Down mixing stereo to mono

[dweeeeb2]

I don’t have a MLP, in general I move about listening to music. I’m thinking that mono could be an advantage here. Down mixing is proving a bit more complicated than what I first thought, primarily cancellation issues from phase differences between channels.
Does anyone here down mix? How do you do it? Any software suggestions?

 

[JaapD]

Just Lmono=(L+R)/2 and Rmono=(L+R)/2. I don't believe so much in possible cancellation issues. If there are cancellation issues then it will also occur in the stereo image during playback. I'm using dbPoweramp myself.

 

[dweeeeb2]

Yeah, I’m thinking purely theoretical and trying to get the best result possible. Thing is any cancellations that do occur during mixing will now apply to the entire field not just the MLP, hence my question. Thanks for your reply.

 

[ernestcarl]

Yeah, I’m thinking purely theoretical and trying to get the best result possible. Thing is any cancellations that do occur during mixing will now apply to the entire field not just the MLP, hence my question. Thanks for your reply.

I thought the whole purpose of listening to mono in the production side was to determine if the mix still sounds good when listening with mono speaker systems -- since not everybody has a proper stereo setup. I've always just summed it left and right with levels reduced as well to listen to my podcasts in mono when at home. Music certainly almost always loses some of that stereo magic but with some equalization it still sounds pretty good for the most part.

 

[TimVG]

Just Lmono=(L+R)/2 and Rmono=(L+R)/2. I don't believe so much in possible cancellation issues. If there are cancellation issues then it will also occur in the stereo image during playback. I'm using dbPoweramp myself.

Not always. For instance pure AB recording with omnis. Lots of comb filter effect when summing to mono. If you're lucky nothing too offensive audibly speaking.
Or M/S recording with figure 8+center mic (usually cardioid). Collapsing to mono cancels out the side channels entirely. So it works in mono, because at that point it is mono - at the expense of losing everything except the center mic.

 

[Chromatischism]

Makes me wonder how do devices like the HomePod handle stereo?

 

[dweeeeb2]

I confess to not knowing much about this at all so I'm just trying to understand it. I've heard people say that EQ can help but wouldn't the eq need to be adjusted for each song? And I assume by ear as the uneven FR is being driven by the downmixed content not the speaker?

 

[kemmler3D]

If there are cancellation issues then it will also occur in the stereo image during playback

Not true, cancellation of out of phase content will be 100% when performed on the signal, much less than 100% when L+R interact acoustically due to 1) two ears each having different distances to the speakers, 2) reflected / scattered sound.

This is why simple stereo widener effects that invert and delay part of the signal sound interesting on speakers, and hollow and lifeless summed to mono.

In general you can get pretty objectionable comb filtering on some content summed to mono in DSP or electrically, or not, depending on how it was recorded and mixed. By and large the effect shouldn't be horrible, as @ernestcarl mentions it's standard practice (or at least used to be) to check mixes in mono for big problems when summed. I still think everything summed to mono sounds kinda lifeless compared to stereo, but in isolation I never think how awful it is. On the other hand I only listen to mono on my JBL Go or countertop radio.

IMO there are no perfect solutions here, you can just listen to L or R alone, which IMO gives cleaner sound but you tend to miss entire sounds that are panned the other way, you can sum with the aforementioned issues, or you can just put stereo speakers close together. I don't think there is much point to using a mid side processor, as the drawbacks are similar to listening to just one channel but it takes more work to get there.

The reason to take a stereo recording down to one channel is for compatibility with a system that only has one speaker. I don't think there are any clear and significant advantages if that's not a restriction you're dealing with.

 

[dweeeeb2]

Thanks @kemmler3D it does feel like I’ trying to do something that the medium is not designed to do. It’s a shame because to my ear the single speaker definitely sounds cleaner whilst moving about, but then down mixing degrades it.

 

[DVDdoug]

Down mixing is proving a bit more complicated than what I first thought, primarily cancellation issues from phase differences between channels.

Are you doing it in hardware or software? You can't just short left & right together but you can get a passive mixer which has resistors so you don't have a "hard short".

Most recordings are made to be mono compatible so phase cancellation problems are rare. But often there is a bit of out-of-phase reverb that's lost.

