Review: Dust removal systems / sensor cleaning

Róbert Irházy (Birdie) - March 20., 2007. 8:00am (CET)
Sensor cleaning is becoming a standard feature on modern digital SLR cameras. While the manufacturers may use different techniques to get rid of the dust from the image sensor, they all promise clean, dust free images. Do they all work equally well? Can we throw our air blower away? We have tested the four different types of sensor cleaning technologies to see which one will most likely to fulfill the expectations.

Kattintson ide a magyar változatért

When a camera we are reviewing arrives to the office, we are very gentle and careful, as not to cause any harm to the camera. They are expensive equipment, and we can't afford to pay the repair bills. Two weeks ago we got three cameras (the fourth one in the test was our own camera) for a test that most of us wouldn't want to do at home.

Our goal was to set up a controlled environment to get the image sensors of digital cameras dirty and test how effective the different cameras' sensor cleaning technologies are. It is well known fact (and it is a bit over dramatized) that digital SLR cameras are very sensitive to the dust particles found in the air. These particles can settle on the image sensor, and will show up on the images as dark spots when the image is taken with a small aperture (high F number). When the aperture is open, the dust is usually not a problem, since the distance between the sensor's surface and the dust is large enough to get the dust out of the focal plane. When the aperture is smaller (usually smaller than F16) the rays of light are parallel to each other, creating shadows to the dust particles. These shadows will show up as clearly on the images. At F22 or smaller apertures, the spots become really visible. How does one prevent this from happening when we can't afford to keep the same lens on all the time? The easy solutions come from the manufacturers, but they first must go through our tests to receive our approval.

How sensor cleaning works

Supersonic Wave Filter

The first manufacturer to offer sensor cleaning technology in their digital SLR cameras was Olympus. The Supersonic Wave Filter became a standard feature of the DSLR cameras using the FourThirds system. The idea is to shake the thin foil in front of the sensor with ultrasonic (above 20 kHz) frequency, and collect the falling dust particles on a adhesive surface. This process happens every time the camera is turned on, or when this function is selected from the menu. This adhesive material should be replaced after a few years to keep the effectiveness of the system.

Olympus was the only manufacturer to support dust reduction for a long time. Since there was no competition, there was no way (and reason) to conduct a comparison test. Although we have never really tested the sensor cleaning / dust removal system of Olympus cameras in particular, we never had any dust related problem with their DSLR cameras. The same cannot be said to the cameras of other manufacturers, although we usually received well-cleaned cameras to review.

Olympus lost its monopoly on dust reduction systems when Sony announced the Sony Alpha DSLR-A100 digital SLR camera in June 2006. This camera used the CCD-shifting technology (originally developed by Konica Minolta) to provide sensor cleaning functionality. In this case, the camera shakes the sensor itself (not a thin foil) with a rather high amplitude and low frequency (we didn't measure, but it is in the 100 Hz range, far from ultrasonic). To help this process, Sony put an indium-tin oxide coating on the sensor to lower the static charge.

Canon EOS Integrated Cleaning System
Canon EOS Integrated Cleaning System

Sony was closely followed by Canon with their introduction of the EOS-400D, which also promised a solution to the dust problem. The press release described a technology very similar to the Olympus SWF system. The filter in front of the sensor is shaken by a piezo crystal, so the dust particles will fall off. The second part of Canon's solution is a dust map that can be recorded in the EXIF header of the images, and this map can be used (even with RAW images) to get rid of the dust spots appearing on the images. This system is the called the Canon EOS Integrated Cleaning System. YouTube has a video to illustrate this function. Nice job by the marketing department, but the efficiency of the system is still to be proven.

The last manufacturer to introduce such a system was Pentax. The technology used in the Pentax K10D is somewhat similar to the one found in the Sony DSLR-A100. The image sensor is equipped with a CCD-shifting functionality, what is also used for sensor cleaning when needed.

Sigma cameras have a protective filter in front of the mirror, preventing the dust particles to get into the camera at the first place. The sample photos we have seen so far didn't really convince us about the effectiveness of this concept.

