing and show limitations of the camera. But I also noticed that it was noisier, despite the larger sensor and lower pixel density. I just want to say that people often judge a camera on their specifications. If a camera gets more megapixels, it must be worse than the one before.
Now be careful about the LX3 launch photographs. As far as I understood these LX3 photographs were all taken with preproduction models. The model might be the same, but because it was not a production camera there may always be differences.
Cheers,
Wouter
Chromatic aberration is not only caused by lens design, but also through microlenses on the CCD. These lenses are used to correctly focus green light, the incorrect focusing of red and blue results in purple fringing around highlights. And it is more a problem in CCD’s with a very small pixel pitch. This is what’s been called “purple fringing”.
On photographs taken using a digital camera, very small highlights may frequently appear to have chromatic aberration where in fact the effect is because the highlight image is too small to stimulate all three color pixels, and so is recorded with an incorrect color. This may not occur with all types of digital camera sensor. Again, interpolation techniques may affect the apparent degree of the problem.
These ugly borders can only been seen at high contrast areas on images from the Fujifilm E900. To avoid it, you could take a color channel in Photoshop and process that one.
Wouter
Re Fuji Supper HR CCD’s;
They really started getting the legendarty status for the low noise with the 6mp sensors in the F10/20/30/31fd’s.
The 9mp ones in the 900 are the same as that in the S9100/9600 which while very nice, are not nearly so well thought of for hi ISO.
As Wouter has shown here, the only really fair way to judge sesnors is by working with the RAW files sans any OEM processing.
In the case of the 6MP Super HR CCD’s barring CHDK style hacks that can only ever be done using the Fuji S6000fd; it uses this sensor in a Bridgecam body with a 28-300mm efl zoom and allows RAW.
Wouter, if you like I could send you some RAW files from mine to work with?
Thank you Lili. The E900 is the first Fujifilm camera I have ever used. I read about the legendary status of the Super CCD all over the place, but I need to see it in real life situations. I love to get a RAW file from your camera. I like to know how it copes with Lightroom and my B&W processing.
I will post a processing method on my blog to overcome the problem of chromatic aberration with Photoshop. It is more tedious and time consuming. But it works.
Cheers,
Wouter