|
|||||||
 |
|
|
Thread Tools |
Mar 3, 2009, 06:26 PM
|
#1 |
|
Tail Razer
Join Date: Jun 2005
Location: Bernyurass, AZ - USA
Posts: 4,027
Rep Power: 0  |
Obscure DANE bug
I have stumbled upon this a few times, consistently.
As the title says its obscure, its mostly an weird annoyance. Anytime I use this plugin: (credited, was posted in these forums somewhere) Code:
name "virtual5.1";
copyright "Copyright (c) 2004. Foolou";
created "10/22/2004";
engine "kX";
guid "b2d90700-13e6-4aba-ae11-fb3a34eb986a";
itramsize 172;
; This plugin does a reduced HRTF-convolution to achieve virtual 5point-surround-sound
; The hrtf is derived from the kemar data : http://sound.media.mit.edu/KEMAR.html
; input for the convolution is 4 mono-streams for the 4 positioned speakers:
; infl = frontleft, infr = frontright, inbl = backleft, inbr = backright
; inm = center
input infl, infr, inbl;
input inbr;
input inm;
input lfe;
; prestored values for the mixers
control front-left=0.83;
control front-right=0.83;
control back-left=0.9;
control back-right=0.9;
control center=0.7;
control lfe-volume=0.7;
; output is a spatialized stereo-mix of the 5 audiosources for
; headphone
output outhl, outhr;
; this is the FIR-filters used to achieve the effect
; backward-filter:
static l10= 0.03326667, l11= 0.28890000, l12=-0.06650000;
static l13=-0.30056667, l14= 0.22286667, l15= 0.07123333;
static l16=-0.17296667, l17= 0.18620000, l18= 0.41426667;
static l19= 0.08590000, l20= 0.07236667, l21= 0.33280000;
static l22= 0.03366667, l23=-0.10180000, l24= 0.00623333;
static l25=-0.04856667, l26=-0.12326667, l27=-0.05936667;
static r22= 0.08553333, r23= 0.09703333, r24=-0.09350000;
static r25=-0.04806667, r26= 0.05866667, r27=-0.00556667;
static r28=-0.00956667, r29= 0.15076667, r30= 0.23810000;
static r31= 0.07446667, r32= 0.07696667;
; frontleft-filter:
static fl09= 0.18696667, fl10= 0.41006667, fl11= -0.38650000;
static fl12=-0.14216667, fl13= 0.04960000, fl14= -0.14953333;
static fl15= 0.48100000, fl16= 0.03460000, fl17= 0.22436667;
static fl18= 0.76460000, fl19= 0.06076667, fl20= -0.40433333;
static fl21= 0.09683333, fl22= 0.07473333, fl23= -0.40000000;
static fl24=-0.12726667, fl25=-0.07180000;
static fr20= 0.11010000, fr21= 0.04873333, fr22=-0.10876667;
static fr23= 0.07416667, fr24=-0.03633333, fr25=-0.04693333;
static fr26= 0.16090000, fr27= 0.04556667, fr28= 0.08066667;
static fr29= 0.23163333, fr30= 0.13646667, fr31=-0.07303333;
static fr32= 0.08133333, fr33=-0.10006667;
; temporal variables needed to do the math
temp outl, outr , x;
; using 5 sections of the itram for 5 audio-delays
idelay write putincuebl at 0x0;
idelay write putincuebr at 0x22;
idelay write putincuefl at 0x44;
idelay write putincuefr at 0x66;
idelay write putincuefm at 0x88;
; needs a lot of registers to access the needed parts of the
; itram for the hrtf-convolution
idelay read w32 at 0x20;
idelay read w31 at 0x1F;
idelay read w30 at 0x1E;
idelay read w29 at 0x1D;
idelay read w28 at 0x1C;
idelay read w27 at 0x1B;
idelay read w26 at 0x1A;
idelay read w25 at 0x19;
idelay read w24 at 0x18;
idelay read w23 at 0x17;
idelay read w22 at 0x16;
idelay read w21 at 0x15;
idelay read w20 at 0x14;
idelay read w19 at 0x13;
idelay read w18 at 0x12;
idelay read w17 at 0x11;
idelay read w16 at 0x10;
idelay read w15 at 0x0F;
