[hts-users:01458] Re: gstream redundancy

["Heiga ZEN (Byung Ha CHUN)" <zen@xxxxxxxxxxxxxxx>]

Hi,

Nicholas Volk wrote (2008/06/24 21:57):

> My version had only one for-loop and much fewer memory allocations that
> the original HTS 2.1 RC2 had.

I understand, thanks.

> Is there some reason for duplicating pstream data to gstream data?

This is because we want to keep un-postfiltered spectral parameters for output.
If we don't duplicate pstream, we lose spectral parameters without postfiltering.

> Meaning that we could reduce memory consumption and time spend of making
> duplicates by removing gstreams and writing HTS_GStreamSet_create() simply
> as shown below. (I use pss->total_frame instead of the API function etc,
> but that is trivial.)

You're right, the current one is not efficient in terms of memory and speed.
So in our internal (fast) version, we use the similar (or same) code as you described.

To address this problem, how about switching to the fast one if HTS_EMBEDDED is defined ?

Best regards,

Heiga ZEN (Byung Ha CHUN)

--
------------------------------------------------
Heiga ZEN     (in Japanese pronunciation)
Byung Ha CHUN (in Korean pronunciation)

Department of Computer Science and Engineering
Nagoya Institute of Technology
Gokiso-cho, Showa-ku, Nagoya 466-8555 Japan

http://www.sp.nitech.ac.jp/~zen
------------------------------------------------