Google Search  |  Meridian Systems  |  Meridian Streaming  |  Restaurant  |  WiKi  |  Duncan's Meridian Info  |  Board Rules  |  Restaurant Rules
Topic Options
Rate This Topic
#266535 - 2017-08-01 22:32 Ambisonic B-format
Registered: 2013-01-27
Posts: 22
Steven Dive Offline
Harmless
Steven Dive Offline
Harmless

Registered: 2013-01-27
Posts: 22
Loc: Brighton, East Sussex, UK
Has anyone else yet managed to try out the direct B-format ambisonic mode in a G series DSP or 806Vx yet? It uses the MA1 inputs on the G series decoders, not digital.

I ask in part because it's a feature that's sort of half hidden in the handbooks and I had to ask Meridian how to use it, such as how to map the B-format horizontal channels, WXY, to specific MA1 inputs (Nb. it's W > L, X > R & Y > C, WAV channels 1 to 3, respectively once converted from the .amb file downloaded). The DSP recognises the 3 channel input as compatible with ambisonic mode and displays a slightly different set of user adjustments than you find in the usual UHJ implementation - no Row setting and the DSP no longer has to apply the (entirely hidden) focus and dominance features.

Having downloaded a good few B-format .amb files, converted them to 3 channel .WAV format in Audacity (channel 4, if present, can be ignored as it's the elevation component, Z, that Meridian DSPs can't use, though not sure of the 806V8), I failed to get my DVD players to output the 3 channels, just 2, even when they displayed the file as multi channel. Then recently I found an eight channel USB asynchronous DAC and played the .WAV files in VLC (Macbook Pro) through the DAC to my G61RSL. I have to say I am pretty impressed by the sound. If anyone has used the ambisonic decoder for UHJ encoded material and liked the sense of space and involvement then, even though never released commercially, it is worth trying B-format for the fun of it. There are some works available for direct legitimate download to which I could provide a link. These are a mix of classical (e.g. Beethoven's 4th Symphony), world music and electronic experimental, plus some test tracks. The sense of focus and immersion is well above that of the UHJ encoded music I have heard so far. The rear & side image phasyness that UHJ has is entirely gone and image stability is even better in relation to head movement. I have a good few Nimbus Records UHJ CDs in my collection for comparison.

It would be good to know if anyone is interested in trying to use the B-format decode mode or has used it.


G61RSL, G98-DH, G01, Quad 606, Quad 306 (x3), Quad-ESL63 (x 2 pairs, F&R), KEF LS50 (S), Quad 10L sub, Martin Logan Motif (CF) + miniDSP
Top
#266537 - 2017-08-01 23:44 Re: Ambisonic B-format [Re: Steven Dive]
Registered: 2009-09-13
Posts: 5,347
Ludwig Offline
President of the Imperial Galactic Government
Ludwig Offline
President of the Imperial Galactic Government

Registered: 2009-09-13
Posts: 5,347
Loc: Europe
Although I could get access to B-Format material quite easily, I've never tried this. Getting a Meridian processor to use B-Format must be possible, as it is mentioned always, but probably only one or two people know how it works. Not me, that's for sure.

Let's ask Bob. Bob, are you there?

(Aside: I have found UHJ unsatisfying and have preferred to play Nimbus albums back in stereo with Trifield.)


Ex-moderator
Sitting room: 818v3, 8kSE
Office: 818v3, Stax SRM-T1, Stax SR404LE
TV room: HD621, 518, D33
Kitchen: MS200, Genelec 6010 actives
Server: Roon/NUC/Touchscreen

Oldest audio file: 1889 Edison cylinder
Top
#266552 - 2017-08-02 12:22 Re: Ambisonic B-format [Re: Ludwig]
Registered: 2013-01-27
Posts: 22
Steven Dive Offline
Harmless
Steven Dive Offline
Harmless

Registered: 2013-01-27
Posts: 22
Loc: Brighton, East Sussex, UK
Well, as I outlined, it is fairly easy to get B-format material into at least a G series box and should also be into the 806Vx. The trick is knowing how to extract and play the files (and the appropriate hardware to do it), once downloaded, which are, more often than not, in .amb format and thus require some way to play through to the DSP. Hence conversion to .WAV as this does make things easier.