idelay read w14 at 0x0E;
idelay read w13 at 0x0D;
idelay read w12 at 0x0C;
idelay read w11 at 0x0B;
idelay read w10 at 0x0A;
idelay read b32 at 0x42;
idelay read b31 at 0x41;
idelay read b30 at 0x40;
idelay read b29 at 0x3F;
idelay read b28 at 0x3E;
idelay read b27 at 0x3D;
idelay read b26 at 0x3C;
idelay read b25 at 0x3B;
idelay read b24 at 0x3A;
idelay read b23 at 0x39;
idelay read b22 at 0x38;
idelay read b21 at 0x37;
idelay read b20 at 0x36;
idelay read b19 at 0x35;
idelay read b18 at 0x34;
idelay read b17 at 0x33;
idelay read b16 at 0x32;
idelay read b15 at 0x31;
idelay read b14 at 0x30;
idelay read b13 at 0x2F;
idelay read b12 at 0x2E;
idelay read b11 at 0x2D;
idelay read b10 at 0x2C;
idelay read p33 at 0x65;
idelay read p32 at 0x64;
idelay read p31 at 0x63;
idelay read p30 at 0x62;
idelay read p29 at 0x61;
idelay read p28 at 0x60;
idelay read p27 at 0x5F;
idelay read p26 at 0x5E;
idelay read p25 at 0x5D;
idelay read p24 at 0x5C;
idelay read p23 at 0x5B;
idelay read p22 at 0x5A;
idelay read p21 at 0x59;
idelay read p20 at 0x58;
idelay read p19 at 0x57;
idelay read p18 at 0x56;
idelay read p17 at 0x55;
idelay read p16 at 0x54;
idelay read p15 at 0x53;
idelay read p14 at 0x52;
idelay read p13 at 0x51;
idelay read p12 at 0x50;
idelay read p11 at 0x4F;
idelay read p10 at 0x4E;
idelay read p09 at 0x4D;
idelay read q33 at 0x87;
idelay read q32 at 0x86;
idelay read q31 at 0x85;
idelay read q30 at 0x84;
idelay read q29 at 0x83;
idelay read q28 at 0x82;
idelay read q27 at 0x81;
idelay read q26 at 0x80;
idelay read q25 at 0x7F;
idelay read q24 at 0x7E;
idelay read q23 at 0x7D;
idelay read q22 at 0x7C;
idelay read q21 at 0x7B;
idelay read q20 at 0x7A;
idelay read q19 at 0x79;
idelay read q18 at 0x78;
idelay read q17 at 0x77;
idelay read q16 at 0x76;
idelay read q15 at 0x75;
idelay read q14 at 0x74;
idelay read q13 at 0x73;
idelay read q12 at 0x72;
idelay read q11 at 0x71;
idelay read q10 at 0x70;
idelay read q09 at 0x6F;
idelay read outm at 0x98; center is just a little delayed
; Code
; putting the 5 inputstreams on the itram
macs putincuebl, inbl, 0x0, 0x0;
macs putincuebr, inbr, 0x0, 0x0;
macs putincuefl, infl, 0x0, 0x0;
macs putincuefr, infr, 0x0, 0x0;
macs putincuefm, inm, 0x0, 0x0;
; and now doing the FIR math with the accum
; backleft-stereo-l-spatialized
macs infl, 0x0, 0x0, 0x0;
macmv x, x, w27, l27;
macmv x, x, w26, l26;
macmv x, x, w25, l25;
macmv x, x, w24, l24;
macmv x, x, w23, l23;
macmv x, x, w22, l22;
macmv x, x, w21, l21;
macmv x, x, w20, l20;
macmv x, x, w19, l19;
macmv x, x, w18, l18;
macmv x, x, w17, l17;
macmv x, x, w16, l16;
macmv x, x, w15, l15;
macmv x, x, w14, l14;
macmv x, x, w13, l13;
macmv x, x, w12, l12;
macmv x, x, w11, l11;
macmv x, x, w10, l10;
macs x, accum, 0x0, 0x0;
macs outl, 0x0, x, back-left; mixed into the outputleftchannel
;back-right stereo-r-spatialized
macs infl, 0x0, 0x0, 0x0;
macmv x, x, b27, l27;
macmv x, x, b26, l26;
macmv x, x, b25, l25;
macmv x, x, b24, l24;
macmv x, x, b23, l23;
macmv x, x, b22, l22;
macmv x, x, b21, l21;
macmv x, x, b20, l20;
macmv x, x, b19, l19;
macmv x, x, b18, l18;
macmv x, x, b17, l17;
macmv x, x, b16, l16;
macmv x, x, b15, l15;
macmv x, x, b14, l14;
macmv x, x, b13, l13;
macmv x, x, b12, l12;
macmv x, x, b11, l11;
macmv x, x, b10, l10;
macs x, accum, 0x0, 0x0;
