Mixing stereo and 5.1 audio

[Licadnium]

I'm new here. I have a quite advanced edit that I'm working on using Premiere Pro. In some scenes I've used the film's soundtrack to fill in some seconds of footage where I've made changes. So we switch from the movie's audio track to just the song and then back to the movie. It sounds smooth because it's synchronized carefully.

I'm a bit concerned though because the soundtrack file is in stereo (L, R channels) while the movie audio is 5.1 (5 channels). It sounds perfectly fine in my computer but would it sound weird for someone with a better equipment? If so, is there a way to fix it?

 

[Phase3]

I've been in exactly the same situation as you, with my previous editing software (which I no longer use after upgrading to Premiere).

Mixing 5.1 and 2.0 audio in general doesn't sound good...

The best thing to do would be to create a 5.1 audio version of the soundtrack music.
I've made 5.1 tracks in Audacity, and they sound great!

 

[addiesin]

Yeah, two options.

1 Export your edit in stereo only

2 Create fake surround with your stereo tracks. You may use a plug-in that does it for you, you may add stereo music to a surround video project and export just that to see if it does an ok job, or you may manually do it;
.....

For the last one, manually faking surround, get the free program Audacity, open your stereo track in it.

1. Duplicate the track and mute the original, leave it untouched and unedited as reference.
2. On the duplicate track, split left and right channels into separate tracks. Name them "L" and "R" (or whatever works for you to convey the same info at a glance). Pan the left one 100% left and pan the right one 100% right. These separate tracks will be used as left and right in your surround mix, and you don't really have to do much to them, they're already left and right.
3. Duplicate the individual L and R tracks again (so you should have the muted original stereo track plus four panned mono tracks now) and name them "BackL" and "BackR", the new ones can be rear left and rear right easily, maybe turn volume down for both rears and/or add some light reverb to them.
4. Then duplicate the original reference track again and don't split the channels this time, and don't pan, instead downmix them to one mono track you can use as the center channel. Rename it to "C". Duplicate it again and name the copy "LFE" which we'll edit on the next step. Turn volume down for "C" compared to the L and R, I think at least minus 6dB, but your mileage may vary. You may also want to use an EQ filter/effect on the center to lower the volume specifically for frequencies that will be shared by dialogue or sound effects in your edit. (Utilize Google to find normal human voice frequency ranges and experiment to find a setting that doesn't hurt the tone of the music but also won't fight with or muffle the dialogue).
5. Then edit your LFE track. cut frequencies above 40hz with a low pass filter.
6. Then select all other tracks, excluding the LFE and the original reference. With your selection, cut frequencies under 40hz with a high pass filter. After this, the lowest frequencies will only be in the LFE.
7. Important step, listen to it. Volume WILL need to be lowered, you essentially have the same frequencies layered three to five times over top of each other playing through probably two computer speakers. Make sure your overall volume is not clipping at any point in the timeline. If it is, adjust individual tracks' volumes until it feels and sounds even and you don't get any dips into red/volume over 0dB.
8. Arrange tracks in correct order, which I think is L, R, Center, LFE, BackL, BackR
9. Export as WAV but change settings to either export as one 5.1 surround file or 6 mono individual files that you can edit now in your video editor.

 

[futon88]

I don't know about Premiere, and maybe I'm doing it incorrectly, but in Vegas it's as simple as ensuring your project is configured for 5.1, and creating a track with the surround panning set in a way that closely mimics where the music lives in the original mix. Most of the time, I'll use a track with the pan somewhat forward, and with the center channel muted:

This places most of the music in FL and FR, and some in BL and BR, with none in C, while letting me work with a simple stereo file. If the music is bleeding into C for creative reasons, I'll sometimes intentionally reproduce that by duplicating the track, moving the pan all the way forward, and muting all channels but center. Then I'll set the volume really low, like -30db.

If you're unsure where the music should go, listen to your source file and try to copy it, but also pay attention to the other sounds. You may have a scene that is mostly music, but chances are there is still some minor ambient noises happening in BL and BR. If in doubt, watch the output monitors while listening to the original mix.

