Converting audio data with ffmpeg to libextract compatible array of doubles

Using ffmpeg libraries in c++ I am decoding live audio streams to signed 16 bit arrays easily using the codec defined within the stream itself. I now want to encode this input into arrays that are compatible with the libextract library that requires arrays of doubles, with values in the range of -1.0 to 1.0.

So I am using avcodec_encode_audio2 to convert to arrays of doubles. But I must not be doing it correctly as the AVPacket data member variable contains doubles that are outside the acceptable range.

Can anyone advise me on my mistake during the encoding?

_audioCodec = avcodec_find_encoder(CODEC_ID_PCM_F64BE);
if (_audioCodec)
{
    _audioCodecContext              = avcodec_alloc_context3(_audioCodec);
    _audioCodecContext->bit_rate    = 128000;
    _audioCodecContext->sample_rate = 22050;
    _audioCodecContext->channels    = 1;
    _audioCodecContext->sample_fmt  = AV_SAMPLE_FMT_DBLP;

    if (avcodec_open2(_audioCodecContext, _audioCodec, nullptr) >= 0)
    {
        int         got_output  =  0;
        AVPacket    outPacket   = {0};

        av_init_packet(&outPacket);

        int err = avcodec_encode_audio2(_audioCodecContext, &outPacket, &frame, &got_output);
        if (!err && got_output)
        {
            double * dbls = (double*)outPacket.data;
            //
            // When casting outPacket.data to double * the data contains values outside 
            // -1.0 to 1.0 and lots of NAN's. I'm obviously doing something wrong.
            //

...

I have used variations of the codec such as CODEC_ID_PCM_F32LE, CODEC_ID_PCM_F64LE, CODEC_ID_PCM_F32BE, CODEC_ID_PCM_F64BE and I have used variations of AV_SAMPLE_FMT_DBLP, AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_FLT and AV_SAMPLE_FMT_FLTP but to no avail.

Thanks in advance.

James