Export & Publish 360VR + 3D Spatial Audio CORRECTLY on YouTube VR | the 2020 Guide

Image for post
Image for post

Today, we are going to learn how to export and publish Top and Bottom Stereoscopic 360 VR video with CORRECT 6-channel ambisonic spatial audio for YouTube VR. We are focusing on VR headset delivery. But it will also work great for browser and phone if your viewers are wearing a pair of headphones. And yes, 6 channels, not just 4 channels. There is no video guide at all on the Internet to show you how to leverage all the great features YouTube built for ambisonic audio. I hope this tutorial answer all your questions.

To tell a good story in VR, visual is only 50%. Immersive audio is as important as 360 visuals for a good VR headset experience. As Oculus Quest, Oculus Go, and most of the VR headsets are consistently sold out, VR is so HOT right now during the pandemic. As a storyteller, you should learn how to get on one of the largest platforms out there, the YouTube VR. Not like Oculus TV. YouTube VR is available on ALL VR platforms, including Oculus, SteamVR, PSVR, and so on. If you are going for reach, you should be on YouTube VR, like yesterday.
But, almost no independent content creator knows how to release high-quality immersive content on YouTube VR. YouTube is only helping major studios like National Geographic, New York Times, or Netflix. So, I will fill the knowledge gap so you can do it yourself. And, it is actually very easy thanks to Adobe and Facebook. You hear it right. Facebook has the FREE 360 Spatial Audio Workstation software that will help you to encode for YouTube VR and Oculus 3rd ambisonic for the Quest and Go.

Which VR camera in 2020 (professional vs consumer)

To capture 360 VR video, you need a professional cinematic VR camera like the Insta360 Pro 2 or Titan right here. Since almost 80% of my followers don’t own this expensive camera system, you can use a monoscopic 360 camera to follow along with this tutorial. A good and a very cheap 360 camera with the First Order ambisonics is the GoPro MAX. Not even GoPro themselves advertise that feature. But, you can use FFMPEG to extract 4-channel spatial audio out from the GoPro MAX to use it as your 3D soundtrack.

Here is the FFMPEG command link for extract the First Order ambisonics audio track from GoPro MAX:

ffmpeg -i YourGoProMAXfile.360 -vn -acodec pcm_s16le outputAmbiXaudio.wav

Image for post
Image for post
360VR cameras in 2020

This is a follow-up tutorial of my first ambisonic workflow tutorial right here. It covers the difference between YouTube and Facebook Oculus, what microphone to use to capture spatial audio, and how to get your sound design from Adobe Premiere into Reaper or other DAWs. We are only going to cover Export and Publish on YouTube VR on this tutorial. So watch that tutorial first and then come back for this one:

Best 360VR export settings, Adobe Premiere 2020

For 3D 360 video in 2020, I recommand this render setting for all delivery.

Image for post
Image for post
Best 360VR Export settings in 2020, Adobe Premiere

For codec, I will recommand to use HEVC H.265 for the smallest file size. For resolution, I think 5.7K, which is 5760 by 5760 is more than enough for final delivery on YouTube VR or Oculus TV. In Premiere 2020, your export resolution of H.265 can not be more than 5968 by 5968. If you want 8K by 8K render in H.265 — I will recommand use AfterCodec — which detailed in this tutorial:

In the post-pandemic era, the YouTube server is getting very slow because of the high viewing traffic. I will not recommend upload 8K by 8K 3D 360 video onto YouTube. This will take forever for YouTube to transcode your video, which will affect your release time. Even you can, technically, upload 8K by 8K video onto YouTube, you should NOT do it.
The Bitrate Settings, I usually do 60Mbps for H.265. But sometimes, this bitrate is not high enough, especially for high action video. Then you will need to use FFMPEG — as Premiere can not do more than that. Here is the FFMPEG command line for encoding:

ffmpeg -y -i “input.mov” -c:v libx265 -preset fast -crf 21 -vf “scale=4096x4096:out_range=full” -pix_fmt yuvj420p -aspect 1:1 -movflags faststart “output.mp4”

If you are using Davinci Resolve, this is what you need. Please refer to this tutorial for detailed instruction. Render for Oculus Quest and for YouTube VR in 2020 is the same.

You should check Video is VR, and pick the correct Frame Layout. Here is Stereoscopic Over and Under. Check use Maximum Render Quality. Okay, not yet, we are getting just the video from Premiere. So uncheck Export Audio right here as we don’t need the audio at all.

Now the video part is done. Easy! Now, let’s spend the major part of this tutorial on 6-channel Spatial Audio.

Design Spatial Audio from Premiere to Reaper

So a good film should have multiple soundtracks — lavalier mono source from the actors, background music, and sound effects as you see right here. This is one of my recent 360VR production, go watch it on your Oculus Quest. I intensionally pan the audio source around so you can hear the audio jump from your front to the back when the rover is moving on this part.
So how to render all these audio layers? Well, you don’t do in Premiere. You bounce all these tracks into a DAW (Digital Audio Workstation) like REAPER. I will also recommand design your spatial audio in 16 channel 3rd order ambisonic for Oculus following the first tutorial right here. When you have the 16 channel done, making into 6 channel is easy. So go watch part 1 and come back when you are done.

