Handling Core Audio buffer size changes in KFR

[geoffw64]

Hi Dan,

Happy weekend!

Background: Core Audio (AudioToolbox) provides buffers containing varying numbers of samples, up to the buffer size configured in AudioToolbox. While capturing samples from Core Audio, I detect a change of buffer size and re-declare the univectors I store audio in. Purpose of this post is to check that my approach is sensible and safe (or not) as I sometimes (rarely) get nasty crackling during playback.

Version: KFR 6.0.3 avx2 64-bit (clang-15.0.0/macos) +in +ve

Declarations:

univector<float> samples_out;	// Samples received from Core Audio as float
univector2d<double> audio_in;	// Samples converted to double using KFR deinterleave

Code run in my callback that captures samples from Core Audio (AudioToolbox):

size_t const size = buf_ref->mAudioDataByteSize / sizeof(float);
univector<float> const samples = make_univector((float *)buf_ref->mAudioData, size);

Code run when number of samples changes:

univector2d<double> const temp_audio_in(num_channels_in, univector<double>(samples_per_channel));
audio_in = std::move(temp_audio_in);

    univector<float> const temp_samples_out(num_channels_out * samples_per_channel);
    samples_out = std::move(temp_samples_out);

Is this a sensible way to handle buffer size changes?

Thanks in advance,
Geoff.

[geoffw64]

I changed my code to use resize() instead.

[geoffw64]

Closed as answered

[geoffw64]

.