Due to the ongoing cooperation between Olympus and Panasonic on FourThirds system cameras, the Lumix DMC-L1 uses the same Supersonic Wave Filter technology found on all Olympus digital SLR cameras.

Similarly, the only Samsung camera providing this feature (GX-10) is equipped with the same system that we can find in the Pentax K10D.

Currently Nikon, Fujifilm and Kodak (although this latter one doesn't even have current DSLR camera model on sale) cameras are missing this useful function.

Sensor cleaning test: How we did it

After that short overview, let's start with the introduction of the participants. I would also like to explain what we did during the test.

Canon EOS-400D
Olympus E-300
Pentax K10D
Sony Alpha DSLR-A100

We had four different cameras from four different manufacturers: Canon EOS-400D (Canon EOS 400D review in English), Olympus E-300 (Olympus E-100 review in English), Pentax K10D (Pentax K10D review in English) and Sony Alpha A100 (Sony Alpha DSLR-A100 review in English).

We started the test with an exhaustive manual sensor cleaning, since not all cameras arrived with entirely clean sensors. After that, we shot a reference image with each camera to see the photo from a clean (or almost clean) sensor. It seems that the sensor of the Canon EOS-400D is statically charged, since our air blower was not able to get rid of all the dust. The reference images are available in this article (below).

The second step was the dusting (not cleaning, but dirtying) process. We locked up the mirrors, and left the cameras laying on their backs for a few minutes. We artificially kept the air dusty, what caused the sensors to become equally dirty.

Since the Olympus E-300 and the Sony Alpha DSLR-A100 don't allow us to turn the camera on without the sensor cleaning process, we powered up the cameras still laying on their backs. This gave us the chance to take the first shots with all cameras without the sensor being cleaned even once. The image taken at this stage is labeled as dusting.

This is when the exciting part started. We set the cameras to normal position (standing on their bottoms in landscape orientation), and performed two sensor-cleaning processes (turned the cameras on and off twice). The images taken after the second cleaning cycle are labeled as 2nd cleaning.

One or two cleaning cycles wouldn't have made this test very informative, so we have repeated the cleaning process 23 more times with each camera. We believe that if the cleaning process is any good, the result should be visible after running it 25 times in a row. The resulting images are carrying the 25th cleaning label.

Since some of the systems failed, we used our good old air blower to clean the image sensors (mirrors locked up, camera facing down, air blown onto the sensor's surface). We did this twice. The images taken after this step are named as 2nd blow.

When we only have a few spots on the sensor, blowing air into the camera twice usually does the trick, but at this time we got a little more dust particles on the sensors than that. The images labeled as final were taken after our extensive chemical (99% clean ethyl alcohol) cleaning.

Sensor cleaning effectiveness

By evaluating the results shown above we can say that only Olympus has a somewhat effective solution. Lets discuss it in more detail:

Canon EOS-400D

Personally, I was expecting a lot from Canon, since their system is very similar to the one used in Olympus cameras. We assume that since the CMOS was statically charged, this prevented the dust from falling of the sensor's surface. We can see that the first two cleaning cycles have cleaned of about a dozen of spots, but that is almost nothing compare to the total number of over 500 spots.

After the second cycle, no more particles fell of the sensor.

Although the system is promising, it is not as effective as we would expect when the sensor is statically charged. We don't know how to prevent the charging, but the manufacturer should address this problem in future releases. ?We estimate that the effectiveness of this system is 5%.

Our air blower has managed to remove most of the spots, but still this camera had the most dust after the air cleaning. This seems to support our static charge theory.

Olympus E-300

The system that was first introduced seemed to work quite well. We must note, that the spots were also less visible on this camera, and the sensor cleaning technology has worked much more effectively than in any one of the other three cameras. The dust particles are less visible because the sensor is smaller, and the protecting foil is probably further away from the sensor than in the other tested cameras.

Although the result after the second cleaning cycle wasn't very impressive, we still counted over 30 spots gone. We were hoping to get more off with the first two cycles.

