Speex dev - Jul 2010 - notification: More than two wideband layers found. The stream is corrupted.

Dear Speex Devs,

One more problem from my side. I had earlier queried about bad quality of
speex.

For this run, I changed the memory mode option in CCStudio to none from
large. (I will explain this if required.)
Now, I am trying to run speex on C55x in real time.

Find a code snippet from my main function below.

Basically, I am getting the error mentioned in the subject line.
Weird thing is that, for the statements marked in bold below, if I enable
printf, I don't get this error, whereas if I disable it, Speex complains
with this error.
Is it because of the delay required?

By having a look at the code, and the code structure which I have mentioned
in the code itself, can you throw some insight?
Also why should this error occur at first place? I searched the forum but
could not find the exact solution.

Awaiting reply.

Vijay


            if(canEncode == 1)   //sets to 1 when DMA is ready with 20ms
samples with 8khz sampling
            {
                canEncode = 0;

                if(busyEncoding == 0)  // to guard encoding and decoding
                {
                    busyEncoding = 1;

                    //copy 20ms data (160 is the frame size) to encode

                    for(i_loop = 0; i_loop < FRAME_SIZE; i_loop++)
                    {
                        if(CurrentRxR_DMAChannel == 2)
                        {
                            inout_byte[2*i_loop] = RcvL1[i_loop];
                            inout_byte[2*i_loop+1] = RcvR1[i_loop];

                        }
                        else
                        {
                            inout_byte[2*i_loop] = RcvL2[i_loop];
                            inout_byte[2*i_loop+1] = RcvR2[i_loop];

                        }
                    }
                    byte2word(inout_byte, in_short, FRAME_SIZE);
                    speex_bits_reset(&bits);
                    returnVal = speex_encode_int(encoderState, (spx_int16_t
*) in_short, &bits);


                    nbChars = speex_bits_write(&bits, cbits,
sizeof(cbits)*BYTES_PER_CHAR) /BYTES_PER_CHAR;
                    speex_bits_rewind(&bits);

                    speex_bits_read_from(&bits, cbits,
TESTENC_BYTES_PER_FRAME);

*                    //printf("%d\n",counterDecode++);
                    counterDecode++; *

                    returnVal = speex_decode_int(decoderState, &bits,
(spx_int16_t *)out_short);

                    if(returnVal != 0)
                    {
                        printf("speex decode int falied %d %d",
returnVal,
counterDecode);
                        exit(0);
                    }

                    speex_bits_reset(&bits);
                    skip_group_delay = 0;

                    busyEncoding = 0;
                    //printf("encode-decode complete\n");

                    //copy decode data to playback through DMA

                    for(i_loop = 0; i_loop < FRAME_SIZE; i_loop++)
                    {
                        if(CurrentTxR_DMAChannel == 1)
                        {
                             XmitL1[i_loop] = out_short[2*i_loop];
                        }
                        else
                        {
                            XmitL2[i_loop] = out_short[2*i_loop];
                        }
                    }

                    for(i_loop = 0; i_loop < FRAME_SIZE; i_loop++)
                    {
                        if(CurrentTxR_DMAChannel == 1)
                        {
                            XmitR1[i_loop] = out_short[2*i_loop + 1];
                        }
                        else
                        {
                            XmitR2[i_loop] = out_short[2*i_loop + 1];
                        }
                    }


                     busyEncoding = 0;

                }
                else
                {
                    //printf("encoder called when busy\n");
                }
            }
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/speex-dev/attachments/20100725/6293f47d/attachment.htm