- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
In IPP 5.2 speech coding sample there is the following fragment of code that performs PLC:
#define G729_SPEECH_FRAME 80
Ipp16s LostFrame[G729_SPEECH_FRAME]={ 0 };
apiG729Decode(DecObj,(const Ipp8u*)LostFrame,(-1),(Ipp16s*)out->pBuffer);
Why LostFrame array is 160 bytes long? Should it be the size of g.729 frame, ie. 10 bytes?
If the decoder is keeping some state in LostFrame array, should it be allocated per-channel?
If no state is kept - in fact I can't see src parameter is used at all if frametype = -1, can't we just pass src = NULL to apiG729Decode()?
Same for g.723.
#define G729_SPEECH_FRAME 80
Ipp16s LostFrame[G729_SPEECH_FRAME]={ 0 };
apiG729Decode(DecObj,(const Ipp8u*)LostFrame,(-1),(Ipp16s*)out->pBuffer);
Why LostFrame array is 160 bytes long? Should it be the size of g.729 frame, ie. 10 bytes?
If the decoder is keeping some state in LostFrame array, should it be allocated per-channel?
If no state is kept - in fact I can't see src parameter is used at all if frametype = -1, can't we just pass src = NULL to apiG729Decode()?
Same for g.723.
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thank you for pointing this out.
Though 'Losframe' zero padded arraysize was alreadyoptimized in IPP 5.3 for some codecs evidently it was overlooked somewhere. NULL also possible to utilize in many cases.
WBR, Vyacheslav
IPP speech
Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page