[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PATCH v2 08/20] asc: generate silence if FIFO empty but engine still ru
From: |
Mark Cave-Ayland |
Subject: |
[PATCH v2 08/20] asc: generate silence if FIFO empty but engine still running |
Date: |
Sat, 9 Sep 2023 10:48:15 +0100 |
MacOS (un)helpfully leaves the FIFO engine running even when all the samples
have
been written to the hardware, and expects the FIFO status flags and IRQ to be
updated continuously.
There is an additional problem in that not all audio backends guarantee an
all-zero output when there is no FIFO data available, in particular the Windows
dsound backend which re-uses its internal circular buffer causing the last
played
sound to loop indefinitely.
Whilst this is effectively a bug in the Windows dsound backend, work around it
for now using a simple heuristic: if the FIFO remains empty for half a cycle
(~23ms) then continuously fill the generated buffer with empty silence.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
---
hw/audio/asc.c | 19 +++++++++++++++++++
include/hw/audio/asc.h | 2 ++
2 files changed, 21 insertions(+)
diff --git a/hw/audio/asc.c b/hw/audio/asc.c
index 336ace0cd6..b01b285512 100644
--- a/hw/audio/asc.c
+++ b/hw/audio/asc.c
@@ -334,6 +334,21 @@ static void asc_out_cb(void *opaque, int free_b)
}
if (!generated) {
+ /* Workaround for audio underflow bug on Windows dsound backend */
+ int64_t now = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
+ int silent_samples = muldiv64(now - s->fifo_empty_ns,
+ NANOSECONDS_PER_SECOND, ASC_FREQ);
+
+ if (silent_samples > ASC_FIFO_CYCLE_TIME / 2) {
+ /*
+ * No new FIFO data within half a cycle time (~23ms) so fill the
+ * entire available buffer with silence. This prevents an issue
+ * with the Windows dsound backend whereby the sound appears to
+ * loop because the FIFO has run out of data, and the driver
+ * reuses the stale content in its circular audio buffer.
+ */
+ AUD_write(s->voice, s->silentbuf, samples << s->shift);
+ }
return;
}
@@ -611,6 +626,7 @@ static void asc_unrealize(DeviceState *dev)
ASCState *s = ASC(dev);
g_free(s->mixbuf);
+ g_free(s->silentbuf);
AUD_remove_card(&s->card);
}
@@ -633,6 +649,9 @@ static void asc_realize(DeviceState *dev, Error **errp)
s->samples = AUD_get_buffer_size_out(s->voice) >> s->shift;
s->mixbuf = g_malloc0(s->samples << s->shift);
+ s->silentbuf = g_malloc0(s->samples << s->shift);
+ memset(s->silentbuf, 0x80, s->samples << s->shift);
+
/* Add easc registers if required */
if (s->type == ASC_TYPE_EASC) {
memory_region_add_subregion(&s->asc, ASC_EXTREG_OFFSET,
diff --git a/include/hw/audio/asc.h b/include/hw/audio/asc.h
index d9412815c3..4741f92c46 100644
--- a/include/hw/audio/asc.h
+++ b/include/hw/audio/asc.h
@@ -68,6 +68,8 @@ struct ASCState {
int samples;
int shift;
+ uint8_t *silentbuf;
+
/* Time when we were last able to generate samples */
int64_t fifo_empty_ns;
--
2.39.2
- Re: [PATCH v2 01/20] q800-glue.c: convert to Resettable interface, (continued)
- [PATCH v2 02/20] q800: add djMEMC memory controller, Mark Cave-Ayland, 2023/09/09
- [PATCH v2 03/20] q800: add machine id register, Mark Cave-Ayland, 2023/09/09
- [PATCH v2 04/20] q800: implement additional machine id bits on VIA1 port A, Mark Cave-Ayland, 2023/09/09
- [PATCH v2 05/20] q800: add IOSB subsystem, Mark Cave-Ayland, 2023/09/09
- [PATCH v2 06/20] q800: allow accesses to RAM area even if less memory is available, Mark Cave-Ayland, 2023/09/09
- [PATCH v2 07/20] audio: add Apple Sound Chip (ASC) emulation, Mark Cave-Ayland, 2023/09/09
- [PATCH v2 08/20] asc: generate silence if FIFO empty but engine still running,
Mark Cave-Ayland <=
- Re: [PATCH v2 08/20] asc: generate silence if FIFO empty but engine still running, Laurent Vivier, 2023/09/25
[PATCH v2 09/20] q800: add Apple Sound Chip (ASC) audio to machine, Mark Cave-Ayland, 2023/09/09
[PATCH v2 11/20] swim: add trace events for IWM and ISM registers, Mark Cave-Ayland, 2023/09/09
[PATCH v2 10/20] q800: add easc bool machine class property to switch between ASC and EASC, Mark Cave-Ayland, 2023/09/09