Okay, I assume you just finish the part 1, let’s recap the Export part in REAPER to make sure you did everything correctly. We need to render the 16 channel 3rd order ambisonic audio without the background music or any sound that should NOT pan following the viewer’s head turn. The background music and not spatialize sound effects should be sent to HL Master right here. Just to show you the routing again because it is important.

  1. In the Mixer view, click on the ’ROUTING’ button
  2. Set Track Channels to 10 or 16 depending on if you are working in 2nd or 3rd order Ambisonics respectively. (in 2nd order mode the Spatial Workstation will only use the first 9 channels out of the 10)
  3. Click on the Add New Send.. drop down menu and select 3D MASTER
  4. Under the options for this newly created send, set Audio to Multichannel source→10 Channels→1­10 for 2nd order Ambisonics or Multichannel source→16 Channels→1­16 for 3rd order Ambisonics.
  5. Uncheck/disable the Master/parent send option at the top­left of the window

For background music, the routing is stereo and to HL MASTER with Stereo 2 channels.
To export your spatialised mix:

  1. Use Reaper’s Time Selection to specify the area you would like to bounce. Make sure the selection starts at 00:00:00:00
  2. Open the Control plugin and disable Decode Binaural
  3. Select the CONTROL PLUGIN track (and no other track)
  4. Go to File→Render in the menu bar
  5. In the Render To File window that shows up, choose Stems (selected tracks) in the Source drop­down menu
  6. In the Bounds drop­down menu, choose Time Selection
  7. Specify the file name and output directory
  8. Under Options, type in 9 or 16 in the Channels option for 2nd or 3rd order Ambisonics respectively
  9. Click Render 1 file..

The spatialised mix should now be exported. Next, export the head­locked stereo mix:

  1. Select the HL MASTER track (and no other track)
  2. Go to File→Render in the menu bar again
  3. In the Render To File window that shows up, choose Stems (selected tracks) in the Source drop­down menu
  4. In the Bounds drop­down menu, choose Time Selection
  5. Specify the file name and output directory
  6. Under Options, specify 2 in the Channels option
  7. Click Render 1 file..
Image for post
Image for post
Fb360 Encoder

Now open Facebook 360 Encoder. Output Select Format is B-format (1st order ambiX). Input Select format is B-format 3rd order ambiX. Load in your spatial audio file 16 channel output from REAPER here and then hit Encode. Now we have a 4-channel 1st order ambisonic track just like what you get from Zoom H3VR or other ambisonic microphones. The only thing different is your spatial soundtrack will sound 100 times better than just directly capture audio using the ambisonic mic. And it is produced in a studio so it is free of noise.

Now we need to combine the 4-channels to the HeadLock audio master background track to create a 6-channels for YouTube VR.

Image for post
Image for post
Use Audacity to mux spatial audio

We are going to use Audacity — because it is open source, and it is free. Download and install it. First, go to file — preference — under Import / Export — choose Use Advanced Mixing Options.
Drop-in both the 4 channels and HL Master into Audacity. You should see 6 channels now. I assume you already level all the tracks and they should sound perfect. So we are only going to export them. Go to file — export — Save as type Other uncompressed files, choose WAV and 32 bit float. Hit save and now you have a 6 channels audio file ready for muxing.

Now we are back to FFMPEG. You can use Premiere, Resolve, or any video editing program to mux your video with audio. I am going to show you FFMEPG, b/c it is free and open source. Put the video file without the audio and the 6 channel audio source into the same folder of your FFPMEG bin folder. Type this command:

ffmpeg -i yourvideo.mp4 -i “your6chaudio.wav” -map 0:v -map 1:a -c:v copy -c:a copy youroutput.mov

Replace video.mp4 to your video file name and audio.wav into your audio file name like so. Output.mov, you can replace it with the name you want.

Spatial Meta Injection

Before you can upload it onto YouTube, you need to inject meta to tell Google this is a 3D 360 top and bottom video with the correct 6 channel ambisonic audio for YouTube VR. So download the Google Metadata Injector with the link down below. Drop in the muxed video file into the injector. Make sure you check EVERYTHING, My video is spherical 360, My video is stereoscopic 3D top and bottom layout and My video has spatial auido ambiX ACN/SN3D format with head-locked stereo. Hit inject metadata.

Image for post
Image for post
Spatial Media Metadata Injector for 360VR video

Now, go ahead and go to YouTube. Upload your video just like any other video. YouTube will recognize your video as 3D and with correct spatial audio. Now if your viewers are wearing a pair of headphone, their minds will be blown by you. Some kids might call it 8D Audio — like the next cool thing after ASMR. But we know it is a 6 channel spatial audio and it will be the next big thing after ASMR or 8D Audio. And now you know how to make it correctly.

Thank you for watching this pretty technical but necessary tutorial. The knowledge I shared here not only applies to 360 VR Video but also game engine design. VR game has to have good spatial audio as well.

Follow me on Social Media:

➜ Oculus TV: https://ocul.us/2PINYMz
➜YouTube: https://www.youtube.com/creatorup
➜Facebook: https://facebook.com/360creator
➜Instagram: https://instagram.com/hugh.hou
➜ TikTok: https://www.tiktok.com/@hughhou/

Image for post
Image for post

Written by

YouTuber at CreatorUp (https://www.youtube.com/creatorup)

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store