Re: [LAU] Sequencer suggestions?

On Tue, Aug 26, 2014 at 05:20:24PM +0200, Philipp Überbacher wrote:

> Thanks Fons, much appreciated. It seems the original author didn't
> touch the code in more than 1 1/2 years, so the bug report would likely
> had no effect.

The following patches seem to fix the problem. They remove the
(inoperative) per patch DC blocker and add one per voice, just
after the oscillators. More testing may be required.

--- engine.h.orig 2013-01-13 08:18:49.000000000 +0100
+++ engine.h 2014-08-26 17:18:13.514909324 +0200
@@ -84,9 +84,7 @@
    sample_t bps; /* beats per second */
    sample_t out1; /* output sample 2 */
    sample_t out2; /* output sample 1 */
-#ifdef ENABLE_DC_REJECTION_FILTER
- sample_t dcR_const;
-#endif
+ sample_t wdcf;
 } GLOBAL;
 
 
@@ -105,6 +103,8 @@
    int portamento_sample; /* sample number within portamento */
    int portamento_samples; /* portamento time in samples */
    int age; /* voice age, in samples */
+ sample_t dcf1; /* DC filter state 1 */
+ sample_t dcf2; /* DC filter state 2 */
    sample_t out1; /* output sample 1 */
    sample_t out2; /* output sample 2 */
    sample_t amp_env; /* smoothed final output of env generator */

--- engine.c.orig 2013-01-13 08:18:49.000000000 +0100
+++ engine.c 2014-08-26 17:54:29.301481175 +0200
@@ -193,9 +193,7 @@
 
    /* (-3dB @ 20Hz) DC blocking filter */
    /* 1.0 - (M_PI * 2 * freq / (sample_t) f_sample_rate) */
-#ifdef ENABLE_DC_REJECTION_FILTER
- global.dcR_const = 1.0 - (125.6 / (sample_t) f_sample_rate);
-#endif
+ global.wdcf = 125.7 / (sample_t) f_sample_rate;
 }
 
 
@@ -487,6 +485,10 @@
            voice->midi_key = -1;
            voice->keypressed = -1;
 
+ /* init dc filters */
+ voice->dcf1 = 0.0;
+ voice->dcf2 = 0.0;
+
            /* initialize moog filters */
            voice->filter_y1_1 = 0.0;
            voice->filter_y1_2 = 0.0;
@@ -780,10 +782,6 @@
 run_part(PART *part, PATCH_STATE *state, unsigned int part_num)
 {
    unsigned int osc;
-#ifdef ENABLE_DC_REJECTION_FILTER
- sample_t tmp1;
- sample_t tmp2;
-#endif
 
    /* generate amplitude envelopes for all voices */
    run_voice_envelopes(part, state, part_num);
@@ -842,19 +840,6 @@
    if (state->delay_mix_cc) {
        run_delay(get_delay(part_num), part, state);
    }
-
- /* output this sample to the buffer */
-#ifdef ENABLE_DC_REJECTION_FILTER
- tmp1 = part->out1;
- part->out1 = part->out1 - part->dcR_in1 + global.dcR_const * part->dcR_out1;
- part->dcR_in1 = tmp1;
- part->dcR_out1 = part->out1;
-
- tmp2 = part->out2;
- part->out2 = part->out2 - part->dcR_in2 + global.dcR_const * part->dcR_out2;
- part->dcR_in2 = tmp2;
- part->dcR_out2 = part->out2;
-#endif
 }
 
 
@@ -1300,6 +1285,12 @@
        run_osc(voice, part, state, osc);
    }
 
+ /* apply DC filters */
+ voice->dcf1 += global.wdcf * (voice->out1 - voice->dcf1);
+ voice->out1 -= voice->dcf1;
+ voice->dcf2 += global.wdcf * (voice->out2 - voice->dcf2);
+ voice->out2 -= voice->dcf2;
+
    /* oscs are mixed. now apply AM oscs. */
    for (osc = 0; osc < NUM_OSCS; osc++) {
        if (state->osc_modulation[osc] == MOD_TYPE_AM) {

-- 
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-user