A link to legitimate B-format downloads from the Ambisonia site : www.ambisonia.com Just click the ambisonic logo buttons to get through to the download page or item.

UHJ does not work well for everyone. Even stereo does not for a few. The rear/side phasiness can be distracting on some material. Do you have any test discs, such as Nimbus Records NI1417 - Ambisonic Surround Sound Sampler - or HiFi News &RR Test Disc 2 HFN 015? These are a really big help and the Nimbus disc is most probably still available, although it may require a direct query to Nimbus.

It took me a while to get the best from UHJ and found my 6 speaker setup needed the sides trimmed by -7 dB, even though sensitivities were set in the calibration step for the G61. I found the same using a 565, so it seems the ambisonic speaker sensitivities are set independently for the sides. Hmm! My fronts and rears are the same brand and model (Quad ESL63). Sides are KEF LS50, so I did expect different setting would be needed, plus use of the HF filters. I do not use my CF speaker as it does odd things to imaging in ambisonic, although it's fine for Trifield and Dolby.

For the uninitiated, it's worth emphasising that for ambisonics all the speakers are on the arc of the same virtual circle, which means keeping the sides, if you use them, 90 degrees directly left and right of the listening position, time delayed if required to place them on the circle. I tried the sides at a more forward angle but the result was unsatisfactory, which can be seen in geometrical terms. Ideally for ambisonic, a 5, 6 or 7 speaker layouts should be regular pentagon, hexagon or heptagon, respectively, with the listener centrally placed in the (virtual) circle.


G61RSL, G98-DH, G01, Quad 606, Quad 306 (x3), Quad-ESL63 (x 2 pairs, F&R), KEF LS50 (S), Quad 10L sub, Martin Logan Motif (CF) + miniDSP
Edited by Steven Dive; 2017-08-02 14:03.
Top
#266553 - 2017-08-02 14:07 Re: Ambisonic B-format [Re: Steven Dive]
Registered: 2003-10-04
Posts: 1,932
spinaltap Offline
Knows where his towel is
spinaltap Offline
Knows where his towel is

Registered: 2003-10-04
Posts: 1,932
Loc: Bournville, UK
I'm intrigued...

I have several QS (UHJ/Ambisonic) discs dating from the 1970's, designed to be played-back on 4 channel/quadraphonic systems.

So, I don't see how such audio files would satisfactorily play-back on more than 4 channels.


My Mac can go all the way to XI - Mac user since 1988.
Top
#266556 - 2017-08-02 15:23 Re: Ambisonic B-format [Re: spinaltap]
Registered: 2013-01-27
Posts: 22
Steven Dive Offline
Harmless
Steven Dive Offline
Harmless

Registered: 2013-01-27
Posts: 22
Loc: Brighton, East Sussex, UK
Simple question but longish answer, so my apologies in advance.

QS is not the same as UHJ although the matrixing has similarities (e.g.. the phase shifters). UHJ, like QS, is encoded to two channels but thereafter the similarities end, really. QS is encoded from and decoded to so-called discreet speaker feeds whilst in ambisonic (UHJ and B-format) the speaker outputs combine around one's head to reconstruct the original soundfield as was around the ambisonic microphone. You can try putting UHJ encoded material through a QS decoder to find how it sounds to you, but it will not be as intended. Likewise putting QS through an ambisonic decoder may give pleasing results, though not as intended. Nimbus Records have produced some G-format (not a typo) material from the UHJ masters that are intended for replay using a DVD player with DTS & MLP (4.0) playback (into a virtual quad array, basically). This format is a pre decoded form for people without an ambisonic decoder. One such release is A Midsummer Night's Dream (NI9009).

