Virtual Reality Streaming Tools and Workflow, Part 2: Production Workflow

Configuring the Output

Now that you've got the input side all together, it's time to focus on the output side. Again, for my project, the sole output was a video to upload to YouTube. To create the file in Vahana VR, you click Configuration on the upper left to access the buttons shown in Figure 9 (below).

Figure 9. Setting output options. Click the image to see it at full size.

Click Panorama to set the size of the video file that you're capturing. When I first captured the video, the resolution was 1280x640, and the result was blurry and grainy. I asked the tech at Mobeon how to boost quality, and he recommended capturing and uploading at the configuration shown in Figure 10 (below).

Figure 10. Here’s where you set the capture resolution.

Obviously, there's a trade-off. The higher the resolution, the higher the data rate necessary to sustain the quality, and the faster the connection you need to actually play the file. Since I knew YouTube would create multiple file iterations at different resolutions and data rates, I went for broke and tried 4K.

Next you set the capture parameters as shown in Figure 11 (below). You access this by closing the Panorama configuration screen and clicking Outputs in Figure 9. As discussed in the first article in this series, you have four options: HDD output, RTMP stream, Oculus View, and DeckLink 4K Extreme. I selected HDD file output. If you study Figure 11 (below), you’ll note that some of the numbers and entries look out of place. Somehow, when I captured the screen, a lot of critical parameters were left blank, and by the time I noticed it, I had returned the system to Mobeon. So I filled in the blanks during post, so to speak.

Figure 11. The encoding parameters used for video capture

Anyway, I didn't do a lot of experimentation on the capture side; I just assumed the 20 Mbps would be sufficient. Given more time, I would boost the data rate until I pushed the limit of the encoder/disk subsystem, since as we all know, video is a garbage in/garbage out medium. Still, at 20 Mbps, the final video looks pretty good on YouTube.

Once you configure your output, and return to the main screen, all configured outputs appear on the right hand toolbar, as shown in Figure 12 (below). Once you press the HDD button, the unit start recording.

Figure 12. Click HDD to start recording.

The software records a file as you've configured it--in my case, an MP4 file. Most producers will likely have to perform some editing work on the file, in my case to sync in the higher-quality audio. My goal during import, editing, and export was to not change the resolution or frame rate in any way. I'm not sure this was necessary, but it seemed like a good idea.

To accomplish this in Premiere Pro, I imported the file, and automatically created a sequence by right clicking and selecting new sequence from clip (Figure 13, below). This made sure that the sequence used the same parameters as the clip.

Figure 13. Letting Premiere Pro create the sequence from the clip

After syncing the audio, I started the export process (Figure 14, below), choosing the H.264 format and the Match Source - High Bitrate preset, which I customized by upping the data rate, using maximum render quality, and conforming the keyframe interval to 90 frames to match the original file. I'm not sure any of this was essential to maintain the VR-ness of the video file, but I was just conforming to my typical upload practice.

Figure 14. Here are the output setting used in the Adobe Media Encoder. Click the image to see it at full size.