Signaling Changer of AAC in MP4 without Demux

0readme.md

Signaling Changer of AAC in MP4 without Demux

guinurmi

Applying the patch (gpac-v0.6.1.diff), MP4Box becomes to be able to accept -sbr and -ps when importing MP4.
In other words, MP4Box will be able to change signaling of AAC in MP4 without demux which may cause out-of-sync.

> MP4Box -info AVC_HE-AAC.mp4 2>&1 | grep LC
MPEG-4 Audio AAC LC - 2 Channel(s) - SampleRate 24000

> MP4BoxSBR -add AVC_HE-AAC.mp4 -sbr -new new_AVC_HE-AAC.mp4
IsoMedia import AVC_HE-AAC.mp4 - track ID 1 - Video (size 320 x 240)
IsoMedia import AVC_HE-AAC.mp4 - track ID 2 - Audio (SR 24000 - 2 channels)
 - backward compatible signaling of AAC-SBR (SBR-SR 48000)
Saving new_AVC_HE-AAC.mp4: 0.500 secs Interleaving

> MP4Box -info new_AVC_HE-AAC.mp4 2>&1 | grep LC
MPEG-4 Audio AAC LC (HE-AAC v1) - 2 Channel(s) - SampleRate 24000 - SBR: SampleRate 48000 Type MPEG-4 Audio SBR

Let's use this patched MP4Box (MP4BoxSBR) for inflexible hardware players.
MP4BoxSBR.exe was built on MINGW32, so you can execute it without additional DLLs in Windows Command Prompt.

gpac-v0.6.1.diff

diff --git a/include/gpac/version.h b/include/gpac/version.h
index f93e487..630b220 100644
--- a/include/gpac/version.h
+++ b/include/gpac/version.h
@@ -41,6 +41,6 @@
 #define GPAC_VERSION_MICRO 0
 
 #include <gpac/revision.h>
-#define GPAC_FULL_VERSION       GPAC_VERSION "-rev" GPAC_GIT_REVISION
+#define GPAC_FULL_VERSION       GPAC_VERSION "-rev" GPAC_GIT_REVISION " (-sbr usable when importing MP4)"
 
 #endif //_GF_VERSION_H
diff --git a/src/media_tools/media_import.c b/src/media_tools/media_import.c
index 783df49..4786942 100644
--- a/src/media_tools/media_import.c
+++ b/src/media_tools/media_import.c
@@ -2023,6 +2023,7 @@ GF_Err gf_import_isomedia(GF_MediaImporter *import)
 	GF_Err e;
 	u64 offset, sampDTS, duration, dts_offset;
 	u32 track, di, trackID, track_in, i, num_samples, mtype, w, h, sr, sbr_sr, ch, mstype, cur_extract_mode;
+	u32 bot = 0, bsi = 0;
 	s32 trans_x, trans_y;
 	s16 layer;
 	u8 bps;
@@ -2114,6 +2115,8 @@ GF_Err gf_import_isomedia(GF_MediaImporter *import)
 			PL = dsi.audioPL;
 			sbr = dsi.has_sbr ? ((dsi.base_object_type==GF_M4A_AAC_SBR || dsi.base_object_type==GF_M4A_AAC_PS) ? 2 : 1) : GF_FALSE;
 			ps = dsi.has_ps;
+			bot = dsi.base_object_type;
+			bsi = dsi.base_sr_index;
 		}
 #endif
 		gf_isom_set_pl_indication(import->dest, GF_ISOM_PL_AUDIO, PL);
@@ -2330,6 +2333,53 @@ GF_Err gf_import_isomedia(GF_MediaImporter *import)
 		}
 	}
 
+	if (origin_esd && (origin_esd->decoderConfig->objectTypeIndication==GPAC_OTI_AUDIO_AAC_MPEG4)
+		&& ((import->flags & GF_IMPORT_SBR_IMPLICIT) || (import->flags & GF_IMPORT_PS_IMPLICIT))) {
+		u16 sbr_sr_idx = 0;
+		GF_BitStream *dsi;
+		dsi = gf_bs_new(NULL, 0, GF_BITSTREAM_WRITE);
+
+		for (i=0; i<16; i++) {
+			if (GF_M4ASampleRates[i] == (u32) 2*sr) {
+				sbr_sr_idx = i;
+				break;
+			}
+		}
+
+		/*regular AAC*/
+		gf_bs_write_int(dsi, bot, 5);
+		gf_bs_write_int(dsi, bsi, 4);
+		gf_bs_write_int(dsi, ch, 4);
+		gf_bs_align(dsi);
+		/*implicit AAC SBR signal*/
+		if (import->flags & GF_IMPORT_SBR_IMPLICIT) {
+			gf_import_message(import, GF_OK, " - backward compatible signaling of AAC-SBR (SBR-SR %d)", 2*sr);
+			gf_bs_write_int(dsi, 0x2b7, 11); /*sync extension type*/
+			gf_bs_write_int(dsi, 5, 5);	/*SBR objectType*/
+			gf_bs_write_int(dsi, 1, 1);	/*SBR present flag*/
+			gf_bs_write_int(dsi, sbr_sr_idx, 4);
+		}
+		if (import->flags & GF_IMPORT_PS_IMPLICIT) {
+			gf_import_message(import, GF_OK, " - backward compatible signaling of AAC-PS");
+			gf_bs_write_int(dsi, 0x548, 11); /*sync extension type*/
+			gf_bs_write_int(dsi, 1, 1);	/* PS present flag */
+		}
+		gf_bs_align(dsi);
+
+		{
+			GF_ESD *esd = gf_isom_get_esd(import->dest, track, 1);
+			if (esd) {
+				if (!esd->decoderConfig->decoderSpecificInfo) esd->decoderConfig->decoderSpecificInfo = (GF_DefaultDescriptor *) gf_odf_desc_new(GF_ODF_DSI_TAG);
+				if (esd->decoderConfig->decoderSpecificInfo->data) gf_free(esd->decoderConfig->decoderSpecificInfo->data);
+				gf_bs_get_content(dsi, &esd->decoderConfig->decoderSpecificInfo->data, &esd->decoderConfig->decoderSpecificInfo->dataLength);
+				gf_isom_change_mpeg4_description(import->dest, track, 1, esd);
+				esd->decoderConfig->decoderSpecificInfo->data = NULL;
+				gf_odf_desc_del((GF_Descriptor *)esd);
+			}
+		}
+		gf_bs_del(dsi);
+	}
+
 	gf_media_update_bitrate(import->dest, track);
 
 exit: