As for the CBT, you got me thinking and it's pretty easy to implement a source array with each source individually shaded (i.e. level or delay controlled). Now I still need to devise a way to bend the whole structure (forming an arc, as it should be), that shouldn't be difficult. This will be a nice feature to have.
Hi mabat
if you plan to implement it in the next beta (would be great feature!) just a comment: I did many direct sound sims for line array. Setting low pass filter individually for each tweeter (lower fc the further away from middle; 1st order butterworth) can give very nice cd behaviour. If array is large additionally delaying helps a lot (in groups is sufficient…see keele cbt papers). So what I want to say: please make filter AND delay possible for every driver
Curving (complex to build) is not necessary at all if you have some active amp channels to spent!
if you plan to implement it in the next beta (would be great feature!) just a comment: I did many direct sound sims for line array. Setting low pass filter individually for each tweeter (lower fc the further away from middle; 1st order butterworth) can give very nice cd behaviour. If array is large additionally delaying helps a lot (in groups is sufficient…see keele cbt papers). So what I want to say: please make filter AND delay possible for every driver
Curving (complex to build) is not necessary at all if you have some active amp channels to spent!
As long as each speaker "mesh" is separate or has an individual tag they can be defined as a separate element and assigned filters and delay by modifying the existing script without much effort, not as convenient as having mabat automate it thoughSo what I want to say: please make filter AND delay possible for every driver
This is an example of a Lumped Element Script, you can specify Delay and Filters in here, the LEM section of the ABEC manual explains the parameters available, here there is Tweeter and Woofer but there could easily be a reference to any Element that exists.
Code:
Def_Driving "Driving"
Value={10^((-25+20)/20)} // -25dBV + 20dB amplifier
IsRms
Def_Driver "Woofer"
Re=6.11ohm Le=0.2mH
// Rg=0.5ohm
fre=18.6kHz ExpoRe=1.633 ExpoLe=0.68
Mms=3.5g //Cms=2.349e-3m/N Rms=0.149Ns/m Bl=1.573N/A
fs=100Hz Qms=4.6 Qes=1.6 //fs=98.6Hz Qes=1.6
Def_Driver "Tweeter"
Re=4.8ohm Le=0.05mH
Mms=0.32g //Cms=2.349e-3m/N Rms=0.149Ns/m Bl=1.573N/A
fs=500Hz Qms=4.6 Qes=1.6 //fs=98.6Hz Qes=1.6
System "S1"
Driver "Drv1" Def="Woofer"
Node=1001=0=2001=0 DrvGroup=1001
Filter 'F1'
f0=1.8kHz
Order=4 LR Lowpass
Filter 'F3'
{b0=0.55;}
RadImp "RadImp Drv1 Front"
Node=2001 DrvGroup=1001
System "S2"
Driver "Drv2" Def="Tweeter"
Node=1002=0=2002=0 DrvGroup=1002
Filter 'F2'
f0=1.8kHz
Order=4 LR Highpass
z0=17mm // delay between drivers
RadImp "RadImp Drv1 Front"
Node=2002 DrvGroup=1002
Hi mabat
if you plan to implement it in the next beta (would be great feature!) just a comment: I did many direct sound sims for line array. Setting low pass filter individually for each tweeter (lower fc the further away from middle; 1st order butterworth) can give very nice cd behaviour. If array is large additionally delaying helps a lot (in groups is sufficient…see keele cbt papers). So what I want to say: please make filter AND delay possible for every driver
Curving (complex to build) is not necessary at all if you have some active amp channels to spent!
YMMV, but I think a big part of the reason that the Danley Synergy Horns do that weird thing where you can't even tell where the speaker is located, is because the drivers are located at various depths but the output of the horn is in-phase.
In other words, there's three ways to produce a wavefront that's in phase:
1) The most obvious option is to have a single point source that covers as much bandwidth as possible.
2) Another option is to use a line source and delay the output at the center so that the wavefront is curved (what you are suggesting)
3) The least obvious option is what Danley is doing, which is that the highs, mids and lows radiate from different locations, but the output of the horn is in-phase. Because our ears can detect the location of where speakers are, the net effect is that the sound is in-phase but it sounds like the depth of the soundstage is indeterminate.
Synergy Horns are the only speakers I've ever heard that pull this trick off, it's trippy. If you read Erin's review of the Danley speakers, he mentions it too.
I think that the output of a properly implemented CBT is in-phase as well. If it's constant directivity both horizontally and vertically, which it certainly has the potential for (within the bandwidth of the array), it can't be otherwise. The question is how low can one get with the array itself. Down to Schroeder would be great, perhaps 300 - 500 Hz is more realistic, I don't know - it will depend on transducers used (without compromising the top end too much).
https://www.at-horns.eu/release/ath-beta-221114.zipbeta version available ?
Two examples attached. You can export FRD data for the individial array groups and than work with that in any crossover simulator (see array1.txt demo). Or you can also work within ABEC manually on the script level (the most simple is to overwrite individual levels and/or delays in observation.txt).
- No LF source/enclosure support yet.
Last edited:
I don't know but with the latest beta it can be tried what happens when the middle array elements are omitted. I guess it just breaks apart.
Please show us, I have my doubts... I don't think it will be as pretty as in the brochure...
(at least not with life size drivers)