1db of attenuation before edcast for low bitrate stream

Depends on the bit rate and encoder. aacPlus with SBR may need a little more. MP3 may need a little less.

Check with an oscilloscope hooked up to the decoder. Breakaway RTA works great for this.

One thing to remember is that aacPlus decoders have built-in limiters, and MP3 decoders don’t.

aacPlus has virtually no peak control (due to SBR) whereas MP3 passes a processed waveform pretty well, with only a few overshoots.
So, you’ll want to make sure you don’t feed the aacPlus encoder to hot that the limiter at the decoder end starts working, as it creates extreme amounts of IM-distortion.

On the other hand, if a sample of MP3-overshoot gets clipped here and there, that doesn’t really matter.

To completely avoid limiter action, you may need 3dB attenuation for aacPlus, but 1dB is almost certainly enough for mp3. You’ll have to try it yourself and find what tradeoff works best for you. Let me know what you come up with 🙂.

///Leif

As far as using RTA as a scope and meter… You say to check for clipping and distortion at the decoder end. The ways I’ve tried to do this, it’s always affected by whatever I set the volume to. so it’s not really "attached" to the decoder say in Winamp. It’s attached to the wave out. On the receiver end, how would I accurately check if there is clipping with RTA? Jesse showed me a cool winamp dsp plugin that gives me a spectrum analyzer and a meter with clipping indicators and it’s attached right to the decoder and is NOT affected by any volume controls. How do I make RTA do that?

Timmy, it’s simple.

You know that the output of Breakaway Broadcast is extremely tightly peak controlled. You’ve seen what the (de-emph) output oscilloscope looks like.

So, let’s say you set the BBP output level (into the encoder) to -1dB. Then, you tune into the stream with Winamp, and turn Winamp’s volume controls to full (and turn off the EQ).

Then, set Breakaway RTA to look at the output of Winamp, and look at the scope. You’ll be able to visually tell where the -1dB boundary is (and if not, hop into the I/O settings in Breakaway RTA and set the Scope Lines to -1.0 dB.

Do you see "too many" peaks go beyond the lines and hit the rails (0.0 dB)? If so, you need more headroom. Otherwise, maybe you could get away with less.

So how many is too many? Well, it depends on when peaks occur (on bass/sibilance/voices/percussion etc), and how often they occur.

For example, if you see lots of peaks appear when there’s S sounds, then that’s horrendously bad. Clipping S-sounds in this manner would sound extremely dirty — even turning S-sounds into F’s.
On the other hand, if you see just a little peak here and there, particularly if they’re of the single sample variety, then that’s nothing to worry about, and perfectly acceptable.

Also, make sure to listen. In headphones. The adjustment between loudness and audio cleanliness is a critical tradeoff wherever it occurs.

Best,
///Leif