I have the original Mike Oldfield Boxed quad LPs, which are QS encoded and sound pretty well using ambisonic playback, but I have never tried it over a quad (QS) system. But this is the only quad stuff I have.

In ambisonics, the soundfield is 'measured' by the mic and 'decomposed' (see Spherical Harmonic Decomposition) into a pressure (omni) signal (W) and three, orthogonal vector signals (XYZ), which are not loudspeaker feeds. This soundfield scene description is used for reconstruction of the soundfield by recombining the 3 (for horizontal, WXY) or four (for spherical, WXYZ) signals appropriate for your particular loudspeaker layout; ambisonics is speaker number and position agnostic and thus you tell the decoder how many speakers you are using and where the speakers are placed. The minimum number of speakers for horizontal reconstruction is four, which works well for smallish arrays. For larger arrays, or where the front-back ratio to left-right ratio is large, 6 speakers may give better stability for side source images. In principle, you can use as many more speakers as you like, although there is next to nothing to be gained from going to more that 8 (horizontal) for a first order (aka. FOA) signal set. Besides, I know of no domestic ambi decoders that cater for more than 7 speakers (inc a CF). The signal feed to each speaker is adjusted according to the number and placement of the speakers. For first order playback in a small space, such as my sitting room, the decoders are dual band using velocity optimised decoding (rV) below 200 Hz and energy optimised (rE) above 700 Hz, with the perceptual cues from each band kept in step to avoid them conflicting (phase and energy cues).

I hope this helps more than it confuses.


G61RSL, G98-DH, G01, Quad 606, Quad 306 (x3), Quad-ESL63 (x 2 pairs, F&R), KEF LS50 (S), Quad 10L sub, Martin Logan Motif (CF) + miniDSP
Top
#266563 - 2017-08-02 18:28 Re: Ambisonic B-format [Re: Steven Dive]
Registered: 2003-10-04
Posts: 1,932
spinaltap Offline
Knows where his towel is
spinaltap Offline
Knows where his towel is

Registered: 2003-10-04
Posts: 1,932
Loc: Bournville, UK
FWIW...

In the 1970's, Tomita recorded for RCA. RCA favoured CD-4 rather than QS.

Undeterred, Tomita recorded almost all of his albums in QS - but the album sleeve didn't convey this information due to RCA's stance on the format. It would be interesting to discover if the present-day CD's by Tomita still conveyed the QS information.

On UHJ, David Bedford recorded 'Instructions for Angels' in this format. Again, it would be interesting to hear a current CD recording via the appropriate kit.


My Mac can go all the way to XI - Mac user since 1988.
Top
#266567 - 2017-08-02 21:01 Re: Ambisonic B-format [Re: spinaltap]
Registered: 2013-01-27
Posts: 22
Steven Dive Offline
Harmless
Steven Dive Offline
Harmless

Registered: 2013-01-27
Posts: 22
Loc: Brighton, East Sussex, UK
I was not aware of either the Tomita QS material nor the David Bedford UHJ album. I have Tomita's Snowflakes are Dancing original LP, which I have not listened to in years so I'll give it a listen on the G61s ambisonic (UHJ) decode and Trifield.

Sometimes re-releases of former UHJ material are not in UHJ. Cowboy Junkies' Trinity Sessions is one of these, it seems:( but I have the original (yay!).

Time to go hunting for any stuff I still don't have as well. Am about to do some .amb to .WAV B-format file conversions in Audacity. Beethoven's 4th is one of these.


G61RSL, G98-DH, G01, Quad 606, Quad 306 (x3), Quad-ESL63 (x 2 pairs, F&R), KEF LS50 (S), Quad 10L sub, Martin Logan Motif (CF) + miniDSP
Top
#271312 - 2017-12-31 23:38 Re: Ambisonic B-format [Re: Steven Dive]
Registered: 2014-03-19
Posts: 15
FuddyDuddy Offline
Harmless
FuddyDuddy Offline
Harmless

Registered: 2014-03-19
Posts: 15
Loc: Mariehamn, Finland
Hi Steven

A rather belated response to your request for inputs, I'm afraid. Thing is, I've only within the last week or so taken my 561M out of mothballs, for the express purpose of resuming experiments with ambisonics. I don't at present have an ambisonic speaker array to play it through but after a bit of elementary carpentry to provide some new shelf-space I shall have a 'square' array (one pair front one back, but unmatched as between pairs blush).

