Binaural audio processing on rendeer.
When saving as a rendered audio file I would welcome Binaural processing of the stereo panning if possible. ?
I think the best place to have a check box for the option may be in the Export section of the preferences.
"Process stereo panning Binaurally when rendering to audio"
Or maybe at the mixer.
I am suggesting the process at render as I believe it could be CPU intensive for realtime binaural effects. ?
Not impossible though. It would require variable millisecond delay of each voice from one channel to the next depending on the extremities of the pan position...
So... The reverb of the synthesiser would have to be employed cross channel for each voice.
Ideally the reverb utilised for the binaural realism independent of the overall reverberation of the room.
Maybe you don't even need reverb to create the effect due to the miliseconds involved ? maybe a very simple track duplication the required miliseconds apart at the required lower velocity mixed into the opposite channel ?#
yes :) ... that's how I would do it (if I could code)
Comments
Thanks for mentioning Binaural audio. I didn't know about this technique. For those who are not familiar with it, binaural audio processing is used to improve headphone listening of stereo recordings.
Not sure, MuseScore is the best place to implement such a feature. You could already save the MuseScore output as a wav file, load it in and use a LADSPA plugin such as http://bs2b.sourceforge.net/download.html to post process it.
To do it in realtime, you could use JACK audio output in MuseScore and redirect the output to the LADSPA or VST plugin, and then record the result. Support of VST or LADSPA plugin would also be a way of getting the result. It has been requested a couple of time.
In reply to Thanks for mentioning by [DELETED] 5
Your absolutely right... but my meaning of binaural relates to my knowledge of the jecklin disk.
The Jecklin disk itself a refinement of a patent regarding binaural sound by Alan Blumlein (so I have read)
I use a Jecklin Disk when recording acoustic ensembles such as Choir, Brass, Big Bands and although I have not done so myself the technique is also deployed when attempting to accurately capture a true stereo image of a full Orchestra
OSS or Optimal Stereo signal.
As I understand it ... when you create a stereo recording utilizing the Jecklin disk process the result is classed as binaural.
In my recording experience (although there are many ways to make great stereo recordings) using a Jecklin disk is the most fuss free 'works every time' in 'any location' mainly because it accurately mimics our ears or more interestingly the mass between them. It is the difference between the same sound arriving at your closest ear to that of the ear further away from the sound source which allows our brain to pinpoint where an instrument is located within a stereo soundscape.
My point being... binaural processing of the stereo panning (inducing relative delay on the weaker side of the pan) will create a more realistic stereo soundstage for ensemble and orchestral compositions. So... its not just a headphone thing.
In fact the headphone thing is quite different.
The Jecklin disk itself is beautifully simple. recreating what it does virtually by a PC mixer could also be simple. but the rewards sonically would be fantastic :)
I have just looked into the
"Bauer stereophonic-to-binaural DSP."
As it deals with fully mixed sound is seems to filter/split frequency's first before it applies delay etc.
So... if this code were incorporated into the MuseScore mixers individual pan you would have no need to first split frequency's
so... I guess this dsp code could be adapted as the whole point of individual panning (soundstage location) is to have the power to place individual voicing as accurately as possible within it.
The original Headphone code variables will probably be more extreme as well...so would have to be recalibrated for loudspeakers.
additionally,
think of a flat semi circle
With your PC loudspeakers either side of the straight edge.
Your head looking centre on.
I also propose a 2nd slider to the pan option 'distance' backward and forward of the left/right pan.
This will enable you to position a voice anywhere within this semi circle.
And all of this is independent of reverb
'distance' in this event would not be calibrated by reducing volume but by combined delay of left and right channels pre cross channel pan delay.
Same concepts as in computer game design.
Its the physics of sound.
This would be a audio physics engine in musescore to accurately position voices in a stereo soundstage.
damn... I wish I could code.
for what its worth the man for the job could be 'Jax' formally of disney and electronic arts. Fluent in assembler and specialized in audio throughout his career.
If its a feature MuseScore wants I could find out if he is available/interested.