From 716c84be13ba425edf49a533ef5ffdfa5f6f4aaf Mon Sep 17 00:00:00 2001
From: Tim Blechmann <timblech@users.sourceforge.net>
Date: Fri, 24 Jun 2005 19:39:18 +0000
Subject: importing source

svn path=/trunk/externals/tb/; revision=3248
---
 block_delay~/block_delay~.c | 99 +++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 99 insertions(+)
 create mode 100644 block_delay~/block_delay~.c

(limited to 'block_delay~/block_delay~.c')

diff --git a/block_delay~/block_delay~.c b/block_delay~/block_delay~.c
new file mode 100644
index 0000000..f5cfa4c
--- /dev/null
+++ b/block_delay~/block_delay~.c
@@ -0,0 +1,99 @@
+/* this external is based on nop~ which was part of zexy-1.x, see the copyright notice below: 
+
+ * ZEXY is published under the GNU GeneralPublicLicense, that must be shipped with ZEXY.
+ * if you are using Debian GNU/linux, the GNU-GPL can be found under /usr/share/common-licenses/GPL
+ * if you still haven't found a copy of the GNU-GPL, have a look at http://www.gnu.org
+ *
+ * "pure data" has it's own license, that comes shipped with "pure data".
+ *
+ * there are ABSOLUTELY NO WARRANTIES for anything
+
+ it does absolutely the same as z~ 64, but since it uses simd instructions, it's faster.
+*/
+
+
+
+/* ------------------------ block_delay~ ----------------------------- */
+/* this will pass trough the signal unchanged except for a delay of 1 block */
+
+#include "m_pd.h"
+
+static t_class *block_delay_tilde_class;
+
+typedef struct _block_delay
+{
+	t_object x_obj;
+	t_float *buf;
+	int n;
+	int toggle;
+} t_block_delay;
+
+
+static t_int *block_delay_tilde_perfsimd(t_int *w)
+{
+	t_float *in  = (t_float *)w[1];
+	t_float *out = (t_float *)w[2];
+	t_block_delay *x = (t_block_delay *)w[3];
+	int n = x->n;
+	t_float *rp = x->buf + n * x->toggle, *wp = x->buf + n * (x->toggle ^= 1);
+
+	copyvec_simd(wp, in, n);
+	copyvec_simd(out, rp, n);
+
+	return (w+4);
+}
+
+static t_int *block_delay_tilde_perform(t_int *w)
+{
+	t_float *in  = (t_float *)w[1];
+	t_float *out = (t_float *)w[2];
+	t_block_delay *x = (t_block_delay *)w[3];
+	int n = x->n;
+	t_float *rp = x->buf + n * x->toggle, *wp = x->buf + n * (x->toggle ^= 1);
+
+	while (n--)	{
+		*wp++ = *in++;
+		*out++ = *rp++;
+	}
+
+	return (w+4);
+}
+
+static void block_delay_tilde_dsp(t_block_delay *x, t_signal **sp)
+{
+	if (x->n != sp[0]->s_n)
+	{
+		if (x->n)
+			freealignedbytes(x->buf, x->n * 2 * sizeof(t_float));
+		x->buf = (t_float *)getalignedbytes(sizeof(t_float) * 2 * (x->n = sp[0]->s_n));
+	}
+	if (simd_check2(sp[0]->s_n, sp[0]->s_vec, sp[1]->s_vec))
+		dsp_add(block_delay_tilde_perfsimd, 3, sp[0]->s_vec, sp[1]->s_vec, x);
+	else
+		dsp_add(block_delay_tilde_perform, 3, sp[0]->s_vec, sp[1]->s_vec, x);
+}
+
+static void block_delay_tilde_free(t_block_delay *x)
+{
+	if (x->buf)
+		freealignedbytes(x->buf, x->n * sizeof(t_float));
+}
+
+
+static void *block_delay_tilde_new(void)
+{
+	t_block_delay *x = (t_block_delay *)pd_new(block_delay_tilde_class);
+	outlet_new(&x->x_obj, gensym("signal"));
+	x->toggle = 0;
+	x->n = 0;
+
+	return (x);
+}
+
+void block_delay_tilde_setup(void)
+{
+	block_delay_tilde_class = class_new(gensym("block_delay~"), (t_newmethod)block_delay_tilde_new, (t_method)block_delay_tilde_free,
+		sizeof(t_block_delay), 0, A_DEFFLOAT, 0);
+	class_addmethod(block_delay_tilde_class, nullfn, gensym("signal"), 0);
+	class_addmethod(block_delay_tilde_class, (t_method)block_delay_tilde_dsp, gensym("dsp"), 0);
+}
-- 
cgit v1.2.1