[tbritton]

I finally found the developer's page for Dolby, and here are a series of articles on the various formats, including Dolby Atmos:
 
http://developer.dolby.com/news.aspx
 
Only part one exists of this 3-part series at the time I am writing this.
 
I'm sure poking around in http://developer.dolby.com will divulge more information as I look around in there, but their article was interesting, even though not vert informative as to our binaural/hrtf interests.
 
Terry
 
EDIT - NOW WE'RE TALKING! A free encoder for small videos or audio-only files to Dolby Digital Plus. Interesting, though not sure what this has to do with the new "Spatial Surround" options.
 
http://developer.dolby.com/tools-tech.aspx#windows  (click the "tools" link or scroll down)

 

[tbritton]

More discoveries at the Dolby developer site.
 
So I am not double posting, please see this post where I discuss that.
 
http://www.head-fi.org/t/844611/dolby-atmos-for-headphones-now-available-in-windows-10/15#post_13449516
 
You might want to follow that thread as well if this stuff interests you!
 
I just received a nice lengthy reply from Richard Furse to my information seeking email sent to Blue Ripple Sound, the high-end Ambisonics encoder/decoder provider who offers that free Core O3A set of tools for Reaper. Some of his reply insights are in that post. Nice to hear from him! He wants to be kept abreast of anything we discover here!
 
Terry

 

[immersifi]

So, the site says the following (shortened to highlight a few key bits, emphasis added by me):
 
"...It includes integrated spatial sound on Xbox and Windows, with support for both surround and elevation (above or below the listener) audio cues. ...The spatial sound APIs allows developers to create audio objects that emit audio from positions in 3D space. Dynamic audio objects allow you to emit audio from an arbitrary position in space, which can change over time. You can also specify that audio objects emit sound from one of 17 pre-defined static channels (8.1.4.4) that can represent real or virtualized speakers. ...Spatial sound with Windows Sonic supports TVs, home theaters, and sound bars that support Dolby Atmos. Spatial sound can also be used with any pair of headphones via Windows Sonic for Headphones or Dolby Atmos for Headphones."
 
This sounds a lot like some of the VST plugins that I have seen from noisemaker, VVAudio et al. As I understand things thus far (and I am a newb to A and B format, so if I make mistakes, forgive me) one can define a multitude orf arbitrary speaker locations once the A-format signals have been converted to B-Format, again, because one needs to start the process with three orthogonal (and signed) transducer arguments (it also helps to think of the three signals in three dimensions themselves...to think of them as surfaces rather than a planar shape) if one is to synthesize arbitrary angles and patters - again, by varying the ratios of the three arguments relative to each other.
 
It's interesting that they (Microsoft) explicitly mention elevation as this suggests rotation about all three othogonal axes, so for head-tracking based material, this would be pretty important (or can be, depending upon how diffuse is the space that is being emulated (synthesized) or used (i.e. where the A-format signals were acquired)). In principle B-format is being used by the VST plugins that I mentioned to facilitate these very things - arbitrary microphone patterns as well as being able to 'steer' the microphone patterns in post - I mentioned this being akin to a flashlight of sorts that could be moved about. In other words, the fact that they state "...which can change over time..." means that one can steer virtual sources in terms of azimuth or elevation so that the sources are in lockstep with where they are presented visually - at least this is my guess.
 
B-format plug ins allow a great deal of manipulation, again because the format comprises three orthogonal velocity signals. Thus, generating arbitrary-angle X-Y pairs, Blumlein and semi-ORTF et al can all be derived from the B-format signals (and with the right plugins using the SOFA data, to emulate what should be pretty accurate binaural renderings (again, depending upon whether the HRTFs / HRIRs are near or far field. In other words, I can see how (in a game for instance) choosing a far field SOFA dataset for a given head when a nearfield SOFA set would make more sense could cause some 'cognitive dissonance' between the auditory and visual cortexes. On the other hand, it seems that such a things should be moot if the head-to-virtual object is on the order of 2 meters or so. Were I synthesizing a live concert binaural playback from a B-Format signal for instance, it would make no sense to use the near field SOFA set for that particular mannequin unless the B-format signals were derived from a soundfield mic placed in extreme proximity to the source.
 
