Review: Matrox Monarch HD

How I Tested

How did I test the Monarch HD? Well, pretty much the way that you would have, hopefully. That means I tested connectivity, audio sync, and quality.

For connectivity, I streamed to YouTube Live, Brightcove Live, and Ustream, all without any problems. Strangely, the older version of Livestream would not work, but that turns out to be problem that Matrox was aware of and is working to fix.

For audio sync, I ran multiple tests for streaming and recording. Starting with streaming, I ran a one-hour test on YouTube Live, where I spoke in the video every ten minutes or so I could check audio sync throughout. You can watch the video below; from what I saw, the sync was rock-solid throughout. I also ran a 30-minute test to UStream, with similar results.

To test the sync while recording, I recorded 30-minute and 52-minute videos to a hard disk connected to the Monarch via USB. In the 30-minute video, which I recorded at 720p, sync was rock-solid. In the 52-minute video, the sync was off almost immediately in the QuickTime and the VLC Players, but perfect when I dropped the MP4 file into a timeline in Premiere Pro.

I asked Matrox about this, and learned, “We have identified that QuickTime expects [an] H.264 file with [an] IBP GOP to have their file headers written in a very specific way. Adobe is more flexible regarding file headers. This issue only arises with IBP files. When B-frames are eliminated from the GOP, QuickTime plays the files in sync. We have adjusted the way we write our IBP files to be compatible with QuickTime and continue to ensure FCP and Adobe also play nice. This fix will be found in the upcoming firmware update.”

To verify that B-frames were the problem, I changed the encoding parameters to eliminate B-frames and recorded a 27-minute file to disk. Audio/video sync was perfect throughout in the QuickTime and VLC players, as well as Premiere Pro.

Quality

For the most part, with a live encoder, you produce either a single, high-resolution, high-quality file for real-time transcoding by the streaming service provider (like YouTube Live and Brightcove) or a single, reasonable-quality, SD file for distribution without transcoding (Ustream). For example, in my YouTube/Brightcove tests, I streamed a 720p file encoded at 3Mbps to the services, which they transcoded into separate iterations. For Ustream, I sent a 640x360 file encoded at 800Kbps.

Neither of these files push the envelope when it comes to H.264 quality; I expected pretty much perfect quality with both files and that’s what I got. Ditto for the 30Mbps 1080p files that I recorded to disk. Of course, I would expect this from any live encoder, so these tests weren’t particularly illuminating regarding Monarch’s comparative quality.

To get a sense for Monarch’s output quality, I encoded some test runs with my standard test file at 720p @ 800Kbps, which is the very aggressive configuration that I use to test VOD encoders. In these tests, Monarch’s quality was slightly behind the best of the VOD encoders, which is expected, given the off-line, non-real time nature of the VOD encode. The file data rates were very accurate and the data rates very flat, which is important since live encoders are typically streaming out of bandwidth constricted connections.

Overall, Monarch feels like a very competent product from a reputable company who has been very responsive to the inevitable problems that new live encoders experience in the field. The dual-stream use case--one for streaming, one for archive or other production--is very common, and the Monarch appears to be the lowest-cost solution. If it’s a use case you need a solution for, give Monarch a shot.