Driver Version Tested: 3.0
Type: Hybrid hardware/software encoder
MPEG Standards: MPEG 1, MPEG 2, VCD, XVCD (2.3 Mbit/s CBR),
SVCD, DVD
Bit Rate Control Modes: average/max-based VBR, CBR
Bit Rates: .001-15 Mbit/s
GOP Formats: I-frame, IP-frame, IPB-frame
MPEG Audio Modes: Full MPEG-1 layer 2 support, linear PCM and AC3
Capture to AVI? No
DirectShow Capture Driver? No
Frame Sizes: In MPEG-2 mode: Full D1, cropped D1,
2/3 D1, half D1, and SIF; In MPEG-1 mode, only SIF
is available
Scene Change Detection? Yes
Macrovision Detection? No
Inputs: Shared S-Video/Composite, stereo audio (minijack)
Outputs: Shared S-Video/Composite, stereo audio (minijack)
Video Pass-through? Yes, but does not show decoded video
during capture
Audio Capture On-Board? Yes
Drivers: Windows 98, Windows ME, Windows NT 4.0, Windows 2000,
Windows XP
Web Page: http://www.vitecmm.com/
Availability: Video specialty stores
Price: US$500
The DCM is a "hybrid" encoder: the capture card preprocesses the video, but does not completely finish the MPEG encoding process. It's then passed down to the CPU to finish encoding. This results in a cheaper capture card and a more flexible encoder: you only have to change the software and possibly also the EPROM chip on the board to change the encoding algorithm. But, this requires that you have a powerful CPU to capture video. For MPEG-2 encoding, Vitec recommends a Pentium III-800 as the minimum configuration. For MPEG-1 encoding (which is limited to SIF frame sizes by the DCM's Imager capture application), Vitec recommends at least a Pentium II-400.
In our testing, we got the CPU nearly peaked on a 933 MHz system when doing full-D1 encodes, and it ranged from 60-75% on an Athlon XP 1700. On that Athlon, the preview window stuttered a bit while encoding, but this didn't affect the output file: it plays back smoothly.
Unlike Vitec's MPEG Maker software MPEG encoder, the DCM doesn't use a CQ-based VBR mode. Instead, it uses a simpler but lower-quality average/max type VBR mode. (See my page on MPEG encoder modes for more information on why CQ-based encoding modes are superior.) At the test rate of 3 Mbit/s, there is very little variability in the output bit rate. Even at 5 Mbit/s, there's less than a megabit per second variance from the average with our test video. This behavior is typical of non-CQ VBR modes.
This encoder does do one very cool thing: it can apparently parse Dolby Surround stereo sound (Pro Logic and older systems, most likely) and turn them into AC-3 surround streams for DVD. No other encoder in this lineup does that, to the best of my knowledge. Unfortunately I didn't try it, so I don't know how well it works.
The user interface can't even be classed as "mediocre". You will likely have to spend some time to figure out the user interface's oddities, and if you run into a snag trying to get the card to work, the software's error messages are unhelpful. Don't bother running to the manual, either: it's very basic, and was clearly written by someone whose native language was not English.
Frame 0a, Frame 1a, Frame 2a, Frame 3a.
Frame 0b, Frame 1b, Frame 2b, Frame 3b, Bitrate and Quantization Data.
The "a" series frame grabs were done without any filtering, and the "b" series were done with filter strength 4. The filter strength tells the encoder how much to blur the input images; this reduces noise and softens hard edges, which makes the job of encoding easier. Level 1 hardly blurs the image at all, and therefore hardly removes any blockiness over no filtering. There is also a level 5, which I didn't try. You might choose not to use any filtering if your video is non-demanding or you can afford a relatively high bit rate. If you have demanding video or need to use a relatively low bit rate, you'll need to experiment with the filtering to find the best tradeoff between sharpness and encoder artifacts.
I've included some additional frame grabs because the two canonical frames don't tell the whole story. Frames 2 and 3 are the ones directly following frames 0 and 1, respectively. The DCM's quality apparently fluctuates between frames, as the encoder oscillates between using too many bits to encode one frame, and then using fewer bits than optimal on the next frame in order to meet the requirements imposed by the bit rate cap. This problem is similar to the one TMPGEnc went through for a while in its early beta versions, and the overall effect is to make the picture look worse than some individual frame grabs would imply. A consistent quality level looks better than a highly-variable quality level, even when the variable-quality encoder has moments of brilliance that outshine the consistent encoder. I think this is because large fleeting artifacts stick in the mind, whereas you eventually begin to ignore consistent levels of smaller artifacts.
This package is a fair value: not wonderful, but not terribly overpriced, either. It would be more compelling if it didn't have all the competition it does. The main reason to choose the DCM is the hardware encoding assistance, which is useful if you have a slower PC. If you have a fast PC, a realtime software encoder will probably give you better results, and will be cheaper besides.
Usability: 6
Functionality: 8
Quality: 6
Core Value: 7
Bundle value: 10
Overall: 6.5
The DCM comes with Vitec's $300 (street) DVD Toolbox. This includes the MPEG Maker off-line encoder, an MPEG editor, and a bunch of other lesser tools. This bundle is very useful, and worth $300, which is why the "bundle" score is so high. The "core value" score comes from the fact that if you subtract the cost of the bundled software you get $200, which isn't bad for a decent real-time hardware MPEG encoder. Compare the quality of the similarly-priced Hauppauge WinTV PVR.
| Updated Mon Sep 22 2008 12:15 MDT | Go back to MPEG Encoder Reviews | Go to my home page |