Meanwhile, one or two comments. I was a bit surprised when I first read your OP that you seemed not to be aware that there are B-format decoders out there, downloadable for free (the Ambisonia website links to them) - or did I misunderstand and you're already fully-informed about, and have reasons for shunning, them? The point being that there's no necessity to transcode the *.amb files you refer-to to multichannel *.WAV files because with a B-format decoder (eg David McGriffy's VVAudio, which I use https://www.vvaudio.com/downloads ) you can just play them as they are. The problem remains of how to get the speaker-feeds correctly routed from there but you've already solved that for your G61RSL.

I got interested a long while back in ambisonics as such and wanted to explore further, in my amateurish way. For a few years now I've been listening, off and on, to whatever *.amb recordings I've downloaded from Ambisonia through only one pair (front of course) of speakers - which provides little if anything more than plain stereo. However I took it no further because:- 1) the *.amb repertoire though free is very sparse - not surprisingly, 2) cost, plus impracticality in a domestic setting (up to 8 identical active speakers... eek), 3) my setup has for a while now been all-analogue so to get speaker-feeds out of my computer I would have needed a multi-analogue-outputs interface of some kind (more cost). So my interest (in B-format) has remained fitful.

All the while I was ignoring UHJ. But having now just decided to have another go, just for fun (this time homing-in on UHJ) I have resurrected the 561M, bought a USB -> SPDIF converter for $39 through E-Bay, and am able to stream PCM audio into one of the 561's digital inputs. That has enabled me to play my one and only UHJ recording through the Meridian ambisonic decoder and out to my (soon to be four) speakers.
Quote:
Has anyone else yet managed to try out the direct B-format ambisonic mode in a G series DSP or 806Vx yet? It uses the MA1 inputs on the G series decoders, not digital.
To (sort of) answer to your question:-

Out of curiosity I did play an *.amb file (using VVMic) and (the 561M having no multichannel analogue inputs) having selected my one and only 561M digital source, with preset 'Ambisonic', and got sound through my speakers assumed I was hearing ambisonic playback. I wasn't: I'd neglected to disable my analogue device in Windows mixer so what I was hearing was stereo played ambisonically via analogue connections which I suppose makes my answer to your question "yes, perhaps".

But why bother? I have to disagree with your low opinion of UHJ ambisonic. To my (leaden?) ears UHJ ambisonic is flawless.

EDIT

...and besides the repertoire is far more extensive and diverse (though not free of course)


561M, Bryston 9B 5-ch power-amp, Technics SL-7 TT with P205CMK4 cartridge, PC (Intel integral HD Audio device)
Edited by FuddyDuddy; 2018-01-01 10:51.
Top

Moderator:  Carl, Duncs, ncpl 
Who's Online
4 registered (Nstzya, Akimo, _M_, JimK), 44 Guests and 0 Spiders online.
Key: Admin, Global Mod, Mod
Newest Members
OneArtist, AndrewE, Obe, PaulRP0, Zetetic
4574 Registered Users
Top Posters (30 Days)
Ratbert 66
_M_ 50
CMB Integrations - Bruce 49
Ian 38
VirusKiller 36
Top Posters
VirusKiller 12230
Fiddler 8664
ncpl 8352
Carl 7563
Ian 7350
Forum Stats
4,574 Registered Members
17 Forums
24,340 Topics
250,179 Posts

Most users ever online: 211 @ 2015-05-05 14:40
Meridian  |  Media Centre  |  Support  |  Firmware Release Notes  |  RSS Systems  |  RSS Streaming  |  RSS Restaurant