[oggonachip] Re: [vorbis-dev] mdct.c

[Pattara Kiatisevi]

[cut]

> > spectrogram of a peak-to-peak square wave.  I seem to remember that the
> > output range is indeed +-1 however I can't recall if it is clamped at this
> > or if it will stray outside given the right kind of input data.
>
> What I did was to insert a code to check maximum and minimum value of
> input and output vector in mdct_backward() and observe it :).
>
> It would be great to know the limit of both input and output (and both
> floating- and fixed- points.

Ok, today I have tried this MDCT_INTEGERIZED option and finally made it to
run. I set the input/output ranges for mdct_backward() to be: floating
point -1 to 1, integer -32768 to 32767. I check the result from both of
methods and found they are slightly different. I can't notice the
difference for the output music though (are there supposed to be some
"noticable" differences?)

My plan (on the software side):

-Add to ov_read an option to use the output integer value if
MDCT_INTEGERIZED is set (no need for conversion).

-Find the way to do it faster for conversion of input vector from float to
int (may use assembly or do in hardware).

Please comment..

Thanks for all helps,
Pattara

ps.BTW I am not clear about that 32-bit * 32-bit  --> 64 bit
multiplication problem, now I can run and get (quite) correct result so it
means that gcc on i386 is doing the right thing for me? Could someone
please help explain...

<p><p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.