macs outr, 0x0, x, back-right; mixed into the outputrightchannel
;backleft stereo-r-spatialized
macs infl, 0x0, 0x0, 0x0;
macmv x, x, w32, r32;
macmv x, x, w31, r31;
macmv x, x, w30, r30;
macmv x, x, w29, r29;
macmv x, x, w28, r28;
macmv x, x, w27, r27;
macmv x, x, w26, r26;
macmv x, x, w25, r25;
macmv x, x, w24, r24;
macmv x, x, w23, r23;
macmv x, x, w22, r22;
macs x, accum, 0x0, 0x0;
macs outr, outr, x, back-left; mixed to outputright
;backright stereo-l-spatialized
macs infl, 0x0, 0x0, 0x0;
macmv x, x, b32, r32;
macmv x, x, b31, r31;
macmv x, x, b30, r30;
macmv x, x, b29, r29;
macmv x, x, b28, r28;
macmv x, x, b27, r27;
macmv x, x, b26, r26;
macmv x, x, b25, r25;
macmv x, x, b24, r24;
macmv x, x, b23, r23;
macmv x, x, b22, r22;
macs x, accum, 0x0, 0x0;
macs outl, outl, x, back-right;
; now the frontspeakers:
; frontleft, stereo-l-spatialized
macs infl, 0x0, 0x0, 0x0;
macmv x, x, p25, fl25;
macmv x, x, p24, fl24;
macmv x, x, p23, fl23;
macmv x, x, p22, fl22;
macmv x, x, p21, fl21;
macmv x, x, p20, fl20;
macmv x, x, p19, fl19;
macmv x, x, p18, fl18;
macmv x, x, p17, fl17;
macmv x, x, p16, fl16;
macmv x, x, p15, fl15;
macmv x, x, p14, fl14;
macmv x, x, p13, fl13;
macmv x, x, p12, fl12;
macmv x, x, p11, fl11;
macmv x, x, p10, fl10;
macmv x, x, p09, fl09;
macs x, accum, 0,0;
macs outl, outl, x, front-left;
; frontleft, stereo-r-spatialized:
macs infl, 0x0, 0x0, 0x0;
macmv x, x , p33, fr33;
macmv x, x , p32, fr32;
macmv x, x , p31, fr31;
macmv x, x , p30, fr30;
macmv x, x , p29, fr29;
macmv x, x , p28, fr28;
macmv x, x , p27, fr27;
macmv x, x , p26, fr26;
macmv x, x , p25, fr25;
macmv x, x , p24, fr24;
macmv x, x , p23, fr23;
macmv x, x , p22, fr22;
macmv x, x , p21, fr21;
macmv x, x , p20, fr20;
macs x, accum, 0,0;
macs outr, outr, x, front-left;
; frontright, stereo-l-spatialized:
macs infl, 0x0, 0x0, 0x0;
macmv x, x , q33, fr33;
macmv x, x , q32, fr32;
macmv x, x , q31, fr31;
macmv x, x , q30, fr30;
macmv x, x , q29, fr29;
macmv x, x , q28, fr28;
macmv x, x , q27, fr27;
macmv x, x , q26, fr26;
macmv x, x , q25, fr25;
macmv x, x , q24, fr24;
macmv x, x , q23, fr23;
macmv x, x , q22, fr22;
macmv x, x , q21, fr21;
macmv x, x , q20, fr20;
macs x, accum, 0,0;
macs outl, outl, x, front-right;
; frontright, stereo-r-spatialized:
macs infl, 0x0, 0x0, 0x0;
macmv x, x, q25, fl25;
macmv x, x, q24, fl24;
macmv x, x, q23, fl23;
macmv x, x, q22, fl22;
macmv x, x, q21, fl21;
macmv x, x, q20, fl20;
macmv x, x, q19, fl19;
macmv x, x, q18, fl18;
macmv x, x, q17, fl17;
macmv x, x, q16, fl16;
macmv x, x, q15, fl15;
macmv x, x, q14, fl14;
macmv x, x, q13, fl13;
macmv x, x, q12, fl12;
macmv x, x, q11, fl11;
macmv x, x, q10, fl10;
macmv x, x, q09, fl09;
macs x, accum, 0,0;
macs outr, outr, x, front-right;
;after calculating the 8 different audio-streams they are mixed
;together for the stereo-output
macs outl, outl, lfe, lfe-volume;
macs outr, outr, lfe, lfe-volume;
macs outhl, outl, outm, center;
macs outhr, outr, outm, center;
end;
As you can see - there is only 1 pin feeding audio to the V5.1 plugin, and the PEAK meter (as *any* peak meter will show) - theres 'no audio' on that one pin. And as seen - there is still audio exiting the plugin; Thus - the audio is actually there, but the peak meter shows nothing. I am confident it is something with the V5.1 plugin and not the peak meters. Can any one explain why this happens? Is there an easy fix? Is it a deep bug with in DANE? I realize its a rather complex DANE only plugin (more than I can understand, tho, I tried looking for common constants used - theres a ton of them so...), and for the most part, it functions. Still...?? |
