Decode AAC-LC (ADTS) framed from NSData

Question

Anders Cedronius OP

Created Sep ’15

Replies 1

Boosts 0

Views 2.2k

Participants 1

Is there a 'simple' way of decoding and playing AAC frames without going the coreaudio route?

Is this even possible?

I get ADTS frames from the network I put in NSData

Then I create a AudioStreamBasicDescription matching what's inside the ADTS header.

        AudioStreamBasicDescription audioFormat;
        audioFormat.mSampleRate = 48000.00;
        audioFormat.mFormatID = kAudioFormatMPEG4AAC;
        audioFormat.mFormatFlags = kMPEG4Object_AAC_LC;
        audioFormat.mBytesPerPacket= 0;
        audioFormat.mFramesPerPacket= 1024;
        audioFormat.mBytesPerFrame= 0;
        audioFormat.mChannelsPerFrame= 2;
        audioFormat.mBitsPerChannel= 0;
        audioFormat.mReserved= 0;

        CMFormatDescriptionRef format = NULL;
        OSStatus status = CMAudioFormatDescriptionCreate(kCFAllocatorDefault, &audioFormat, 0, nil, 0, nil, nil, &format);

After that I remove the ADTS header and retain the data.

Then i create a CMBlockBufferRef:

            CMBlockBufferRef frameBuffer;
             OSStatus status = CMBlockBufferCreateWithMemoryBlock(NULL, (void *)[audioData bytes], [audioData length],
                                               kCFAllocatorNull, NULL, 0,[audioData length], 0, &frameBuffer);

And finally a CMSampleBufferRef :

            CMSampleBufferRef sampleBuffer;
            status = CMSampleBufferCreate(kCFAllocatorDefault, frameBuffer, true, NULL /readyCallback*/,
                                 NULL, format, 1, 1, &timing, 0, NULL, &sampleBuffer);

Then I just put the audio data into the "AVSampleBufferDisplayLayer" where i put the H264 video.

[renderLayer enqueueSampleBuffer:sampleBuffer];

No sound of cource.

Question 1: Can a "AVSampleBufferDisplayLayer" deal with audio?

Question 2: Is there a non coreaudio method of decoding AAC audio... audioengine? Something simple without going down the 200 lines of code route?

Thanks for helping..

/A

Boost

Answer 1

Anders Cedronius OP

Sep ’15

The most lightweight solution I came up with was.

    NSData *AACdata=[weakSelf getAACFrame];
    memcpy(buffer->mAudioData, (unsigned char*)[AACdata bytes], (UInt32)[AACdata length]);
    buffer->mAudioDataByteSize=(UInt32)[AACdata length];
    AudioStreamPacketDescription packetDescriptor;
    packetDescriptor.mStartOffset = 0;
    packetDescriptor.mVariableFramesInPacket = 0;
    packetDescriptor.mDataByteSize = (UInt32)[AACdata length];
    OSStatus err= AudioQueueEnqueueBuffer(weakSelf.audioQueue, buffer, 1, &packetDescriptor);

Thanks for helping 😉.

/A

0