Analysis of Fader ND (mark I) filter image quality
In February 2010, I purchased an 82mm Fader ND Filter (ND2 to ND400) from the official eBay store for Light Craft Workshop. The label on the filter reads "82mm Fader ND [W] HL-OPTICS".
The purpose of a Neutral Density (ND) filter is to reduce the amount of light, therefore allowing for slower shutter speeds. A variable ND filter allows you to vary the amount of light by rotating the front of the filter.
When shooting video, the shutter speed is normally fixed - it is tied to the frame rate giving the required Shutter Angle (e.g. 180 degrees) controlling motion blur. When shooting video, the only way to prevent over exposure in bright light is to Stop Down the aperture. This limits your ability to use the aperture to control depth of field.
A variable ND filter is particularly useful for video, since it allows you to control exposure independently of the aperture and ISO. Specifically you set your fixed shutter speed, choose your depth of field, then set exposure using a variable ND filter or ISO gain (for bright or dim light respectively).
Variable ND filters can be constructed by stacking a Circular Polariser (CPL) filter and a Linear Polariser (LP) filter. However I heard some people had issues with colour shifts, and I preferred to just buy a variable ND filter ready to use.
I believed that I would be able to use the variable ND filter for normal photography and not just video. I've found though that image sharpness is severely affected, and out-of-focus (Bokeh) areas of the image have a crazing pattern.
Note, I've only tested the Light Craft Workshop Fader ND filter, and not the Vari-ND filter sold by Singh-Ray. Additionally Light Craft Workshop now no longer sell this model of variable ND filter, the new one is named Fader ND mark II.
These test images were taken under these conditions:
Test 1
Canon 7D with Sigma 24-70 f/2.8 HSM. ISO 100, f/8, 45mm. Manual focus on the front corner of the battery.
This was just supposed to be a general photo which should be reasonably sharp for most of the objects. The image sharpness is degraded across the whole image, and the bokeh on the lens appears like a ghosted double image.
Test 2
Canon 7D with Canon 70-200 f/4 L USM. ISO 100, f/4, 70mm. Manual focus on the front corner of the battery.
For this image I've use a shallow depth of field (f/4). Image sharpness is again degraded and shows ghosting.
Test 3
Canon 7D with Canon 70-200 f/4 L USM. ISO 100, f/11, 70mm. Manual focus on the front corner of the battery.
In this image, I've increased the aperture to f/11 to see if things improve when bokeh is taken out of the equation. Sharpness is better but still far from the original image. Chromatic aberration is more apparent on the lens hood.
Test 4
Canon 7D with Sigma 24-70 f/2.8 HSM. ISO 400, f/2.8, 70mm. Manual focus on the pencil label. All images have the same white balance applied (photos were taken under consistent downlight lighting).
In this image, I've tried to create lots of bokeh with an aperture of f/2.8. Now the Sigma 24-70 doesn't have the most pleasant bokeh because the circles have a hard circle edge. In this case the bokeh has a strange noise pattern which clearly demonstrates the problem of the ND filter.
Conclusion
This model Fader ND filter is unsuitable for true photographic work. It is questionable whether it is entirely suitable for video work. Specifically the times you'll use a ND filter for video are when you want to use a large aperture to create large/pleasant bokeh. This filter will result in harsh ugly noise patterns in the bokeh even at the reduced resolution of video.
The reduced sharpness makes it more difficult to manually focus - even when using live-view.
I can only assume the Mark II version of the Fader ND filter, and the Singh-Ray Vari-ND filter both perform better. However at this point I'm unable to test them.
The purpose of a Neutral Density (ND) filter is to reduce the amount of light, therefore allowing for slower shutter speeds. A variable ND filter allows you to vary the amount of light by rotating the front of the filter.
When shooting video, the shutter speed is normally fixed - it is tied to the frame rate giving the required Shutter Angle (e.g. 180 degrees) controlling motion blur. When shooting video, the only way to prevent over exposure in bright light is to Stop Down the aperture. This limits your ability to use the aperture to control depth of field.
A variable ND filter is particularly useful for video, since it allows you to control exposure independently of the aperture and ISO. Specifically you set your fixed shutter speed, choose your depth of field, then set exposure using a variable ND filter or ISO gain (for bright or dim light respectively).
Variable ND filters can be constructed by stacking a Circular Polariser (CPL) filter and a Linear Polariser (LP) filter. However I heard some people had issues with colour shifts, and I preferred to just buy a variable ND filter ready to use.
I believed that I would be able to use the variable ND filter for normal photography and not just video. I've found though that image sharpness is severely affected, and out-of-focus (Bokeh) areas of the image have a crazing pattern.
Note, I've only tested the Light Craft Workshop Fader ND filter, and not the Vari-ND filter sold by Singh-Ray. Additionally Light Craft Workshop now no longer sell this model of variable ND filter, the new one is named Fader ND mark II.
These test images were taken under these conditions:
- Camera mounted on a good tripod.
- Shutter release using a cable.
- Photos were taken using live-view, silent shooting mode 2.
- The lighting varies for the first couple since the light source is light from the windows.
