Yes, that’s logical.
But I was thinking of how the waveforms combine in the room, when not mixed.
Consider a stereo program, with distinct and uncorrelated signals in the two channels. When played in stereo over two speakers, those signals combine in the room in power mode, it seems to me. Yes, you could make an argument that it is always vector edition, amplitude mode, but unless the recording was perfectly time aligned and the speakers are perfectly time aligned and you are locked into the sweet spot, that won’t be perfect and on average you would get power combination. Isn’t that reasonable? I’m not sure, haven’t seen anybody make this argument, but it seems plausible to me.
So if the use case is that I have a lot of stereo music, and I normally play it over two speakers which (per the above) does power mixing in the room, and I want to mimic that by combining the signals in DSP which is amplitude mode and I intend to play that through a single speaker, I would need to do -3 dB.
In contrast, if I have a mono program, with the same signal in both channels, they would vector-add in the room, more or less depending on seating position, and the analysis would lead to -6 dB as you say.
I’m not sure if this thinking is right. I should do a test, but I’m otherwise busy at the moment.
Yes, amplitude is a scalar, but the signal is a vector, having amplitude and phase. The signal is a complex number.
Adding two amplitudes to get 6 dB is valid only if the two signals are in phase, and for them to stay in phase they have to be strongly correlated, in fact identical and not delayed.
If you have two uncorrelated signals and add them electrically, you would on average get 3 dB, power addition.
But it is true that if the concern is with clipping, the electrical addition of the uncorrelated signals could on occasion be in phase and add amplitude, for 6 dB.
So it comes down to what the concern is. It is true that to avoid clipping we need to think like that.
I agree. I hadn’t been really clear in my mind what the purpose was.
Doing 6 dB will avoid clipping; it will also lower the sound level with stereo signals in but that is not a problem, that’s why we have volume control.
I need to change it, I use mono mixing to feed a single speaker in the kitchen, and I used 3 dB as someone had recommended.
Yes I understand the uncorrelated argument, but I still don’t think it holds. Think again of a mono recording, if the level changes in this case the mix is at the incorrect level. This is an edge case you must satisfy - if the signal is not changed at all, the level shouldn’t change either.
Yes, for the mono recording, it’s obviously 6 dB.
An important use case.
The other use case is a stereo recording with one speaker, that’s where I was thinking 3 dB in order to get the right sound level. But as I said, that’s wrong because it can occasionally cause clipping.
I don’t know how big an issue the occasional, bad luck-alignment clipping is. But people are very concerned with intersample overs…
But it’s tricky, if you set it up with -6 dB, mono signals will sound right but stereo signals will be 3 dB too low, compared to stereo playback.
All ok if you are the meticulous type who changes DSP preset for each album, but the UI does not make that convenient.
Clipping is the issue. I don’t agree that the signal type matters at all. Even if you feel that the stereo blended to mono is lower than the original, you still have the clipping problem, which would happen with any signal shared between channels.
So why not avoid the clipping altogether and raise the output volume by 3 dB if you feel it is too low?
Geoff, I’ve visited that KB article and only see a very brief description of Procedural EQ. No examples of how to build a mono preset. Am I going to the wrong link?
I don’t see it either, however, if you go back up to the top of the thread, just after post 10 or so, you will a long post by Brian on how to set things up.
@Bart_Stephens - I wrote the article up on the old KB site, which has since been replaced by the new Help Portal. Looks as though the article wasn’t transferred. @dylan - did this get missed?
I have one of my speakers I built set up with the above method works really well. Make sure you give plenty of DSP headroom though as it can clip quite easily on some music.
Stereo is a trap! I discovered it many years ago. I realized that “the beauty of stereo” fools my ears and steals my attention, so that my focus is unknowingly shifted to how it sounds instead of what and how is being played. Since then, I listen to music downmixed to mono. I only have stereo in headphones and even this using binaural DSP.
Many people I know do the same. Most are musicians or former musicians. The crowd of people who go back to mono for everyday listening is growing all the time.
I used this method to break in a pair of speakers. Set it up as mono but L/R out of phase. Speakers facing each other. It is amazing how the volume just diminishes radically. Even with the speakers several inches apart. Angle one out 45 degrees and the volume jumps back up! Very handy.
