AAC | ADPCM | Convolution | DAFx | FFT | IIR | Mixer | MP3 | MPEG | MPEG-4

Nasim-

>            i followed ur discussion abt two q15 number multiplication and addition,
> i have one doubt on two q15 format multiplication. as u told that in multiplication
> of two q15 format we hav two sign bit :
>  So we are left with,
> > > > 0 I's 2 S's 14 F's
> > > > So now you left shift it by 1 to get rid of sign,
> > > > 0 I's 1 S 15 F (lsb fraction bit being 0)
> i am agreed with that , but question is which sign bit is considered as extra. i
> mean if one number is +ve ( so 0 sign bit) and one number is -ve(1 sign bit) then
> after multiplication first number sign bit 0 comes first and second number sign bit
> 1 comes second then as per the rule second bit that is 1 would be considerded as
> extra bit and decimal point would be placed between 0 and 1, 0 would be considered
> as a sign bit. that mean result will be a +ve number but result must be -ve number.
> i hope u understand my doubt and u will clear my doubt. thanks in advance.

The most significant sign bit (msb) is the extra.  This is why many fixed-point
processors include an option for "automatic left-shift after multiply".

-Jeff

(You need to be a member of audiodsp -- send a blank email to audiodsp-subscribe@yahoogroups.com )

both signs bit will always be same...you r doing something wrong if they ar=
e=20
different...
>From: Jeff Brower <j...@signalogic.com>
>To: Nasim Ahmed <m...@yahoo.com>
>CC: a...@yahoogroups.com
>Subject: [audiodsp] Re: q15 multiplication
>Date: Wed, 17 May 2006 12:08:16 -0500
>
>Nasim-
>
> >            i followed ur discussion abt two q15 number multiplication=20
>and addition,
> > i have one doubt on two q15 format multiplication. as u told that in=20
>multiplication
> > of two q15 format we hav two sign bit :
> >  So we are left with,
> > > > > 0 I's 2 S's 14 F's
> > > > > So now you left shift it by 1 to get rid of sign,
> > > > > 0 I's 1 S 15 F (lsb fraction bit being 0)
> > i am agreed with that , but question is which sign bit is considered as=
=20
>extra. i
> > mean if one number is +ve ( so 0 sign bit) and one number is -ve(1 sign=
=20
>bit) then
> > after multiplication first number sign bit 0 comes first and second=20
>number sign bit
> > 1 comes second then as per the rule second bit that is 1 would be=20
>considerded as
> > extra bit and decimal point would be placed between 0 and 1, 0 would be=
=20
>considered
> > as a sign bit. that mean result will be a +ve number but result must be=
=20
>-ve number.
> > i hope u understand my doubt and u will clear my doubt. thanks in=20
>advance.
>
>The most significant sign bit (msb) is the extra.  This is why many=20
>fixed-point
>processors include an option for "automatic left-shift after multiply".
>
>-Jeff

=20

=20

(You need to be a member of audiodsp -- send a blank email to audiodsp-subscribe@yahoogroups.com )