[vlc-devel] karaoke filter

[Konstanty Bialkowski]

Hi All,

+ for( i = 0; i < i_samples; i+=i_channels) {
+      for( i_chan = 0; i_chan < i_channels; i_chan++) {
+          p_in[i*i_channels+i_chan]

Correct me if I am wrong but the index for p_in should just be i+i_chan -
the size of i_channels is already taken into account in the samples loop
with i+= i_channels (and i_samples is the total number of samples for all
the channels).


Konstanty

2011/4/21 郭宇宏 <yuhong.guo68 at gmail.com>

> the latest patch
>
> 在 2011年4月20日 下午9:50，郭宇宏 <yuhong.guo68 at gmail.com>写道：
>
> Modify the patch according to all your suggests.
>>
>> Thanks
>>
>> 2011/4/15 Juha Jeronen <juha.jeronen at jyu.fi>
>>
>>> Hi,
>>>
>>> On 04/15/2011 02:25 PM, Steinar H. Gunderson wrote:
>>> > On Fri, Apr 15, 2011 at 07:16:04PM +0800, 郭宇宏 wrote:
>>> >>     for(i=0;i<len;i+=nch)
>>> >>     {
>>> >>         a[i] = (a[i] - a[i+1]) * 0.7;
>>> >>         a[i+1]=a[i];
>>> >>     }*
>>> > My guess is that this is really 1/sqrt(2), which makes sense from an
>>> energy
>>> > point of view (or at least would, if the channels were independent; if
>>> not,
>>> > it depends on the correlation), but can cause clipping.
>>>
>>> Ok. Thanks. It did look a lot like 1/sqrt(2) :)
>>>
>>> Maybe for float samples (if we support those), 1/sqrt(2) would make
>>> sense, and for integer (16-bit) samples, 0.5?
>>>
>>>  -J
>>>
>>> _______________________________________________
>>> vlc-devel mailing list
>>> To unsubscribe or modify your subscription options:
>>> http://mailman.videolan.org/listinfo/vlc-devel
>>>
>>
>>
>
> _______________________________________________
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
> http://mailman.videolan.org/listinfo/vlc-devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/vlc-devel/attachments/20110421/4af4e339/attachment.html>