Topic: Post-processing, mixing, and mastering w SBD feed

[checht]

Happy Friday!

Curious as to what folks are up to in post when recording ambient sources plus a line or SBD feed.
Post-processing?
How do you mix it?
Any mastering effects?

Inquiring mind wants to know...

[grawk]

if they were recorded on a single recorder, it's fairly straight forward.  Normalize each source individually, then align tracks by zooming in to match the start times to account for the speed of sound, then listen, adjusting the balance between soundboard and audience mics til you like the result.  Listen to any edits using as many different playback methods as possible so you can make sure the results sound good everywhere.

[adrianf74]

if they were recorded on a single recorder, it's fairly straight forward.  Normalize each source individually, then align tracks by zooming in to match the start times to account for the speed of sound, then listen, adjusting the balance between soundboard and audience mics til you like the result.  Listen to any edits using as many different playback methods as possible so you can make sure the results sound good everywhere.

QFT.  Also make sure you've got your sources mapped correctly (i.e., channel 1 of the ambient source is the same as channel 1 of the board feed).

[rocksuitcase]

if they were recorded on a single recorder, it's fairly straight forward.  Normalize each source individually, then align tracks by zooming in to match the start times to account for the speed of sound, then listen, adjusting the balance between soundboard and audience mics til you like the result.  Listen to any edits using as many different playback methods as possible so you can make sure the results sound good everywhere.

I use Audacity.
As an example of a stereo SBD pair and a stereo AUD pair:
I open/load the SBD first, then name it. Then import the AUD and name it.
DO NOT MAKE ANY edits to either pair yet.
SAVE (export selected audio) the SBD as bandnamdDATESBD (raw or take00 I use here)
SAVE (export selected audio) the AUD as bandnamdDATEAUD (raw or take00 I use here)

Now you can add delay to the SBD if your AUD mics were more than 25-30 feet from stage. (the formula is .9ms for each foot of distance. I just did one where the AUD was 60 feet away, i used 54 ms delay)
Then adjust the levels of each pair to the other. I typically start with the AUD down between 6-9dB from the SBD.(of course, THIS is the ratio you play with, listening carefully as you make changes- it is never the same ratio between any two recordings. And, as GB says below, you may wind up reversing course and using the AUD as primary, it ALL depends on your program material)
play with the slider, adjusting the AUD up and down to get the proper amount of room ambiance you prefer. ( I try to get the AUD so I can JUST barely distinguish it using the mute or solo buttons)
Before you decide to save the mix, check it over more than one playback type. Headphones AND speakers etc.

To save the mix, select both pairs and do Export selected audio, it will warn you that you are exporting to two tracks, Say OK, name the track properly, e.g. bandnameDATEsbdaudmics.wav

Now, open your mixed track and see what else needs done. This is where you equalize, adjust final L/R ratio, final levels (I prefer amplify up to -0.2DB than normalization) and name it.
this is when I cut out before and after chatter/silence. DO not do cut anything when looking at all four tracks as you can cause odd issues.

Phew! Hope this helps.     

[nulldogmas]

I'm fortunate to have two A10s with well-matched clocks, so I can usually just do:

1) Bring the AUD source into iZotope RX and reduce close clapping with the de-click filter.
2) Import both sources into Audacity, trim the start of one so that they're aligned, and adjust the mix to taste.

If I'm having trouble getting a good balance of instruments, sometimes I'll bring one source or the other or both into RX, use Music Rebalance to create separate stems, and then bring the whole mess into Audacity to noodle with the gain sliders for a while. But I try to avoid it when possible, though my new Macbook does at least mean the stems render in an hour or two rather than taking overnight.

[Gutbucket]

Same as the folks above with regards to the front end work of aligning and normalizing the AUD and SBD.  Add to that any repair fixes needed for either source alone in isolation, such as level corrections dropout fixes, noise reduction, obvious EQ needs, stuff like that.

When it comes to combining the sources, I tend to use the AUD as my primary source and foundation of the mix, bolstering that as needed with the SBD feed.  So I first work on getting the AUD balanced and sounding its best.  For most folks that means L/R balance and possibly EQ.. maybe a Mid/Side decode to L/R.  However for me running OMT rigs consisting of four, six, or eight mic channels, there is actually a lot of inter-channel balancing and back and forth going on in the course of this step.  This step is more or less the same as when I only have AUD without SBD.  However when I do have SBD, I don't concern myself as much with the vocal and instrument clarity / articulation as much as I otherwise would.

