Added new AudioProcessing fuzzer
This is a high-level fuzzer that creates an AudioProcessing instance.
All possible combinations of publicly visible components are
fuzzed. Input and output sample rate, call order and use of the
float/fix interface is fuzzed. Sample rate may change between calls.
To fuzz floating point numbers, raw data is converted to floats, and
filtered for special values like 'inf', 'nan' and very large
values. Note that the default use case of APM is to only allow values
between +/- 2^15.
BUG=webrtc:7820
Review-Url: https://codereview.webrtc.org/2876793002
Cr-Original-Commit-Position: refs/heads/master@{#18678}
Cr-Mirrored-From: https://chromium.googlesource.com/external/webrtc
Cr-Mirrored-Commit: 8c51282f7fd7e712628aba1677a9733d903eb219
4 files changed
