Getting, from the highly compressed format the GO-HD records, back to any kind of an editable format is extremely processor intensive. For example, I used the latest version of Final Cut Studio Compressor on an Intel Macbook Pro to convert the raw video from the GO-HD to an editable MOV file, and a 30 second clip took 30 minutes to render. On an HP laptop, using QuickTime Pro, it took a whopping 1-hour to render only 6 seconds of video to an editable HD MOV. This is NOT an editing-friendly codec. In addition, although the latest version of Vegas and Final Cut supposedly support native editing of AVCHD, it is still going to take some significant processing power to get to a practical system for editing. [Editor’s Note: Vegas’ support of AVCHD is limited exclusively to Sony cameras, so that’s one less solution that could even make a pass at it.]
I suddenly saw the financial advantages slipping away. The GO-HD is inexpensive, but I was going to have to compensate for the compression with a much more powerful computer. And I had to ask myself whether the extra money I would have to invest in processing power might be better invested in a video camera with a more editing-friendly codec.
Another problem with the codec that the GO-HD uses is the illusion of added detail and information as it relates to the entire editing and exporting process. That detail really gets blown away when you try to get to an editable format. This brought up real questions, after looking frame-by-frame at my editable MOV files: Whether the quality could hold up all the way through the requisite workflow? After all, the image had been highly compressed, then it had been expanded and decompressed for editing, and finally. it would be compressed again for output to DVD or for the web. This can quickly become an issue because there is simply a limit to how much crunching and stretching the images can take before they start to fall apart visually.
Automatic Exposure Cleverness
During one test, I was shooting traffic with the GO-HD. The cars were moving at a constant speed. But on playback, the cars would start out moving smoothly and then suddenly become jerky, almost like a bleach-bypass effect. It took me some experimentation to finally understand what was happening.
See, when the clouds moved apart and direct sunlight shone on the cars, the exposure changed. Rather than allowing the video to get blown out, the camera adapted. There is only so much that it can do to shut down the iris, and because there are no neutral density filters (and no simple way to attach them, that I could see), the camera would increase the shutter speed to compensate for the added light. As a result, when I was shooting in direct sunlight, the footage looked like I was shooting with perhaps a 1/500th or 1/1000h of a second shutter. No motion blur. Just sharp, sharp images from one frame to the next. When the light in the environment changed, the shutter speed would slow, and the footage would look normal.
I understand why the camera did this – to preserve a usable exposure. And it is certainly the behavior I would want from a casual-use camera – where I would not want to miss the shot just because the light was changing. However, in a filmmaking camera, having the motion qualities of the video change during a single clip is not usually desirable.
