From tandelkaran at gmail.com  Tue Jul 14 06:29:22 2015
From: tandelkaran at gmail.com (karan tandel)
Date: Tue, 14 Jul 2015 18:59:22 +0530
Subject: [Tremor] Tremor Digest, Vol 67, Issue 2
In-Reply-To: <mailman.3.1383073202.13557.tremor@xiph.org>
References: <mailman.3.1383073202.13557.tremor@xiph.org>
Message-ID: <CA+_q5_qzsPi9OmHADyUTQN4RupseFFp-kimAh8DpOgEUAZZG_w@mail.gmail.com>

Hi all,
   I have started working on vorbis again and this time We want to port
vorbis's fixed point encoder implementation. Do we have any fixed point
vorbis encoder? If no then I am going to work on it. Please suggests the
steps to do that.

Thanks,
Karan

On Wed, Oct 30, 2013 at 12:30 AM, <tremor-request at xiph.org> wrote:

> Send Tremor mailing list submissions to
>         tremor at xiph.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         http://lists.xiph.org/mailman/listinfo/tremor
> or, via email, send a message with subject or body 'help' to
>         tremor-request at xiph.org
>
> You can reach the person managing the list at
>         tremor-owner at xiph.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Tremor digest..."
>
>
> Today's Topics:
>
>    1. fixed_point_encoder (karan tandel)
>    2. Re: fixed_point_encoder (Ralph Giles)
>    3. Re: fixed_point_encoder (Timothy B. Terriberry)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Tue, 29 Oct 2013 18:58:35 +0530
> From: karan tandel <tandelkaran at gmail.com>
> Subject: [Tremor] fixed_point_encoder
> To: "tremor at xiph.org" <tremor at xiph.org>
> Message-ID:
>         <
> CA+_q5_qcRdP3NZdC4OPGTcwanWXxWgs+Km7xE+csepBgp0ynbA at mail.gmail.com>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Hi,
>    I have to port libvorbis encoder on Android. But as it is floating point
> only Its not showing good timing performance. So for that I need to port
> fixed point encoder. I have done so much searching  but I didnt find any
> fixed point vorbis encoder. So please provide some pointers regarding
> these issues.
> 1. Is there any fixed point encoder  available?
> 2. If not then what should I do to make it fixed point implementation?
>
> Thanks in advance,
> Karan
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> http://lists.xiph.org/pipermail/tremor/attachments/20131029/3b3308c5/attachment-0001.htm
>
> ------------------------------
>
> Message: 2
> Date: Tue, 29 Oct 2013 10:52:49 -0700
> From: Ralph Giles <giles at thaumas.net>
> Subject: Re: [Tremor] fixed_point_encoder
> To: tremor at xiph.org
> Message-ID: <526FF5F1.3070005 at thaumas.net>
> Content-Type: text/plain; charset=ISO-8859-1
>
> On 2013-10-29 6:28 AM, karan tandel wrote:
> > Hi,
> >    I have to port libvorbis encoder on Android. But as it is floating
> > point only Its not showing good timing performance. So for that I need
> > to port fixed point encoder. I have done so much searching  but I didnt
> > find any fixed point vorbis encoder. So please provide some pointers
> > regarding  these issues.
> > 1. Is there any fixed point encoder  available?
>
> Not that I'm aware of, no. It's unfortunate.
>
> > 2. If not then what should I do to make it fixed point implementation?
>
> Write fixed point forward versions of the inverse coding functions in
> tremor, like the forward mdct, floor curve generation, etc. Then port
> over the analysis and mode choice code from libvorbisenc. The main issue
> is to ensure you maintain enough dynamic range during calculations.
>
> Then submit your code for review an incorporation upstream, of course.
> :-)
>
> See also https://bugzilla.mozilla.org/show_bug.cgi?id=932390
>
>  -r
>
>
> ------------------------------
>
> Message: 3
> Date: Tue, 29 Oct 2013 11:03:07 -0700
> From: "Timothy B. Terriberry" <tterribe at xiph.org>
> Subject: Re: [Tremor] fixed_point_encoder
> To: tremor at xiph.org
> Message-ID: <526FF85B.6080507 at xiph.org>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> Ralph Giles wrote:
> > over the analysis and mode choice code from libvorbisenc. The main issue
> > is to ensure you maintain enough dynamic range during calculations.
>
> The good news is that because you're writing an encoder, not a decoder,
> you don't have to support everything (in particular, floor0, which is
> the hardest part to do with enough precision). So in that sense encoding
> is much easier than decoding.
>
>
> ------------------------------
>
> _______________________________________________
> Tremor mailing list
> Tremor at xiph.org
> http://lists.xiph.org/mailman/listinfo/tremor
>
>
> End of Tremor Digest, Vol 67, Issue 2
> *************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.xiph.org/pipermail/tremor/attachments/20150714/7de78241/attachment.htm 

From tterribe at xiph.org  Tue Jul 14 07:03:30 2015
From: tterribe at xiph.org (Timothy B. Terriberry)
Date: Tue, 14 Jul 2015 07:03:30 -0700
Subject: [Tremor] Tremor Digest, Vol 67, Issue 2
In-Reply-To: <CA+_q5_qzsPi9OmHADyUTQN4RupseFFp-kimAh8DpOgEUAZZG_w@mail.gmail.com>
References: <mailman.3.1383073202.13557.tremor@xiph.org>
	<CA+_q5_qzsPi9OmHADyUTQN4RupseFFp-kimAh8DpOgEUAZZG_w@mail.gmail.com>
Message-ID: <55A516B2.6060100@xiph.org>

karan tandel wrote:
>     I have started working on vorbis again and this time We want to port
> vorbis's fixed point encoder implementation. Do we have any fixed point
> vorbis encoder? If no then I am going to work on it. Please suggests the
> steps to do that.

As in 2013, there is still no fixed-point Vorbis encoder. Ralph's advice 
earlier in the thread is still a good starting point for what to do.