 

[futon88]

(Part of me worries Vegas is down-mixing the song to mono, then spreading that mono sound across the various channels and stomping out the true L/R nature of the original recording...and that part is speaking very loudly right now, so I need to go do some experiments...)

 

[Phase3]

(Part of me worries Vegas is down-mixing the song to mono, then spreading that mono sound across the various channels and stomping out the true L/R nature of the original recording...and that part is speaking very loudly right now, so I need to go do some experiments...)

I had a similar problem with Power Director & Audio Director, where 5.1 was downmixed to 2.0 during editing.
When the audio was transferred back to the editing software, the audio would spill over / bleed into the LFE channel, leaving me with awful sound!

 

[M4_]

I use premiere and you can configure it pretty easily so that stereo audio left/right corresponds to left/right surround channels (usually rear is quieter) and then center is a mono mix- but it can be confusing at first to set up and get used to. One thing I would recommend though for people who don’t edit in a 5.1 studio (most of us lol) is if you have your audio levels pulled up (the green flashing bars) you can solo individual channels by clicking under them so for every transition I do I give a listen with just front channels, then just center channel, just rear, etc and that way you can 90% ensure it will sound perfect with a surround speaker set up even if you just have headphones. Usually you just have to balance the volumes.

Sometimes soundtrack has more vocals mixed in center and rear channels have reverb, depends on the movie though and if the soundtrack has vocals/choir in it. You can try to fake this with some EQ and other effects.

 

[futon88]

I had a similar problem with Power Director & Audio Director, where 5.1 was downmixed to 2.0 during editing.
When the audio was transferred back to the editing software, the audio would spill over / bleed into the LFE channel, leaving me with awful sound!

I tested my process on a song with some clear stereo dynamics, and those were maintained after mixing across six channels in Vegas (whew). Crisis averted. Carry on.

 

[krausfadr]

In Premiere it’s easy to turn stereo into surround. You just duplicate it as a 5.1 track to get a total of 5 tracks and re-assign each LR channel to the following channels

L R > L R (normal volume)
L > C (lower the volume)
R > C (lower the volume)
L R > Ls Rs (slightly lower the volume)
L R > LFE (add a high pass filter and lower the volume)

I found this process tedious when adding a lot of music, so I use the plugin Nugen Halo Upmix to do it automatically (and better with no possibility of phase cancellation).

 

[Phase3]

In Premiere it’s easy to turn stereo into surround. You just duplicate it as a 5.1 track to get a total of 5 tracks and re-assign each LR channel to the following channels

L R > L R (normal volume)
L > C (lower the volume)
R > C (lower the volume)
L R > Ls Rs (slightly lower the volume)
L R > LFE (add a high pass filter and lower the volume)

I found this process tedious when adding a lot of music, so I use the plugin Nugen Halo Upmix to do it automatically (and better with no possibility of phase cancellation).

That's a good way to do it - I'll look into the plugin you mentioned too!

 

[Licadnium]

Thanks a lot for your answers! I followed your instructions in Audacity, @addiesin and it works quite well, I think the end result is basically the same as what @krausfadr is suggesting. Can you explain more about how much you lower the volume? Right now I'm trying -8db for C and -16dB for LFE and it seems fine but I have no idea what's the norm.

 

[addiesin]

Thanks a lot for your answers! I followed your instructions in Audacity, @addiesin and it works quite well, I think the end result is basically the same as what @krausfadr is suggesting. Can you explain more about how much you lower the volume? Right now I'm trying -8db for C and -16dB for LFE and it seems fine but I have no idea what's the norm.

It's going to be really fluid unfortunately. There is no real set answer. In film audio the music levels dip and spike all over the place to fit the tone and fill the scene and duck away when dialogue or SFX comes in, etc.