- Three images in each set are using the ND filter (low, medium, high) and the remaining image is taken without a filter.
- Photos are taken in raw, and converted to jpg typically using the default settings. That is, they are un-processed.
- Crops are displayed at 100%.
Test 1
Canon 7D with Sigma 24-70 f/2.8 HSM. ISO 100, f/8, 45mm. Manual focus on the front corner of the battery.
This was just supposed to be a general photo which should be reasonably sharp for most of the objects. The image sharpness is degraded across the whole image, and the bokeh on the lens appears like a ghosted double image.
 | Shutter: 2 sec No filter Original Image |
 | Shutter: 4 sec Fader ND (low) Original Image |
 | Shutter: 8 sec Fader ND (mid) Original Image |
 | Shutter: 30 sec Fader ND (high) Original Image |
 | Shutter: 2 sec No filter Original Image |
 | Shutter: 4 sec Fader ND (low) Original Image |
 | Shutter: 8 sec Fader ND (mid) Original Image |
 | Shutter: 30 sec Fader ND (high) Original Image |
 | Shutter: 2 sec No filter Original Image |
 | Shutter: 4 sec Fader ND (low) Original Image |
 | Shutter: 8 sec Fader ND (mid) Original Image |
 | Shutter: 30 sec Fader ND (high) Original Image |
Test 2
Canon 7D with Canon 70-200 f/4 L USM. ISO 100, f/4, 70mm. Manual focus on the front corner of the battery.
For this image I've use a shallow depth of field (f/4). Image sharpness is again degraded and shows ghosting.
 | Shutter: 0.7 sec No filter Original Image |
 | Shutter: 2 sec Fader ND (low) Original Image |
 | Shutter: 3 sec Fader ND (mid) Original Image |
 | Shutter: 10 sec Fader ND (high) Original Image |
 | Shutter: 0.7 sec No filter Original Image |
 | Shutter: 2 sec Fader ND (low) Original Image |
 | Shutter: 3 sec Fader ND (mid) Original Image |
 | Shutter: 10 sec Fader ND (high) Original Image |
 | Shutter: 0.7 sec No filter Original Image |
 | Shutter: 2 sec Fader ND (low) Original Image |
 | Shutter: 3 sec Fader ND (mid) Original Image |
 | Shutter: 10 sec Fader ND (high) Original Image |
Test 3
Canon 7D with Canon 70-200 f/4 L USM. ISO 100, f/11, 70mm. Manual focus on the front corner of the battery.
In this image, I've increased the aperture to f/11 to see if things improve when bokeh is taken out of the equation. Sharpness is better but still far from the original image. Chromatic aberration is more apparent on the lens hood.
 | Shutter: 4 sec No filter Original Image |
 | Shutter: 10 sec Fader ND (low) Original Image |
 | Shutter: 15 sec Fader ND (mid) Original Image |
 | Shutter: 30 sec (pushed 1 stop) Fader ND (high) Original Image |
 | Shutter: 4 sec No filter Original Image |
 | Shutter: 10 sec Fader ND (low) Original Image |
 | Shutter: 15 sec Fader ND (mid) Original Image |
 | Shutter: 30 sec (pushed 1 stop) Fader ND (high) Original Image |
 | Shutter: 4 sec No filter Original Image |
 | Shutter: 10 sec Fader ND (low) Original Image |
 | Shutter: 15 sec Fader ND (mid) Original Image |
 | Shutter: 30 sec (pushed 1 stop) Fader ND (high) Original Image |
Test 4
Canon 7D with Sigma 24-70 f/2.8 HSM. ISO 400, f/2.8, 70mm. Manual focus on the pencil label. All images have the same white balance applied (photos were taken under consistent downlight lighting).
In this image, I've tried to create lots of bokeh with an aperture of f/2.8. Now the Sigma 24-70 doesn't have the most pleasant bokeh because the circles have a hard circle edge. In this case the bokeh has a strange noise pattern which clearly demonstrates the problem of the ND filter.
 | Shutter: 1/15 sec No filter Original Image |
 | Shutter: 1/4 sec Fader ND (low) Original Image |
 | Shutter: 1/3 sec Fader ND (mid) Original Image |
 | Shutter: 1.5 sec Fader ND (high) Original Image |
 | Shutter: 1/15 sec No filter Original Image |
 | Shutter: 1/4 sec Fader ND (low) Original Image |
 | Shutter: 1/3 sec Fader ND (mid) Original Image |
 | Shutter: 1.5 sec Fader ND (high) Original Image |
Conclusion
This model Fader ND filter is unsuitable for true photographic work. It is questionable whether it is entirely suitable for video work. Specifically the times you'll use a ND filter for video are when you want to use a large aperture to create large/pleasant bokeh. This filter will result in harsh ugly noise patterns in the bokeh even at the reduced resolution of video.
The reduced sharpness makes it more difficult to manually focus - even when using live-view.
I can only assume the Mark II version of the Fader ND filter, and the Singh-Ray Vari-ND filter both perform better. However at this point I'm unable to test them.
Comments