Hi J?rn (not sure what the character '?' is as it always displays that way)
Wow, thanks for al the info! > still holds for ambisonics. try to get as many different room modes as > possible. This is good news, and obviously what I presumed but it is the idea of the same response from each speaker that is threw me off that train of thought. > my setup has its front speakers close to a wall, and the remaining short > reflections are compensated with FIR filters to some degree. the sides > are against a bookshelf and free-standing in the room, with very > different acoustic loading and hence vastly different FIRs. the rears > are wedged between sofas. > > that makes my front direction the most "analytical", and the system > nowhere near isotropic. but it sounds very good. i just know that when i > want to dissect something in detail, i rotate it to be in front. > > unless you can afford a purpose-built room like the (mostly heptagonal!) > listening room at CCRMA (which, despite its very modest speakers, is > quite amazing - goes to show the importance of the room), some > pragmatism is called for :) Yes I think that it still will be front centric, and it seems that having a larger area behind will make it even more so. I actually hadn't thought of the rotation trick, (even though I use it regularly while mixing). That really means I can be a bit more pragmatic with the space. It will be purpose built though, it was just the shape I was having a problem with. It will be completely sound proofed and have all walls treated with sound absorbing materials. > symmetry between left and right of the most frequent listener > orientation is still a good thing. Agreed > central to the speaker system, yes. there is no benefit to being in the > exact center of the _room_, though. i'd go for some front-back asymmetry. Central to the speaker position but not the room would be fine in a cheese wedge space that I have but symmetry of back and front is where it gets difficult, and probably means it isn't such a good idea to have the back higher than the front. Maybe heptagonal is the better, wasted space option.... > a sphere would be absolutely disastrous, unless it is anechoic, and then > the shape does not matter anyways. and as aaron pointed out, overly dead > listening rooms lack proper masking of interference artefacts and will > be very irritating to work in. > > the way i approach it is: > * keep the early reflection paths clean for every speaker, like you > would for stereo. no reflections < 10 ms is a good thing, if possible. > * keep the diffuse field under control. off-axis mud adds up as you add > more speakers, so proper bass absorption and diffuse reflection in the > treble and upper midrange are important. > * use mild digital room correction in addition to acoustic treatment, it > can do wonders for bass problems, where mechanical measures are difficult. > * if you have to make compromises, keep the frontal direction as perfect > as possible, and use it as a "magnifying glass" to work on details even > if the respective sound later moves elsewhere. This is interesting, as I have had various opinions on this. Some people say that spheres are the best as they have no parallel sides, so reflections are reduced. They also only have one room mode, that can be predicted and treated. Or not excited (depending on the size). I do know of speaker box technology that uses this thought to it's advantage, but I have never considered it for studio construction, due to complexity and space. It would also have to be very large for the lowest fundament not to excite it! I must say, I like dead rooms, although I do agree that they are not the best places to work. In fact quite disorientating. Listening to ones own body internally is very off putting. As a consequence I generally make the front complete dead with absorb-tion materials and then have the back handle for reflections via random breakup reflectors. Is this still a good idea? A little room correction will of course be needed, especially for bass. > an off-the-cuff suggestion: > * four subs in the corners. > * the fullrange speakers on a horizontal ring, with one speaker in > front, for a decent approximation of ITU 5.1 and 7.1, if necessary. > * the satellites in a lower ring-of-eight, an upper ring-of-eight, > another ring of six, one zenith speaker. > then you have two spares, and they will come in handy some day. > > the bass management will be tricky. first of all, each speaker needs to > be perfectly delay-compensated to the listening spot. then i'd try to > create different layers of decoding: > > * separate first-order decode for the subs, low-passed at 60, 24dB/oct > * fourth-order decode for everything else > * horizontal speakers high-passed at 120/24 > * satellites high-passed at 120/24 > * a separate horizontal-only decode (of the same full-sphere input > signal) for the range from 60 to 120 hz, again at 24dB/oct > > this lets you drive all speakers to the best of their abilities, and > puts the missing bass frequencies in the correct direction. $DEITY help > you if anything is not perfectly phase-aligned, though. > > disclaimer: i've toyed with such hacked-up multiband setups, but none of > them ever went to production (or had to), so there may be pitfalls i've > overlooked. First order decode for the four subs in the corners was what I was thinking. Didn't think about going to fourth order on everything else though, as I didn't think the increase in channel count was worth the little improvement. I also want to leave some processing power for mixing plugs (I use a lot) :) Agreed on the full range horizontal ring. I was more thinking of a dodecahedron for the satellites, either only 20 on the vertices, or get 5 more, and would it be possible to use the edges? Is it better to use platonic solids, or doesn't it matter? The Satellites actually go down to 80hz really, i've just been using them at 120 in the current set up, generally due to the response of the room, so I could actually cross them all here. This is of course where I may get some dodgy response. I will also delay compensate to the sweet spot Thanks again, and needless to say I will be asking a few more questions as I progress. The build won't start for another month, and when it's finished I would love for all you ambisonic heads to have a listen. Cheers Steve -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://mail.music.vt.edu/mailman/private/sursound/attachments/20140310/8173e2f6/attachment.html> _______________________________________________ Sursound mailing list [email protected] https://mail.music.vt.edu/mailman/listinfo/sursound
