Disable AudioOptions in Scream and L4S BWE integration tests

WebRtcVoiceEngine::ApplyOptions disables software noise suppression,
echo cancellation, and auto gain control on iOS (WEBRTC_IOS) while
leaving them enabled by default on other platforms. In PeerScenario BWE
integration tests using TestAudioDeviceModule::CreatePulsedNoiseCapturer,
this platform difference caused un-attenuated pulsed noise on iOS to
drive Opus and Audio Network Adaptor to higher bitrates, increasing
bottleneck queue delay in
ScreamTest.AudioOnlyBothDirectionsLinkCapacity40To90KbpsRtt200ms
(causing ICE STUN ping RTTs around 653 ms vs. the 600 ms limit) and
previously requiring CcFbSendRateAdaptation to be disabled on iOS.

Explicitly disable echo_cancellation, auto_gain_control,
noise_suppression, and highpass_filter in SendMedia and
CcFbSendRateAdaptation so all platforms run with identical audio
processing settings, relax the upper RTT bound in
AudioOnlyBothDirectionsLinkCapacity40To90KbpsRtt200ms from 600 ms to
700 ms, and re-enable CcFbSendRateAdaptation on iOS.

Bug: webrtc:42225697, webrtc:447037083
Change-Id: I36660d614cc67b17c9cc56ac610fb332f243c8d6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/516140
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#49263}