core: APU channel disable - fix bitmasks and LQ-mode volume application

Two related bugs in the user-facing fceumm_apu_1..5 channel disable
toggles. Pre-existing since commit 448a231 (set_apu_channels) and at
least as far back as the LQ rewrite for the second one.

set_apu_channels (libretro.c)
-----------------------------

The bitmask built by check_variables uses bit i for fceumm_apu_(i+1)
enabled, ordered SQ1, SQ2, Triangle, Noise, DMC. The old code:

    FSettings.SquareVolume[1] = (chan & 1) ? 256 : 0;
    FSettings.SquareVolume[0] = (chan & 2) ? 256 : 0;
    FSettings.TriangleVolume  = (chan & 3) ? 256 : 0;
    FSettings.NoiseVolume     = (chan & 4) ? 256 : 0;
    FSettings.PCMVolume       = (chan & 5) ? 256 : 0;

crossed SQ1/SQ2 and used non-power-of-two masks for the remaining
three. Practical effect: apu_1 muted SQ2 and apu_2 muted SQ1; apu_3
through apu_5 left their channels audible in any normal config (the
masks happened to be true any time another bit was set).

Fixed to read each channel's own bit:

    FSettings.SquareVolume[0] = (chan & 0x01) ? 256 : 0;  /* SQ1  */
    FSettings.SquareVolume[1] = (chan & 0x02) ? 256 : 0;  /* SQ2  */
    FSettings.TriangleVolume  = (chan & 0x04) ? 256 : 0;  /* Tri  */
    FSettings.NoiseVolume     = (chan & 0x08) ? 256 : 0;  /* Nse  */
    FSettings.PCMVolume       = (chan & 0x10) ? 256 : 0;  /* DMC  */

RDoTriangleNoisePCMLQ (sound.c)
-------------------------------

The combined Triangle/Noise/DMC LQ mix uses a non-linear lookup
table:

    totalout = wlookup2[lq_tcout + noiseout + RawDALatch];

The old code took the noise envelope (EnvUnits[2], not Triangle -
Triangle has no envelope; EnvUnits[0]=SQ1, [1]=SQ2, [2]=Noise) and
scaled it by FSettings.TriangleVolume, despite the comment claiming
to modify Triangle. Meanwhile lq_tcout (Triangle) and RawDALatch
(DMC) went into wlookup2 unscaled, so neither TriangleVolume nor
PCMVolume had any effect in LQ mode.

Fixed:
* amptab[0] (noise envelope) is now scaled by FSettings.NoiseVolume.
* lq_tcout is scaled by FSettings.TriangleVolume into a local
  scaled_tcout fed to wlookup2.
* RawDALatch is scaled by FSettings.PCMVolume into scaled_dmc.

The volume scalars are loop-invariant within this function, so the
Triangle volume is cached once in tri_vol at function entry to keep
the inner loops tight; the != 256 fast paths preserve byte-exact
behavior at full volume.

Verification
------------

Built per-channel-only test ROMs (test_sq1_only, test_sq2_only,
test_tri_only, test_nse_only, test_dmc_only) and an apu_test
runner that measures output RMS with each fceumm_apu_N toggle.
After the fix, in both LQ and HQ modes, each apu_N toggle mutes
exactly the channel whose ROM is the only sound source, and no
other toggle has any effect on it.

Default (all-on) audio output is bit-exact identical to the
baseline across the existing test_silent / test_idle / test_active /
test_stress ROMs (1500 frames each), so games and homebrew that
never touch the apu_N options see no change. No performance
regression in the default path; the != 256 guards skip the multiplies.

src/drivers/libretro/libretro.c

@@ -1975,11 +1975,23 @@ void retro_reset(void)
 
 static void set_apu_channels(int chan)
 {
-   FSettings.SquareVolume[1] = (chan & 1) ? 256 : 0;
+   /* Bitmask layout (set up by check_variables): bit i = fceumm_apu_(i+1)
-   FSettings.SquareVolume[0] = (chan & 2) ? 256 : 0;
+    * is enabled. Channel order matches the libretro core options:
-   FSettings.TriangleVolume  = (chan & 3) ? 256 : 0;
+    *   bit 0  = Square 1   (apu_1)
-   FSettings.NoiseVolume     = (chan & 4) ? 256 : 0;
+    *   bit 1  = Square 2   (apu_2)
-   FSettings.PCMVolume       = (chan & 5) ? 256 : 0;
+    *   bit 2  = Triangle   (apu_3)
+    *   bit 3  = Noise      (apu_4)
+    *   bit 4  = PCM / DMC  (apu_5)
+    *
+    * The previous masks crossed the SQ1/SQ2 indices and used non-power-of-
+    * two values for the remaining channels, which silently broke the
+    * apu_3..apu_5 toggles entirely (any combination of bits 0/1/2 left
+    * those channels audible) and made apu_1 mute SQ2 and vice versa. */
+   FSettings.SquareVolume[0] = (chan & 0x01) ? 256 : 0;
+   FSettings.SquareVolume[1] = (chan & 0x02) ? 256 : 0;
+   FSettings.TriangleVolume  = (chan & 0x04) ? 256 : 0;
+   FSettings.NoiseVolume     = (chan & 0x08) ? 256 : 0;
+   FSettings.PCMVolume       = (chan & 0x10) ? 256 : 0;
 }
 
 static void check_variables(bool startup)

src/sound.c

@@ -733,6 +733,9 @@ static void RDoTriangleNoisePCMLQ(void) {
 	uint32_t amptab[2];
 	uint32_t noiseout;
 	int nshift;
+	uint32_t scaled_tcout;
+	uint32_t scaled_dmc;
+	const uint32_t tri_vol = FSettings.TriangleVolume;
 
 	int32_t totalout;
 
