Camera Review: Sony NEX-VG10
- APS-C CMOS Sensor
- 1080-60i (30PF) AVCHD format
- interchangeable lens
- 2lb 12oz w/lens
The only way I can understand Sony's NEX-VG10 camcorder is to think of it as an experiment. In general, I'm all for experimentation, but there is a difference between previewing a prototype and asking people to pay money for a finished product, and in my opinion, the NEX-VG10 tilts resoundingly toward prototype.
What's so bleeding-edge about the NEX-VG10? It puts a DLSR-sized sensor in a camcorder body, maintaining interchangeable lens functionality. All things being equal, a larger sensor size translates into better image quality and reduced depth of field. The VG-10's 23.4mm x 15.6mm sensor is huge: some twenty times bigger in terms of physical area than a typical camcorder's 1/3" chip.
The NEX-VG10's big sensor, paired with its 18-200 (35mm equivalent) lens, does indeed let you create images with magnificently shallow depth of field. Professional (ie Hollywood) cinematographers covet this look, because it allows them to isolate the subject within the frame, calling attention to the scene's most important element. Creatively, this is simply a must-have feature for serious filmmakers.
And here is where things start to go off the rails, because in nearly every other way you can think of, the NEX-VG10 is simply wholly inappropriate for serious filmmaking. The NEX-VG10 is trapped in a netherworld between camcorder and still camera: intended to be both, it is neither. The lack of expected camcorder functionality is simply astonishing. And what, I ask you, are you getting in the still-photo department that an NEX-5 can't already do—and do better?
In terms of video quality, the NEX-VG10 essentially matches the NEX-5A (but not the newer 5N!), shooting 30 1920 by 1080 progressive frames per second wrapped as 60i AVCHD. Yes, the VG10 shoots a 24Mbps data stream (versus the NEX-5's 17Mbps), but realistically you're not going to see much difference. In both cases, the picture is noticeably marred by aliasing artifacts, which remain the bane of DSLR-sensor videography.
Why does this happen? Because these large format sensors are not native 1920x1080—the VG10's sensor, for example, is a 14 megapixel chip, meaning the camera has to resize (downsample) the raw image to fit video's 1920x1080 frame size. Resizing is a processor-intensive business. Given current technology and pricing, doing it 30 times a second requires quite a lot of compromising when it comes to algorithmic purity. End result: banding and other artifacts.
If you know how to work around the camera's limitations, it is possible to get good and even gorgeous footage, particularly shots that isolate simple subjects against blurred backgrounds. But in general, video quality proves disappointing, especially in documentary/run-and-gun situations where you're not setting things up in advance.
Something else to worry about: a predicable concern given the NEX-VG10's interchangeable lens design is sensor dust. I tried a used camera for my tests, and noticed footage-wrecking dust motes on the images. Note that with video, you can't simply clone out the spots in photoshop. There is no real fix. That leaves VG10 owners needing to constantly worry about dust infiltration, and subsequent sensor cleaning hassles.
What is the future of the VG10? Sony has already announced a successor, due late 2011, called the VG20H. That camera will shoot 1920x1080-60p video using the new AVCHD 2.0 spec. But will it resolve any of the VG10's vexing liabilities? At this point, I wouldn't expect any miracles. Until we see either native tri-color DX-sized HD chips, or much improved processing, the outlook for DSLR sensor-based video will remain fuzzy.