Video Accessories Misc
- Written by John E. Johnson, Jr.
- Published on 30 June 2008
On the Bench
First, the lens falloff test results. Below is the analysis at wide angle. It's very good, with only 0.1 f/stop falloff.
For telephoto, falloff maxed out at 0.5 f/stop.
Resolution, using the MTF50 criterion, was 296 LW/PH (with the correct amount of sharpening, it would be 335). This is lower than we measured for the Canon HV20, which records a different type of compression algorithm onto tape rather than hard disk. The HV20 had an MTF50 resolution of 627 LW/PH. The MTF50 criterion is a definition of 50% contrast between juxtaposed white and black lines, so the number can be affected by increasing the contrast and edge enhancement in the video, but this does not necessarily improve the subjective quality of the image as viewed by the consumer. We will discuss resolution vs. sharpness in the Primer section shortly.
Resolution is not as good with video cameras as it is with still cameras for several reasons. One is that even the best video cameras have a 1,920 x 1,080 sensor, which is 2.07 megapixels, while all modern digital still cameras are least 5 megapixels in size, and many are 10 megapixels. Secondly, the compression for video is usually very high so as to make it easier to record 60 interlaced fields or 30 frames per second. Compression for a still camera photograph does not have to be very high, and in fact, can be recorded in RAW format, maximizing the quality. The bitrate for AVCHD, such as in the HG10, is lower than the bitrate for MPEG-2, such as used with the HV20, so the compression has to be higher when lower bitrates are used in order to deliver 60 fields per second within the confines of the bitrate. Third, video cameras these days are 1080i, which means they are recording 60 interlaced fields per second. To put together a complete picture, you need to join two fields, and because they are taken at different times, it is just about impossible to make them align perfectly, no matter how steady the camera is. Even if it is on a tripod, there are vibrations in the room that will disturb the exact alignment between fields.
Chromatic (CA) aberration was 1.15 pixel. This is about twice the CA we measured in the HV20. CA will be worse in the corners, but we measure it in the center because lens manufacturers tend to offset the aperture to minimize curvature of field, and as such, the center of the image has been produced from light that passes to one side of the lens center, and there will be some CA even at small apertures.
The ChromaDuMonde test chart, which has a very tough set of colors for cameras to reproduce, looked slightly too blue.
The gray scale test. First, the gray scale itself, as photographed with the HG10. It recorded just a bit too warm (red) to my eye.
Here is what the analysis grid looks like.
And, the results. Noise reaches a maximum of about 0.5%, in the blue channel. This is actually somewhat better than the Canon HV20, which had noise up around the 0.9% level. The stepchart shows that the gray levels follow the standard closely. At the white end, the stepped black line is below the dotted line. This means the camera reduces the level here, keeping the whites from being blown out, and it is a common technique with digital cameras.
Results using the ColorChecker SG chart indicate that color reproduction is slightly undersaturated. This is shown by the fact that the standard swatch, shown in the top left corner of each colored square, is somewhat darker than the reproduction of that swatch color from the camera, shown in the bottom right corner of each square. This color accuracy is not as good as the Canon HV20.
The Canon HG10 is very easy to use and has the advantage of hard disk storage of video files which makes it simple to review them while they are in the camera to make sure you got the video you wanted. These files will take some getting used to in terms of dealing with them during the download process to your PC because there are several files for each video. The noise issue has improved a bit over their HV20, but the resolution is not as good. In any case, the picture is still very high quality. As high definition video cameras go, I like this one.