Consider all the numbers in this paragraph as guesses you can use as a starting point but not specific recommendations or in any way a bible for what to do. In general, music in the center track is quiet but somewhat common so I would keep that very low in comparison to the rest, at least -12dB but maybe up to -32dB, it really depends on how your source film is mixed. I wouldn't lower LFE volume (while making the track) since there's only one "copy" of those frequencies in the setup I described, it's not getting doubled up. If you must lower LFE, max it out at -3. Front L and R should be loudest and their volume changes should match (anything done to L should be done to R so if L is -5dB then set R to -5 as well), and back L and R should be quieter but not as quiet as the center channel (-9dB to -18dB?). You could have some fun with it and use an AI audio track separator/isolator to fiddle with the center track and only include some sounds, or you could use an audio envelope, which allows you to fluidly change the volume over time, and have the center channel only come in louder when the music picks up in overall volume. Lots of options to experiment for that. You could even just exclude it if your source film tends to have a clean center channel (clean meaning no trace of music).

I have a few ideas you can try to figure out appropriate volumes. The first is pull into Audacity some existing already-professionally-mixed 5.1 audio from your source film where it's the loudest (like a title theme or end credits music or a montage or something) and try the manual fake surround process on a stereo version of the same track. You could try to closely match the levels in your audio editor, since the actual music in this case would be a direct 1:1 comparison. Then you could roughly follow the volumes that match nicely as a guideline for all your upmixed audio for that project.

The second idea is, don't even worry about it that much, because if you don't get the levels right when you make the wav in Audacity it's not the end of the world (as long as there's no clipping) you can always lower the volume in your video editor to match the surrounding music levels. A worse scenario would be lowering volume too much then trying to crank it back up in your video editor, that would increase baseline noise level and introduce some kind of white noise, like a hiss.

 

[ParanoidAndroid]

A previewer noticed that an edit of mine had audio channels that appeared to be swapped, think this was a case of DaVinci Resolve having some sound settings as 5.1 and others as 5.1 Film (different channel ordering).

With only a stereo setup to hand, are there any ways to check that the surround audio track of the exported edit would be correctly interpreted by the hi-fi, so the centre channel is not playing through the right speaker and the LFE not in the left-surround!

 

[addiesin]

A previewer noticed that an edit of mine had audio channels that appeared to be swapped, think this was a case of DaVinci Resolve having some sound settings as 5.1 and others as 5.1 Film (different channel ordering).

With only a stereo setup to hand, are there any ways to check that the surround audio track of the exported edit would be correctly interpreted by the hi-fi, so the centre channel is not playing through the right speaker and the LFE not in the left-surround!

The super low tech way is just throw that bad boy (your exported file) into Audacity and look at the waveforms. It should load as the equivalent of six mono wav files. L,R,C,LFE,BL,BR I think is the standard 5.1 order. You can mute or solo any of them and spot check through the whole timeline.

 

[Possessed]

Back in the days of Womble DVD architect it would automatically do what @addiesin just described for you automatically, minus the LFE step. That's literally the only step you have to take though. If it was a Dolby surround source it would even do an actual proper upmix where the proper sound effects wemt to the surround channels etc. I'm really surprised that isn't a common feature in nle's to this day. Or maybe it is in some? I've never tested it. Womble is so primitive and it could do it so you'd think it would be a commonplace feature.

 

[addiesin]

Back in the days of Womble DVD architect it would automatically do what @addiesin just described for you automatically, minus the LFE step. That's literally the only step you have to take though. If it was a Dolby surround source it would even do an actual proper upmix where the proper sound effects wemt to the surround channels etc. I'm really surprised that isn't a common feature in nle's to this day. Or maybe it is in some? I've never tested it. Womble is so primitive and it could do it so you'd think it would be a commonplace feature.

It either is or should be, but this is just another way to do it but while being sure of what you're going to get.

 

[Possessed]

It either is or should be, but this is just another way to do it but while being sure of what you're going to get.

Agreed. The pro logic upmix should be included though, in case you want to edit in let's say a deleted or alternate scene or audio from an older source if the new mix is inferior then you could still get proper surround for the sound effects

 

[Catlit]

There's a foobar2000 plugin called FreeSurround that has highly customizable upmixing features. I've only tried it a couple times but there are other members of the forum who know way more about it (@Hydra Spectre comes to mind)