Re: [PATCH] sbc: fix for overflow bug in quantization code

[Siarhei Siamashka <siarhei.siamashka@nokia.com>]

[-- Attachment #1: Type: text/plain, Size: 993 bytes --]

On Monday 22 December 2008 13:37:53 ext Siarhei Siamashka wrote:
[...]
> If we try to use more realistic settings similar to the recommended high
> quality settings from SBC specification (Table 4.7):
> ./sbcenc -j -S -b 51 BigBuckBunny-stereo.au >BigBuckBunny-stereo.sbc
>
> bluez 16-bit fixed point:
> stddev:   43.82 PSNR: 63.48 bytes:114491016/114491308
>
> bluez 32-bit fixed point:
> stddev:   43.78 PSNR: 63.49 bytes:114491016/114491308
>
> reference encoder:
> stddev:   43.37 PSNR: 63.57 bytes:114491016/114491308
[...]
> PS. I still wonder why there is a loss when compared to reference encoder.
> 32-bit fixed point version should be even more precise than single
> precision floating point. Maybe there could be another minor bug in the
> code, or it is just a random deviation and there could be a win for other
> audio files.

Found what's the matter. It's a problem in subbands selection criteria for
joint-stereo. The following patch fixes it.

Best regards,
Siarhei Siamashka

[-- Attachment #2: 0001-Fixed-subbands-selection-for-joint-stereo-in-SBC-enc.patch --]
[-- Type: text/x-diff, Size: 1312 bytes --]

>From c1e98a7130377c126b6cdf93812016eab9151729 Mon Sep 17 00:00:00 2001
From: Siarhei Siamashka <siarhei.siamashka@nokia.com>
Date: Sat, 27 Dec 2008 19:36:14 +0200
Subject: [PATCH] Fixed subbands selection for joint-stereo in SBC encoder

Signed-off-by: Siarhei Siamashka <siarhei.siamashka@nokia.com>
---
 sbc/sbc.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/sbc/sbc.c b/sbc/sbc.c
index a3a3ac1..d3dcd9a 100644
--- a/sbc/sbc.c
+++ b/sbc/sbc.c
@@ -1051,15 +1051,15 @@ static int sbc_pack_frame(uint8_t *data, struct sbc_frame *frame, size_t len)
 			}
 
 			/* decide whether to join this subband */
-			if ((scalefactor[0][sb] + scalefactor[1][sb]) >
-					(scalefactor_j[0] + scalefactor_j[1]) ) {
+			if ((frame->scale_factor[0][sb] +
+					frame->scale_factor[1][sb]) >
+					(scale_factor_j[0] +
+					scale_factor_j[1])) {
 				/* use joint stereo for this subband */
 				joint |= 1 << (frame->subbands - 1 - sb);
 				frame->joint |= 1 << sb;
 				frame->scale_factor[0][sb] = scale_factor_j[0];
 				frame->scale_factor[1][sb] = scale_factor_j[1];
-				scalefactor[0][sb] = scalefactor_j[0];
-				scalefactor[1][sb] = scalefactor_j[1];
 				for (blk = 0; blk < frame->blocks; blk++) {
 					frame->sb_sample_f[blk][0][sb] =
 							sb_sample_j[blk][0];
-- 
1.5.6.5