25 cleaning cycles have removed more than half of the particles from the sensor. If we take into consideration that in real life we almost never get more than 3-4 spots on the sensor at once, we can imagine this system working very well (ever since we received the Olympus E-300, we haven't had any dust problems). This camera has produced the best results. We would rate the effectiveness to 50%.

The air blower managed to clean the sensor well, although we still had 20-30 spots left. Can we blame this on static charge as well? We don't know.

Pentax K10D

Part of the reason we started to organize this test was that we had a bad experience with the dust removal system of Pentax K10D that we had for our review.?Although the amplitude of the CCD movement is quite high (one can feel the shake on his hand), the frequency is rather low.

After the second cleaning cycle surprisingly we have found more spots on the image than we had before it. This could be caused by the dust stuck in the mirror area. One thing is sure though: this system is not working as promised. Even after the 25th cleaning cycle we had all the spots on place. I tried really hard to find one disappearing spot, but I failed. So did the Pentax dust removal system. Effectiveness: 0%.

After using the air blower the remaining spots decreased to a dozen. This shows, that the sensor's surface is not statically charged. Knowing this increases our disappointment even more.

Sony Alpha DSLR-A100

After seeing the results the Pentax K10D produced last year, I wasn't expecting too much from Sony's dust removal system. Unfortunately my suspicions were correct.

The first two cleaning cycles has increased the number of spots on the sensor, just like we observed in the case of Pentax K10D. After the 25th cleaning we had exactly the same number of spots as we observed after the second cleaning. Effectiveness: 0%

The air blower did a good job again, only about two dozens of particles remained on the sensor. The CCD is probably not charged, so the indium-tin layer is not just marketing-speak.

Dust removal conclusion

Here is our ranking according to effectiveness:

  1. Olympus: good
  2. Canon: poor (we are disappointed)
  3. Pentax and Sony: useless (we are very disappointed)
If you are looking for a camera, have the dust removal as an expectation only at the end of your list. If sensor cleaning / dust removal is a must, the choices are limited to Olympus and Panasonic cameras.
We guess that theoretically it is possible to get better results with Canon cameras when the CMOS sensor is not charged, but Canon must find a way to discharge the sensor first (if this is really the problem). If the system is so unreliable, the EOS-ICS is not much more than a good marketing name.
In case of Sony, Pentax and Samsung cameras, the dust removal function should be considered as almost non-existent.

The final conclusion: keep your air blower handy, it is still your best tool against dust.

First look: Panasonic DMC-FZ18 preview

Manufacturers are predictable at most of the times. The replacement of a camera usually comes 10-12 months after it's announcement, while it is very rare that a camera is replaced after only 5 months. While the new Panasonic Lumix DMC-FZ18 is officially not a replacement model of the DMC-FZ8, this latter model has to get an eye catching price tag to survive the in house competition.

Nikon D40X review

If we look at the naming convention of Nikon, the X means enhanced, improved version of a previous model. We have seen this before even in the digital era. Now Nikon's successful entrly level camera, the Nikon D40 received it's brother, the Nikon D40X. Despite the enhanced version, the D40 is staying on the market, so the D40X will have a tough job to win the customers. Does Nikon D40X have a chance? We will try to look at every detail to find out.

Nikon D40X image samples

We have uploaded the first 15 images taken with the brand new Nikon D40X and the Nikkor AF-S 55-200 mm F4-5.6 ED DX VR lens. Our detailed review is in editing right now, will be online soon.


Panasonic Lumix DMC-FZ8 image samples

See how the Panasonic Lumix DMC-FZ8 has performed on our tests. Our first test images are online. Detailed review is to follow soon.


Review: Fade resistance test

Do you want your grandchildren to sit around you and see your nice, old, color photographs in fifty years from now? If so, you better watch how you get those digital images printed. While some printer manufacturers promise that the photos will last for over 100, in some case 200 years, don't forget to read the footnotes. Those numbers are usually guarantied only when the prints are stored in albums locked away from sun, heat and air. Since we don't have 200 years to test which prints will last for as long as the manufacturer promises we did a not so scientific, but very demonstrative test to find out which printing technology is the winner these days.