Actually, B-format has a lot in common with M-S in the sense that one is relying upon various polar arguments (either added or subtracted) to generate other polar arguments, that is, patterns. There's a great paper written way back when by Wes Dooley about M-S, and reading it really helped me understand how all first order microphone patterns are derived from a pressure component (an omni mic) and a velocity component (a figure-8 mic). If you look at the gain matrix to take A format to B, it is not rocket science. 
 
​Fascinating stuff. As always Terry, thanks for moving the thread along.
 
Mark
PS: If you want to see the noisemaker VST plug in that uses SOFA and B-format, here's a brief 'tease' video from them:
 
http://www.noisemakers.fr/ambi-head/
 
http://www.noisemakers.fr/binauralizer/

 

[tbritton]

Interesting seeing Noisemakers includes the SOFA data in its Binauralizer and Ambi-Head! Loved that video! Counting the input channels, this would be a "classic" 1st-order Ambisonics tool, similar to how the "Reaper Ambisonics Toolkit" is 1st order (4 channels).
Blue Ripple Audio has something similar to Binauralizer, but I did not see any mention of SOFA tools. They do boast having 3rd order (16 channel) tools, though. http://www.blueripplesound.com/hoa-introduction

Since they are so responsive to email, I'll write and ask him about that later today. (I just wrote them last night, which Richard kindly answered this morning, so I don't want to seem a pest!)

Terry

 

[immersifi]

Cool.
 
By the way, here's a nice paper (a link to it anyway) on the whole Lebedev procedure, and the caveats. This is for the KU 100 subjected to such measurements, but for near field. A lot of stuff of which to be aware when making such measurements:
 
https://www.researchgate.net/publication/310600534_Measurement_and_Perceptual_Evaluation_of_a_Spherical_Near-Field_HRTF_Set
 
You can download the pdf from here as well.

 

[immersifi]

  Here is the link to the Dolby Atmos Creators Update page.
 
Interesting statement here:
 
And here is Microsoft's "Spatial Sound" page with the API stuff for programmers. Wondering how any of this ties in or will tie in to the Ambisonics stuff, especially since YouTube and Facebook video support the SN3D formats already.
 
Terry

I quoted Terry here simply because of the spatial aspects of this thread.
 
I found this link to the paper by Wes Dooley and Ronald Streicher: http://www.aes.org/e-lib/browse.cfm?elib=3814
 
If you look around the web you can probably get it free of charge (it is free to AES Members and those subsribing to the AES Library).
 
It really is a great read, because the Ambisonics aspects are definitely related; the idea of deriving one polar response from constituent parts is addressed. Mind you, these are for first order microphones only, but the underlying mathematical principles are the same.
 
I love this paper in that there are pages of charts that show what happens when a pressure transducer and a velocity sensor are combined in various ratios. Unfortunately the plots are not in terms of dB (they are linear, not logarithmic relationships), but if you start thinking in terms of logarithms it will probably make more sense.
 
M-S is indeed a remarkably powerful 'simple stereo' technique but again...there's a lot of similarity between it and Ambisonics, in terms of the underlying mathematics.
 
I highly encourage people to find the paper and read it. It won't necessarily help people better understand HRTFs, but nevertheless it's very well written and also, pretty important.
 
Enjoy...
 
Mark

 

[MagnetMadness]

And here is Microsoft's "Spatial Sound" page with the API stuff for programmers. Wondering how any of this ties in or will tie in to the Ambisonics stuff, especially since YouTube and Facebook video support the SN3D formats already.

Terry

I've been playing with the Sonic Spatial Sound stuff today. On the page you linked above, the following piqued my interest: Atmos Headphone can support 16 static audio objects and 16 dynamic objects, whereas Window's Sonic Headphone can support 16 static objects and 112 dynamic objects.

