Skip to contentSkip to conversation

SBC codec, joint stereo mode

Bluez-devel Siegbert Baude 4 participants 4 messages Jan 4, 2007 Bluetooth drivers

lore
Share & tools
lore
#0 lore raw

Hi,

I have a question with regard to joint stereo mode in SBC. According to the A2DP spec. p. 53, sec. 12.4 "Syntax" in the frame_header there is one bit RFA after the join[sb] bits. In the example implementations of the SBC codec I got (one from TI, one from iAnywhere AV SDK, one from the ALSA-Bluetooth project), this bit is however never accounted for. Also in the sbc_test_14.sbc example you can download from bluetooth.org, I find at offset 04FB the beginning of the frame as 9C7F797BBC87, where the first bit of "87" should be this RFA bit (as this example uses 8 subbands). P. 11, sec. 1.3.2.1 "RFA" however says, RFA bits should be set to zero, which is not the case for "87".

So is this RFA bit actually used or not?

Ciao
Siegbert

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Bluez-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel

#1 lore raw

Hi,

page 53 says join[sb] is nrof_subbands-1 bits wide =3D> wich is 7 bits. =

The 8th bit is the RFA which is 0.
0xBC =3D 10111100, The first 7 bits are the join table, the last bit is =

the rfa.
Then 87 is already the scale factors.

Frederic

Siegbert Baude a =E9crit :

Quoted 13 lines

Hi,

I have a question with regard to joint stereo mode in SBC. According to the A2DP spec. p. 53, sec. 12.4 "Syntax" in the frame_header there is one bit RFA after the join[sb] bits. In the example implementations of the SBC codec I got (one from TI, one from iAnywhere AV SDK, one from the ALSA-Bluetooth project), this bit is however never accounted for. Also in the sbc_test_14.sbc example you can download from bluetooth.org, I find at offset 04FB the beginning of the frame as 9C7F797BBC87, where the first bit of "87" should be this RFA bit (as this example uses 8 subbands). P. 11, sec. 1.3.2.1 "RFA" however says, RFA bits should be set to zero, which is not the case for "87".

Quoted 8 lines

So is this RFA bit actually used or not?

Ciao
Siegbert

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share y=

our

Quoted 2 lines

opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3D=

DEVDEV

Quoted 5 lines

_______________________________________________
Bluez-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel
=

-- =

Frederic

Without the wind, the grass does not move.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3DDE=
VDEV
_______________________________________________
Bluez-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel

#2 lore raw

Hi Fr=E9d=E9ric,

Quoted 1 line · Fr=E9d=E9ric DALLEAU schrieb:

page 53 says join[sb] is nrof_subbands-1 bits wide =3D> wich is 7 bits.

Oh, my blindness. Thanks a *lot* for the hint, maybe I didn't read it =

carefully enough, because the decoding code sample I looked first at, =

was evaluating nrof_subbands instead of nrof_subbands-1.

Quoted 5 lines

The 8th bit is the RFA which is 0.
0xBC =3D 10111100, The first 7 bits are the join table, the last bit is =

the rfa.
Then 87 is already the scale factors.

Yep, now it makes sense.
BTW, is this the right list to ask SBC-specific questions? I have some =

more general questions, e.g. I don't quite understand the rationale =

behind the codec definition: If you choose joint mode for encoding, it =

happens that the encoder does not use joint mode for all subbands. But =

why? Saving bits? If anybody has a pointer to some documentation for =

SBC, which explains the design decisions I would be really happy.

Ciao
Siegbert

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=3Djoin.php&p=3Dsourceforge&CID=3DDE=
VDEV
_______________________________________________
Bluez-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel

#3 lore raw

Siegbert

Quoted 7 lines

Yep, now it makes sense.
BTW, is this the right list to ask SBC-specific questions? I have some more general questions, e.g. I don't quite understand the rationale behind the codec definition: If you choose joint mode for encoding, it happens that the encoder does not use joint mode for all subbands. But why? Saving bits? If anybody has a pointer to some documentation for SBC, which explains the design decisions I would be really happy.

yes this is the right place.

I haven't worked on it in a while so your question doesn't ring a bell... I do need to get back into it to re-roll the encoder and eliminate the inner loop memory move.

Brad

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Bluez-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bluez-devel