Smyth Research Realiser A16

Discussion in 'High-end Audio Forum' started by jgazal, May 7, 2016.
  1. sander99
    Sahmen, I get the feeling you possibly misunderstand something here. As if you are thinking the filter-information (PRIR) is somehow moved to the Yarra. Just to make it absolutely clear: this is not the case. The A16 does what it always does, generate a 2 channel signal using the PRIR, the 2 channel signal goes to the Yarra (via spdif, analog stereo or stereo over HDMI), and the Yarra does not do any binauralisation (in this case, using it with the A16), it only beams 1 channel to the left (and only the left) ear, and the other channel to the right (and only the right) ear effectively replacing the headphones. (Although the channel-seperation will not be as perfect as with headphones, but good enough for it's purpose). As a consequence, the limitations of the Yarra's decoding capabilities are unimportant, because the A16 does the decoding and the binauralisation all formats that the A16 can handle, all upmixing of the A16, and up to 16 simulated speakers (on every conceivable position, at the ceiling, on ear level, on the ground if you wish, just measure it where you want it) are possible.
     
    sahmen likes this.
  2. sander99
    Nice update by Smyth. Ok, some more delay but we already expected that (at least I did), but they are making progres. The user interface is looking good I think.
     
  3. raband
    Quick query guys on what supporting hardware to look at getting.

    Am probably looking at getting an Oppo Blu Ray player to feed the new toy and was wondering if I'd be missing much/anything if I chose the 203 over the 205 for movies (rather than music)

    Or would there be a different blu ray player with all the bells and whistles I should be looking at?
     
  4. sander99
    I can not look in your wallet, and I don't know exactly what all the other options are but for myself I find even the 203 rather expensive for a device that mainly has to deliver digital data over hdmi. You would be paying for high quality audio and video DACs and analog circuitry that you will never use is my feeling with expensive players.
     
  5. J.Villain
    That link to the UI PDF finally gives us some thing to talk about other than other peoples products. Probably the thing I found most interesting is this

    "The screen directly below this shows the information for a 32-loudspeaker dummy head BRIR
    extracted from a BBC public domain SOFA depository. This file contains 32 different loudspeakers, as
    defined by ITU-R each captured over 24 look angles covering +/-180 degrees. Use of this type of PRIR
    file with the A16 will therefore provide full 360 degree head tracking"

    24 look angles? Didn't know the A16 could handle that. How many did you guys do when you got measured at one of the audio shows? My understanding is theA-16 head tracker won't go any where near that far. But maybe Smyth has some thing up their sleeve for after the units finally ship. And what other measurements are out in the wild. I know the main Smyth dude (forget his first t name) once worked at the NRC in Canada and they did some amazing work on audio. I wonder if they did any thing there that could be used as a source for BRIRs?

    Need to read that PDF again.
     
  6. jgazal
    Afaik, that is with two A16 units working together. The A8 did work in tandem up to 16 channels:

     
    Last edited: Nov 14, 2017
  7. Sanctuary
    I am getting the 203 as well once the A16 ships. The A16 is literally the only thing preventing me from having a UHD player right now and getting the full use out of my B6 OLED. I feel like the 203 might be somewhat overpriced though for what I'll be using it for, but within its price range, it's still said to have the best picture quality of all of the players. So I'd rather pay the premium than pay slightly less and not get what I could be. Plus, I think it's one of the few players right now that also supports Dolby Vision. The 205 doesn't seem worth it at all, unless you aren't going to be bitstreaming and need its audio components to be used for whatever reason.

    The LG UP970 also supports DV and has its own separate audio out over HDMI as well, and it's almost half the price.
     
    Last edited: Nov 14, 2017
  8. sahmen
    I haven't received this e-mail, even though my pre-order of the A16 has been complete since October 13th. Do I still need to request that my address be included in Smyth's mailing list so that I can receive such updates? In every other case I have been involved in, completing the preorder has automatically gotten me placed on all important mailing lists for updates.

    I have sent a note to James Smyth, about this but I have received no response. I have also addressed a similar concern to the realiser-a16@smyth-research.com address, and I am hoping I get some kind of response from there.

    In the meantime, if someone can post a copy of the recent e-mail update on this cite, it would be highly appreciated. Thanks..
     
  9. raband
    Copy below

    Hi,



    Realiser A16 Update 14 Nov 2017



    Hardware.



    We received 25 sets of assembled PCBs as production samples a few weeks ago, and have now completed testing them. Some minor changes were deemed necessary and therefore full production and assembly of the boards is now scheduled to be completed by the end of the first week of December, with delivery about a week later.



    We recently received production samples of the metal enclosures, both the HS and 2U versions. Overall we’re pleased with the quality of the finish of both designs. All of the production metal-work has now been completed and is awaiting inspection before being shipped to us. Shipping will take 3-4 weeks.



    We’re still waiting on samples to arrive of the plastic enclosures for the headtracker and binaural mics etc. We’re aiming to visit this supplier soon when we travel to China to inspect the metal-work and test some of the assembled production pcbs.



    All of the production hardware (with the exception of the plastic parts) should be in the UK before the Christmas break, which will allow us to begin the assembly and testing of all the Realisers.



    Firmware



    We are continuing to make steady progress on the Realiser firmware – primarily on the user interface. Almost all of the underlying dsp and micro-controller code has been written and tested extensively, and the UI is simply trying to organise this and reduce its complexity. Screenshots and captions of some of the menus are shown below and more information is available on our website.



    https://eur01.safelinks.protection....gXgszONrjOKqfhAbVH1Rj76yL8oSekVR0=&reserved=0



    Whilst it may look somewhat complicated, our goal is that on power-up the Realiser should simply work – using either the user’s own personal PRIR or a factory default.



    Once the firmware and UI is complete we can use the sample hardware we currently have for compliance testing and certification. We estimate that compliance testing could begin in mid-December. Other samples will be available for sending to reviewers and selected dealers. When compliance testing is completed successfully we can begin shipment.



    Exchange Website



    The core functionality of the Realiser Exchange website is nearing completion and will be ready to go live when the Realiser hardware begins shipping.



    Once all the production hardware has been received we’ll be in the home straight, since we are in sole control of the firmware and software development. The production samples we have received of the metal-work and assembled pcbs have been excellent and, once the firmware is completed, these samples will allow us to do all necessary compliance testing. One issue is our plastic parts supplier and we’ll be addressing this shortly – however at the moment this isn’t delaying the project.



    Regards

    Smyth Research
     
    sahmen likes this.
  10. sahmen

    Thx.
     
  11. sander99
    It depends how you count them. In my demo they measured a 7.1.4 speaker system for three lookangles (only azimuth 0°, 30° left, 30° right, no elevation), in a way you could say they measured 12 x 3 = 36 lookangles. The A16 can do 16 speakers, and for azimuth and elevation 5 lookangles per speaker, so in a way you could say it can do 5 x 16 = 80 lookangles. Note I don't say at this point what you can do with this or how it is related to what is being said in the update, I am just brainstorming here.
    In the normal A16 operation as they described so far the compensation for head movement works for up to 30°, but the actual head-tracking device of the A16 can handle [Edit: detect] much more: 360 degree tracking on both axes (mentioned on the kickstarter page).
    I don't know what exactly they are gonna do with this, but in any case: even if the A16 does not do something with all these together at once, at least it is kind of a set of building blocks with wich you can put together a virtual speakersystem with certain lookangles (choose a suitable subset). But who knows indeed what the Smyth brothers have up their sleeve.

    Brainstorming further: Now I am thinking for example if you put only 6 speakers around you at ear height and regular angles between them (60°), and the speakers are all identical and the acoustics of the room symmetrical (let's for simplicity say a round (standing cylinder shaped) room. Then you could simulate a sound coming from such a speaker from all horizontal directions. starting with the position exactly at one speakers location, you can interpolate using that speakers HRTF up to 30° right to it's position (where effectively you use it's measurement 30° right), then switch to the next speaker (that sounds identical) now using it's 30° left measurement, going further interpolating up to it's 0° measurement, etc. This idea can be used to provide 360° azimuth headtracking [Edit: or rather: to provide the correction for 360° head motion, for in the context of this post the difference between tracking head movement and compensating for it is important]. What I am now describing for 1 speaker can also be done for 9 speakers simultaniously (the 9 "ground" speakers of a 9.1.6 setup). All the necessary underlying ingredients are there in the A16, it is just a matter of programming it. Maybe this idea can be generalised to 3D to 360° on both axis. If you use 12 speakers oriented as being in the mid-points of the planes of a regular dodecahedron.
    [Edit: The measurements for both the 6 and 12 virtual speakers in these examples can of course be done with the single-speaker measurement method, measuring the same speaker with your head in different orientations. In fact thinking about this makes you aware that for example 3 look angles per speaker for 6 speakers or 18 lookangles for one speaker can in fact be the same thing, it is just a matter of how you look at it. (Only there can be a difference if speakers are in different locations in the room, where the acoustics play out differently.)][Edit: actually we don't need 18 but only 12 lookangles in the example with 6 speakers, the lookangle 30° right for each speaker equals the lookangle 30° left for the next speaker.]
    Now I know this is not exactly the same as what you are talking about, it is just an observation of something that could be expoited in some way: since one A16 can do a full simulation for 2 persons it could in fact simulate 32 speakers (just add the two headphone signals together), only with the limitation that in total only 16 input channels can be used (so for example at any given moment just use 16 of the 32 speakers, but not the same 16 speakers all the time, or send each input signal to 2 speakers).
     
    Last edited: Nov 15, 2017
  12. jgazal
    I see. Fascinating.

    https://github.com/bbc/bbcrd-brirs
    https://www.itu.int/dms_pubrec/itu-r/rec/bs/R-REC-BS.1116-3-201502-I!!PDF-E.pdf
    https://www.itu.int/dms_pubrec/itu-r/rec/bs/R-REC-BS.2051-1-201706-I!!PDF-E.pdf

    Definitely will try to measure one speaker at 32 coordinates to personalize such type of PRIR.

    So as I understood a 32 PRIR is set just to play Atmos/DTS/Auro, but increasing head tracking range to its maximum.

    Does the “Illusonic N:M channel upmixer” upmix from surround codecs (Atmos/DTS/Auro) or second order ambisonics to 32 channels?

    Could the “Illusonic N:M channel upmixer” upmix regular stereo to 32 channels?

    Did someone heard a demo with Illusonic upmixer on with Atmos/DTS/Auro input, 32 channels upmixing and then auralisation to binaural?

    I think in the interview Stephen Smyth gave to @Hugo S, he mentioned trying the Realiser with 4 lower speakers, 8 horizontal speakers and 4 height speakers, but I don’t remember if the content was ambisonics. Since the Realiser does not have yet an ambisonic decoder I am sure such decodification must have been done with an external processor.

    If you are right, I hope then in the future we can load a b-format ambisonics and decode it to a third order ambisonics.
     
    Last edited: Nov 15, 2017
  13. sander99
    I looked at the illusonic website and only see up to 16 channels mentioned. So I expect the answer is no. Anyway, for clarity: I just meant to say that the A16 in theory has the capability to simulate 32 loudpeakers - otherwise it could not do simultanious simulations of two 16 channel systems using different PRIRs - but that doesn't mean that it is automatically possible to simulate 32 speakers playing 32 different signals (either by some upmixing or by somehow using more input channels). And I just speculate/hope that maybe, just maybe Smyth could implement something to somehow make use of that capability.
     
    jgazal likes this.
  14. jgazal
    Got it. With two A16 units and head tracking slave output, I believe you can handle 32 analog channels with external decoding. Not sure if two units could handle 32 digital channels nor that one unit could handle an ambisonics b-format that encompass third order information. Fingers crossed!
     
    Last edited: Nov 15, 2017
  15. Erik Garci

Share This Page