Can’t figure out what I’m doing wrong. Would like to record a quadraphonic source and create an Ambisonic mix from this. I’ve created 4 mono tracks, and have tried different types of Ambisonic buses for the Main Mix (1st, 2nd, 3rd). I’ve assigned the Ambisonic bus to my card (RME AIO) and enabled Control Room monitoring…
Each mono channel has a Multipanner. Each panner is set to send to it’s respective quadraphonic corner (Left Front, Right Front, Left Rear, Right Rear). When this is monitored, the audio seems to come appear at many places in the output channel (Ambisonic bus, meter indicator). The headphones also have no true sense of front and back, as if something wrong is happening in the decoding.
Will do. For some reason, I just feel like either my hardware routing config or my use of the Multi Panner has a user error. Will make a file, and send it over soon. I’m on 10.2 though. You on 8.3?
That link will download a zip with static waveform recordings, and gated waveforms in percussive patterns. Basically… I feel like I’m losing a sense of separation. It feels mostly just… stereo with a tiny bit of ambisonic placement if any.
So, here’s your NPR file with my slight adjustments:
In the multipanner, I placed each mono source on the “unit circle”, i.e. with a radius of 100%.
Also in the multipanner, I set the source size at 0%. This maybe counter-intuitive, but at 100%, the sound will come from everywhere.
I monitored with the IEM BinauralDecoder inserted on the output bus.
On your end now, you just have to set up a decoder to fit your loudspeaker array. I suggest the IEM AllRADecoder. Ambi Quad Test Kewl.npr.zip (27.8 KB)
It’s the binaural decoder I use across a few apps, but mainly in Nuendo and Reaper, and it works reliably. In Reaper, I work at fifth order. I prefer editing in Nuendo, but Reaper routing is more flexible for Ambisonics. Right now, for Ambisonics, the big plus in Nuendo is the native panner integrated to the MixConsole.
In any case, for decoding, I use IEM BinauralDecoder for headphones, and either IEM AllRADecoder or IEM SimpleDecoder in conjunction with IEM DistanceCompensator if necessary (when loudspeakers are not equidistant from the array center). Much more flexible than any native Ambisonic decoder in Nuendo.
With this panner setting, I only hear audio from the far left. As I rotate the panner position around the headspace, it only bleeds the signal into the other directions (Left, Right, Forward, Rear)… and doesn’t create any sense of location. I am monitoring a live mono input but have similar effect if using recorded audio.
On a side note… I was able to take a standard quadro recording and insert the Flux IRCAM HEar plugin to make a decent binaural mix. It doesn’t have quite the same depth or definition of the original but the results are much better than what I’ve accomplished using native Nuendo tools
OK, maybe it’s related to expectations. In one of your replies, you talk about how “it feels mostly just… stereo with a tiny bit of ambisonic placement”: for you, what is “ambisonic placement”?
Given my previous post/example… I’m voting for configuration problem, but haven’t found any information that helps with that. The files I downloaded from you produced a strange panning phenomenon… so not sure what to change.
“Ambisonic placement” would refer to any panning experience beyond 2D left/right panning. Clearly forward or rear, or any height… essentially any 3D audio experience. Does that make sense?
Audio Technica headphones. Trying to create binaural stereo from a quadraphonic mix… the native format is being recorded in multiple mono, and is monitored in 5.1 on a Dynaudio rig. The goal is to create a binaural mix for use in VR video content (no head tracking, but 3D audio experience)
I see. So the effect is there on the speakers, but not on the cans, if I understand you correctly. Then you will definitely need something that encodes your Ambisonics output to a binaural two-channel audio stream - hence the IEM Decoder.
If the ultimate goal is binaural without headtracking, I would bypass Ambisonic encoding altogether and use a binaural panner. One that works well for me is the Sennheiser AMBEO Orbit.
If that still doesn’t “work” for you, then you’ve probably hit the limit of binaural stereo: if the HRTF used by the binaural panner is too different from your own ears and head, the results will always be disappointing for you. But maybe not for somebody else…
You’ll see in previous posts, I was able to get something that “works” using Flux HEar. I also downloaded and tried AMBEO… similar results.
This thread is more about “Why can’t I get Nuendo to do this?” — It has Ambisonic tools, and a native 3D panner… but it’s not working for me and we haven’t figured out “why” — The technical question of how to configure and use Nuendo for these tasks remains. Documentation is limited. Most posts suggest using 3rd party solutions for successful Ambisonic mixing. In general… using Nuendo’s toolset for this task isn’t going very well.
That said… the limits of binaural stereo with both AMBEO and HEar plugins are not as detailed as I expected either. I’m hearing much better spatialization in AAA titles like Fortnite and Red Dead. Perhaps they are using an Atmos encoding or some other function to deliver 3D via headphones?
Nuendo’s MultiPanner is fine: in Ambisonic mode, it works as it should. But Nuendo’s Ambisonic decoders are a bit underwhelming, that’s why I’m suggesting 3rd-party software. As I said, if the target is fixed binaural, don’t bother going through an Ambisonic encoding: you’re not gaining anything.
The limits of binaural stereo are something that transcend a specific binaural panner or another: reading Jens Blauert’s Spatial Hearing can certainly put that in perpective.
Successful binaural panning depends a lot on the choice of sounds, the location acoustics, and expectations about the sound scene. Atmos for headphones or DTS:X are not more miraculous than, say, Sennheiser AMBEO Orbit. Some sounds, in real life or binaurally panned, will always remain difficult to localize.