|
|
|
Mar 3, 2009, 08:35 PM
|
#2 |
|
HardwareHeaven Extreme Member
Join Date: Jan 2005
Posts: 5,561
Rep Power: 62  |
Re: Obscure DANE bug
It is not really a Dane bug, the problem is that the plugin is writing to it's 'infl' input (for some unknown reason?).
i.e. You will see this in a few places: macs infl, 0x0, 0x0, 0x0; The problem is that inputs do not really exist, they are virtual. You have to think of the GPR's as addresses/pointers. When nothing is connected to an input, it would basically just be a value of zero, but when it is connected to an output, it is basically the output GPR's address. So what is happening here is basically that MX6's 'MainL' output is getting zero'd out by the virtual5.1 plugin, so you do not see anything on the peak meter (if the peak meter was loaded before the virtual5.1 plugin, then you would probably see the signal since the peak plugin would see it before it is zero'd out). Last edited by Russ; Mar 10, 2009 at 05:57 AM. Reason: typo |
|
|
|
|
Mar 3, 2009, 08:57 PM
|
#3 |
|
h/h member-shmember
Join Date: Dec 2002
Location: Evil Empire
Posts: 2,639
Rep Power: 69 |
Ответ: Obscure DANE bug
yep, not a dane thing.
(well, actually - this bug in the code is indirectly comes out of one of major dane bugs - "input" registers are not marked as read-only (e.g. assembler should not compile such code at all)) as Russ said you can fix the "virtual5.1" by replacing all macs infl, 0x0, 0x0, 0x0; by macs 0, 0, 0, 0
__________________
|
|
|
|
|
Mar 3, 2009, 09:03 PM
|
Thread Starter
#4
|
|
Tail Razer
Join Date: Jun 2005
Location: Bernyurass, AZ - USA
Posts: 4,027
Rep Power: 0 |
Re: Obscure DANE bug
Ah, ok - I should have caught that, as I knew that is a 'no no'...
So, is that how these plugins are 'connected' by these 'input' and 'output' getting getting assigned a common register? Plugin #1 Output register address = plugin 2 Input register address. Is that how these 'virtual connections' are made internally (literally) - or is all that just 'figuratively'? |
|
|
|
|
Mar 3, 2009, 09:06 PM
|
Thread Starter
#5
|
||
|
Tail Razer
Join Date: Jun 2005
Location: Bernyurass, AZ - USA
Posts: 4,027
Rep Power: 0 |
Re: Ответ: Obscure DANE bug
Quote:
Quote:
Clearing a flag or something maybe..?? |
||
|
|
|
|
Mar 3, 2009, 09:10 PM
|
#6 |
|
HardwareHeaven Extreme Member
Join Date: Jan 2005
Posts: 5,561
Rep Power: 62 |
Re: Obscure DANE bug
It is supposed to clear the accumulator for the macmv instructions, but that would only seem to work if nothing is connected to the input (if at all).
But to answer your question, yes the instructions are needed (and should be as Max wrote). [EDIT] Oops, I did not see your other post... Yes, that is how the plugins are connected together. The raw instructions just contain an opcode and addresses (addresses/offsets of constants, or tram, or physical input/output registers, or general purpose registers (GPR), etc). Last edited by Russ; Mar 3, 2009 at 09:43 PM. |
|
|
|
|
| Thread Tools | |
|
|