Some "amateurs" use phase inversion for a widening effect so they aren't mono compatible. I've heard of cases where someone plays a song on their (mono) phone speaker and the vocals are gone!

...One weird thing happens when mixing-summing. The left & right amplitude (voltage or digital data) is summed, not the power. That increases the center by +6dB, whereas when the same power comes from both speakers the acoustic level sums to +3dB. So when you sum digitally or electrically the center goes-up by +3dB compared to stereo.

 

[JaapD]

For a moment stand still how the mono (and R-L) signal at a FM stereo radio station will be created. (Background info: the radio station transmits a mono signal as well as a R-L signal on a 19 KHz. carrier, for compatibility reasons to mono and stereo radios). The mono signal will be created as I described. So if you listen to an FM radio station with a radio having a single speaker then you’ll hear the signal as I described.

I’m not denying that you ‘can’ get comb filter effects, depending on the recording itself, but it’s not that you by default ‘will’ get comb filter effects. The risk of comb filter effects will be largely in case during the recording the microphones are very close (within a foot or so) to each other, almost negligible in case the microphones are multiple feet apart. So ‘it depends’.

 

[teched58]

For a moment stand still how the mono (and R-L) signal at a FM stereo radio station will be created. (Background info: the radio station transmits a mono signal as well as a R-L signal on a 19 KHz. carrier, for compatibility reasons to mono and stereo radios). The mono signal will be created as I described. So if you listen to an FM radio station with a radio having a single speaker then you’ll hear the signal as I described.

I’m not denying that you ‘can’ get comb filter effects, depending on the recording itself, but it’s not that you by default ‘will’ get comb filter effects.

In strict technical terminology, that mono signal you mention is a L + R. So an FM stereo station transmits a L+R and an L-R. A mono receiver will play the L+R. In a stereo receiver, if the stereo receiver detects the 19 kHz carrier, the stereo receiver will demux (use the demultiplexer circuit) to generate the stereo output by adding and subtracting both signals from each other.

So the sum of L+R and L-R = 2L. That's your left channel

And the difference, that is, L+R minus L-R = 2R. That's your right channel.

 

[JaapD]

No de-multiplexing in a single speaker radio. Just the mono (L+R) signal. The L-R signal on the 19 KHz carrier will be completely excluded as there is there is no de-multiplexing circuitry built in.

 

[dweeeeb2]

I’m using a miniDSP to sum channels, I did attenuate both channels initially by 3dB but found I should actually use 6dB. It is very track dependent, obviously mono recordings sound fine.

 

[dweeeeb2]

One thought I’ve been playing around with is whether the ATMOS format would work better for downmixing. If it truly is object based then could a ATMOS decoder deal with panning better?

 

[kemmler3D]

One thought I’ve been playing around with is whether the ATMOS format would work better for downmixing. If it truly is object based then could a ATMOS decoder deal with panning better?

I think summing Atmos down to mono should be less problematic, at least to the extent that there is less overlapping out-of-phase content in each 'object'. As long as you are sure the device you're using is mixing down to mono using that information and not just feeding you L+R again, it's probably preferable when you have the option.

 

[dweeeeb2]

Yes, I was hoping I could use the ATMOS app on pc and output to external DAC but it needs to be recognised as ATMOS capable.
Thanks for your help I don’t think I’ll be getting the result I’m after. Bugga

 

[gasolin75]

It's not software you can by but wiim amp,streamer has that options, stereo,mono,left or right

 

[Multicore]

Move the speakers close together.

 

[dweeeeb2]

Move the speakers close together.

Out of all my options this is probably my best. I just need to try it and see how it sounds moving around the room.

It's not software you can by but wiim amp,streamer has that options, stereo,mono,left or right

Cheers, down mixing to mono is actually the easy bit, and it appears that its almost always done the same way Mono = (L+R)/2. The problem is that if an instrument is panned somewhere between L and R it can have its L sound out of phase with it R sound. When listening to stereo our brains detect this but it varies as you move your head, therefore we can localise the sound to where the producer intended. But if you downmix to mono this phase difference leads to varying levels of cancellation which gets baked into the mono track. This interference does not change when you move your head so it can significantly change the tone.