I've downloaded and been running some of their demonstration applications and can confirm that it works quite well, and that the spatial audio through headphones does NOT use 7.1 virtualization (turning it on makes it sound much worse). The intention is having Sonic Headphone on is that it will decode spatial audio. the 7.1 virtual surround is for content that is not supported by spatial data. If there is interest I can record a video of 128 sound objects whizzing around. The default demo is a bunch of bugs over the rainforest thunderstorm atmos demo, but more than 10 bugs / frogs / random jungle noises flying around gets overwhelming as many move very fast. I may try to slow them down or replace the sounds with bullet effects or hums of different frequencies.

 

[howsbasaurus]

I am nouveau-mélomane. Can you update this thread with your preferred YouTube/Spotify/Apple Music/Google Play Music/Deezer/Tidal recordings?

 

[tbritton]

I am nouveau-mélomane. Can you update this thread with your preferred YouTube/Spotify/Apple Music/Google Play Music/Deezer/Tidal recordings?

As I receive recommendations of interesting Binaural recordings, I place them into the first post in the topic. Go mining in there and you should find many surprises. Sorry if some links are dead - I haven't checked in several months.

Terry

 

[cleg]

well, it's definitely not binaural recodings, so I'm sorry for an off-topic, but this thread is probably most alive one here, so I'll post this link: http://www.polyphonyproject.com/en/introduction

it's a nice collection of different Ukrainian folk songs, collected in different regions and having pretty good recording quality with nice staging. so, may be they'll be interesting at least for someone

btw, if you'll scroll song's page to the end, there is a dedicated player, that allows to select separate performer's voice

 

[PETEBULL]

Ariana Grande 3D Experience Album
https://alletrailer.wixsite.com/leakgrande
downloadable

 

[tbritton]

Ariana Grande 3D Experience Album
https://alletrailer.wixsite.com/leakgrande
downloadable

Thanks, Pete! I look forward to listening to this!

Terry

 

[immersifi]

I have finally decided to make much of my catalog of sounds - many recorded binaurally (KU 100 mannequin head mic), others recorded in conventional stereo - available via bandcamp. Here's a link to one that is probably timely for this time of the season - at least in the northern hemisphere - the sound of a wood-burning fireplace with a 'healthy' fire burning. This one is binaural and as such the track title (and also file name) bear the identified [headphones required].

You can audition it up to three times but after that if you still would like to hear it you will need to pay for the file.

I'll be adding more as the days progress as a means of helping to recover my investment in the mannequin head and associated recording gear. I have also placed a collection of aviation sounds (recorded at EAA Airventure 2016 in Oshkosh, Wisconsin). Some of those tracks are pretty great in that they feature significant dynamic range as well as being binaural.

Anyway, I hope you like it. Here goes:

 

[immersifi]

Thanks, Pete! I look forward to listening to this!

Terry

More binaural flac goodness...

These just posted on Bandcamp. I put together a natural 'sounds of summer' record that you can download as a whole, or shop tracks a la carte.

I decided to start it out with a track of fireworks, and while this is track is not a fireworks finale of epic proportions (it isn't), what is cool is that they are spaced far enough apart (in time) so that you can hear the concussive 'thud' of the launch, and then a few seconds later, the actual firework exploding. Lots of cool immediate but also echoed / decaying sounds.

NERD BONUS: All of the art for each track is based on a STFT (Short Time Fourier Transform, another name for a sonogram / spectrogram) looking at one ear's response. I changed each track's FFT parameters (FFT size, percent overlap etc) to sort of 'showcase' interesting elements of each track. Again...I did that for the nerds (though it does kind of render a cool (and unique) image for each track.

The firework track is followed by a track of wave sound (made at the shore of the Lake Huron near the 'thumb' of Michigan) and then four fifteen-minute segments of crickets, frogs, and what have you, recorded on a summer night a few years ago. The tracks were recorded somewhat close to a creek that is at the far end of my lot line.

So no, no Beethoven's 9th in these, but the firework track is a good demo of spatial as well as low frequency extensions: note: not every launch of fireworks are concussive, but they are wildly different and do sound pretty cool.

The album's cheap ($4.99) and the tracks are a standard fare at $0.99. You can audition each track up to three times before they will no longer stream.

Next up...I'll be releasing about an hour or so of thunderstorms recorded binaurally. Stay tuned for that.

Mark