Stereo Tool

Hans, would it be possible to create a variable de-emphasis level for the Processed Output? Maybe a slider or preset checkbox levels. This way we can stream lower bit-rates as well as have FM output. I am already using the processed output for streaming but it lacks the pre-emphasis needed.

_________________
J0HN II

If FM output is enabled, you're already getting de-emphasized output as normal (non-FM) output (pre-emphasis value is 50 or 75 us). The only way to have a variable pre-emphasis here would be to do all the processing twice - in fact you could then just as well run 2 copies of Stereo Tool...

If you can use 75us de-emphasis on a 75us pre-emphasized ouput to make it a normal output, can't you just apply less de-emphasis to the normal output to create a 25us output or am I way off in my thinking here? I had just assumed that the de-emphasis was applied at the end of the normal output.

_________________
J0HN II

No.

Ideally, pre-emphasis + de-emphasis should cancel each other out. So if you use pre-emphasis of 75 us and de-emphasis of 25 us, you're left with a pre-emphasis of 50 (which results in extremely loud highs).

Or is that what you want (and if so: Why? Lack of highs?)

Currently, pre-emphasis for normal output (in Loudness&Limiting window) is used to pre-emphasize, then limit, then de-emphasize, which results in a reduction of loud high frequencies, which in turn avoids distortion in lossy compression.

OK, now the light has gone on I have never streamed low bit-rates before (but I am planning to in the future) and was under the impression that the highs had to be pre-emphasized to reduce the quantization noise of the high frequencies (which never made sense to me because that just would make it worse). So the pre-emphasis/compression-limiting/de-emphasis is used to control the peakiness of the high frequencies which would reduce the quantization noise?
Then NVM this suggestion.

_________________
J0HN II

Not so much the quantization noise, but highs are 'changed' more than other frequencies (because they usually contain hiss, and more is filtered than on other freqs). This leads to bigger (different) peaks in the decompressed signal than in the original signal. Which, if the signal was already close to the maximum amplitude, results in more clipping, hence more distortion, than lower frequencies. This leads mainly to (badly) distorted 'S' sounds.

Pre-Amp + clipping + De-amp results in lower peak levels for highs, which reduces this effect.

Alternatively, the whole signal would need to be lowered in amplitude - this basically avoids having to do that.