Hi Niclas,
Thank you so much for the prompt response. I've couple of more
questions on the same issue.
1. Can the application request the algorithm to decode multiple frames at
the same time (assuming the inputBuff contains data for multiple frames).
2. In the demo application, the buffer allocation is done based on some
pre-determined sizes (especially for audio and speech decoders). Can this
take care of samples/frames of varying sizes ?
Warm Regards,
Anil
-----Original Message-----
From: Anderberg, Niclas [mailto:[EMAIL PROTECTED]
Sent: Saturday, September 02, 2006 2:57 AM
To: [EMAIL PROTECTED];
[email protected]
Cc: Pitney, Gilbert
Subject: RE: Question on DaVinci EVm Demo apps
Anil,
The reason why XDM_GETBUFINFO is used in vidoe decode is to figure out
what the worst case size of a video buffer is. We are decoding
elementary streams with no headers so there is no way knowing ahead of
time what the size of the next encoded frame is. Instead we pass down a
buffer of size X and the decode algorithm returns how many bytes Y it
actually used using outArgs.bytesConsumed. We have to make that *at
least* one frame is passed to the decode algoritm for each process call,
otherwise the decoder will fail and we will have to drop a frame or
possibly quit the app. To make sure we pass at least one full frame, the
decoder is asked what it's worst case encoded frame size is. This value
is used in X above.
If you are using a transport stream where you know the size of the next
encoded frame this logic is not needed, as you know the exact amount of
data to pass to the VIDDEC_process() call every time.
As to why this isn't done for audio, it's basically because audio frames
are a lot smaller and we are pretty sure we'll give the decoder at least
a frame every time. But perhaps we should for consistency.
Best regards,
Niclas
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On
Behalf Of Anil
Sent: Friday, September 01, 2006 8:38 AM
To: [email protected]
Subject: Question on DaVinci EVm Demo apps
Hi,
I have got a question on Codec Engine demo application in
DaVinci-Linux. I was reading through the decoder application source code
to get an idea about the buffer allocation schemes. The demo application
has support for audio, speech and video decoder. For audio decoder and
speech decoder the inputBuf size is decided by the application itself,
whereas for the video decoder the inputBuf size is decided/queried using
the VISA function "VIDDEC_control"
with the cmd_id "XDM_GETBUFINFO". Why is this not done for the other two
decoders (audio and speech) ??
Also the from the header file "xdm.h" it is evident that in the
structure "XDM_AlgBufInfo", minOutBufSize field can give the outputBuf
size, then why is not being used in any of the decoders ??
Warm Regards,
Anil
-------------------------------------------------------DISCLAIMER-------
-----------------------------------------------
The information transmitted herewith is confidential and proprietary
information intended only for use by the individual or entity to which
it is addressed. If the reader of this message is not the intended
recipient, you are hereby notified that any review, retransmission,
dissemination, distribution, copying or other use of, or taking of any
action in reliance upon this information is strictly prohibited. If you
have received this communication in error, please contact the sender and
delete the material from your computer.
------------------------------------------------------------------------
--------------------------------------------------------
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
-------------------------------------------------------DISCLAIMER------------------------------------------------------
The information transmitted herewith is confidential and proprietary
information intended only for use by the individual or entity to which it is
addressed. If the reader of this message is not the intended recipient, you are
hereby notified that any review, retransmission, dissemination, distribution,
copying or other use of, or taking of any action in reliance upon this
information is strictly prohibited. If you have received this communication in
error, please contact the sender and delete the material from your computer.
--------------------------------------------------------------------------------------------------------------------------------
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