@@ -754,13 +757,18 @@ static void RDoTriangleNoisePCMLQ(void) {
 	else
 		amptab[0] = EnvUnits[2].decvolume;
 
-	/* Modify Triangle wave volume based on channel volume modifiers
+	/* Apply per-channel volume modifiers (set via fceumm_apu_N options).
-	 * Note: the formulat x = x * y /100 does not yield exact results,
+	 *
-	 * but is "close enough" and avoids the need for using double vales
+	 * EnvUnits[2] is the Noise envelope (not Triangle - Triangle has no
-	 * or implicit cohersion which are slower (we need speed here)
+	 * envelope; EnvUnits[0]=SQ1, [1]=SQ2, [2]=Noise). The previous code
-	 * TODO: Optimize this. */
+	 * scaled amptab[0] by FSettings.TriangleVolume, which crossed the
-	if (FSettings.TriangleVolume != 256)
+	 * Triangle and Noise mute toggles in LQ mode and left Triangle
-		amptab[0] = (amptab[0] * FSettings.TriangleVolume) / 256;
+	 * itself never muted. Triangle's contribution enters wlookup2 via
+	 * lq_tcout below; PCM enters via RawDALatch. Scale each input
+	 * channel by its own volume before the non-linear DAC mix - 0 in
+	 * still produces 0 out, and 256 leaves the value unchanged. */
+	if (FSettings.NoiseVolume != 256)
+		amptab[0] = (amptab[0] * FSettings.NoiseVolume) / 256;
 
 	amptab[1] = 0;
 	amptab[0] <<= 1;
@@ -775,7 +783,14 @@ static void RDoTriangleNoisePCMLQ(void) {
 	else
 		nshift = 13;
 
-	totalout = wlookup2[lq_tcout + noiseout + RawDALatch];
+	scaled_tcout = (tri_vol != 256)
+	             ? ((lq_tcout * tri_vol) / 256)
+	             : lq_tcout;
+	scaled_dmc   = (FSettings.PCMVolume != 256)
+	             ? ((RawDALatch * FSettings.PCMVolume) / 256)
+	             : RawDALatch;
+
+	totalout = wlookup2[scaled_tcout + noiseout + scaled_dmc];
 
 	if (inie[0] && inie[1]) {
 		for (V = start; V < end; V++) {
@@ -792,7 +807,10 @@ static void RDoTriangleNoisePCMLQ(void) {
 				lq_tcout = (tristep & 0xF);
 				if (!(tristep & 0x10)) lq_tcout ^= 0xF;
 				lq_tcout = lq_tcout * 3;
-				totalout = wlookup2[lq_tcout + noiseout + RawDALatch];
+				scaled_tcout = (tri_vol != 256)
+				             ? ((lq_tcout * tri_vol) / 256)
+				             : lq_tcout;
+				totalout = wlookup2[scaled_tcout + noiseout + scaled_dmc];
 			}
 
 			if (lq_noiseacc <= 0) {
@@ -808,7 +826,7 @@ static void RDoTriangleNoisePCMLQ(void) {
 				nreg &= 0x7fff;
 				noiseout = amptab[(nreg >> 0xe) & 1];
 				if (lq_noiseacc <= 0) goto rea2;
-				totalout = wlookup2[lq_tcout + noiseout + RawDALatch];
+				totalout = wlookup2[scaled_tcout + noiseout + scaled_dmc];
 			}	/* noiseacc<=0 */
 		}	/* for(V=... */
 	} else if (inie[0]) {
@@ -825,7 +843,10 @@ static void RDoTriangleNoisePCMLQ(void) {
 				lq_tcout = (tristep & 0xF);
 				if (!(tristep & 0x10)) lq_tcout ^= 0xF;
 				lq_tcout = lq_tcout * 3;
-				totalout = wlookup2[lq_tcout + noiseout + RawDALatch];
+				scaled_tcout = (tri_vol != 256)
+				             ? ((lq_tcout * tri_vol) / 256)
+				             : lq_tcout;
+				totalout = wlookup2[scaled_tcout + noiseout + scaled_dmc];
 			}
 		}
 	} else if (inie[1]) {
@@ -845,7 +866,7 @@ static void RDoTriangleNoisePCMLQ(void) {
 				nreg &= 0x7fff;
 				noiseout = amptab[(nreg >> 0xe) & 1];
 				if (lq_noiseacc <= 0) goto area2;
-				totalout = wlookup2[lq_tcout + noiseout + RawDALatch];
+				totalout = wlookup2[scaled_tcout + noiseout + scaled_dmc];
 			}	/* noiseacc<=0 */
 		}
 	} else {