SOUND PLAN FOR STREAMING

Like your on-air signal, audio streams need to be processed for level consistency and absolute peak control. But unlike your on-air signal, you’re dealing with the streaming codec, whose job is to remove bits from the audio stream to get the best sound possible across the internet. 

Problems with aggressive processing. Codecs use bit reduction to reduce the size of the bitstream and then perceptually hide the fact that those details have been taken away. Problems arise when introducing noise, hum, and distortion byproducts, which the codec can mistake for perceptible audio. Not only does the codec often pass these on as listenable content, it can multiply these artifacts to the exclusion of listenable audio. 

This is why aggressive AGC, limiting and clipping techniques used in broadcasting are especially problematic for streaming. The better plan for streaming is to anticipate overshoots earlier in the processing chain to avoid the thick program density that can set off issues with the codec. The five-band AGC section used in our Streamblade, Wheatstream and now new Wheatstream Duo streaming appliances is designed to eliminate aggressive RMS attack times that can interfere with codec performance. Instead, sophisticated algorithms utilize current and historical data to steer processing to exactly what is required at that instant.

Hard no for clipping. Clipping by its very nature creates additional harmonics that the codec reappropriates as distortion and is therefore not recommended for peak control of streamed content. For the occasional peak that might go over the recommended peak input level of –3 dBFS for most codecs, our Streamblade/Wheatstream and new Wheatstream Duo appliances rely on dual-band limiting rather than clipping to do the heavy lifting.

Separate limiting bands. Final limiting occurring separately in two bands can be helpful when balancing perceived quality/loudness against finite bitstreams. For this reason, we separate final limiting in two bands, above and below 180Hz. For low bitrate streams especially, separately managing low frequencies means being able to selectively remove low frequencies in the stereo difference channel (L-R). This leaves more bits available for encoding those frequencies that are more perceptible.

Stereo, how much is too much?  Big swings in dynamic L-R can skew the codec algorithm in favor of L-R information at the expense of listenable content. For very low bitstreams, monaural might produce the best outcome. Wheatstone streaming appliances have bass boost features for adding depth to low bitrate streams and a stereo width management section to get just the right stereo that doesn’t set off the codec. 

Vocal correction, EQ, and noise filters needed. Streamed programming can often benefit from EQ adjustments, especially at low bitrates. And like broadcast audio, streams often require some noise filtering and asymmetry correction on vocals to make the most of the listening experience. Streamblade, Wheatstream and new Wheatstream Duo have a two-stage phase rotator to correct voice asymmetry, selectable high- and low-pass filters for removing noise and hum, and four-band parametric equalizer with peak and shelf functions for streaming the best quality possible at any variety of bitrates.

Finally, any sound streaming plan will take into consideration your studio and your budget. Wheatstone streaming appliances come in a WheatNet IP audio network Blade option (Streamblade) or standard option without Blade I/O (Wheatstream) for provisioning up to 8 streams, each with four outputs for a total of 32 streams. More recently, we introduced Wheatstream Duo with all the audio processing tools mentioned above, but for one- or two-channel applications.