Once I'm happy with that I play around with adding SBD to support vocal and instrument clarity, articulation, and sometimes a closer sense of proximity to the stage.  I never really pay much attention to what the relative levels between the two are, I just determine the most appropriate relative levels is by ear. That's partly because the best relative mix always seems to vary each time.  Usually it ends up being primarily AUD with just enough SBD to add whatever is needed yet underrepresented in the AUD.  In support of that I'll EQ the SBD to help bring out what I need from it.  However, on occasion when the AUD suffers from some problem like way too much chatter, or just doesn't sound all that great regardless of my efforts, I'll go the other way and start with the SBD as primary source, adding AUD to that to add a sufficient sense of ambiance, depth, and audience reaction.

Once I've achieved a an appropriate balance of the two sources, I then decide if I want or need to go further or not.  Going further might include working on overall dynamics.  It might also include dynamic manipulation of just one source or the other.  It most often involves playing one source off the other to some additional benefit.  Doing things like subtracting EQ from one source - more than would be desirable when listening to that source alone - to allow for contributing a bit more in that region from the other source.  The overall response and energy distribution remains the same, but one source is allowed to step in a bit more strongly in certain areas while the other makes room for that to happen.  The vocal presence range is a good example - I may scoop the AUD a bit there to allow for adding more clear SBD in that range without making the overall sound overly forward or bright. That can work quite nicely but can be somewhat fidgety to balance as neither source will then works quite as well on its own in isolation as it previously did.  Besides EQ, this may also involve different manipulation of dynamics in some subset of channels within one of the two primary sources - a good example is the dynamics of my rear-facing mic channels.  I like to sort of tie that to the overall energy level, such that when things get louder, the contribution from the rear-facing channels is less than when things are quieter.  I currently do that with envelopes, but may try side-chaining compression on the rear facing channels to the front channels or to the sum of all channels.

^ The playing one against the other thing is where I'd really like to find some tool capable of providing "differential" control of level, EQ or dynamics across multiple channels - some function or routine that keeps the overall output the same while by automatically varying one source as I change the other.  Could be applied to level, EQ, dynamics or whatever.  Level is the easiest to conceptualize- as I lower the level on one source it should automatically raise the other to compensate such that the output level remains the same.  Without that kind of tool it requires being especially careful to re-balance things between each tweak so as not deceive oneself.

I do use some of the RX-like features in Samplitude, and plan to get in to RX proper at some point, but doing all this stuff is already a lot!  Most would probably say, too much.

[checht]

Many thanks Gut and everyone for your input, much appreciated. I generally follow the procedure outlined above, then I go off piste:

I’ll use Steely Dead as my primary example, because those recordings represent at least half of the material I’ve worked with lately.
Using my MP-6, I record 2 onstage pairs - mk41s ortf 10' high in uber-overhead position on the house R front corner of the drum riser. MK22s split omni, 10” above the deck centered on the stage, 1’ foot upstage from the lip. SD use in-ear monitors, which means no wedges and no vocals onstage. Therefore I use the mp6's 2 line-in channels for vocals.. I pull from the monitor mixer so I can mix my own using the web interface and also don't have to deal with the various FOH staff at venues.

Before mixing the vocals w the mics, I do a fair amount of processing on the vocal channels these days:
1. I use rx music rebalance to effect a debleed on them, so I have isolated  vocals w/out more instruments.
2. I use Audacity’s compressor in a 3:1 or 4:1 setting to even out the vocals a bit.
3. Finally a little  reverb using Audacity’s lightest preset (vocal 1) to reduce the dryness of the vocals. Without reverb, they stand out in a bad way, incongruously, much too present and really felt different from the ambient  input.

Next I bring the vocals into the mix. I start by mixing my 2 pairs, then slowly add vocals.

Finish up with a light dose of NY compression on the master, and it’s done.

An example from Steely Dead Hawai‘i tour:  https://archive.org/details/sd2024-02-23/sd20240223.matrix.1648-01.flac

Interested in whether anyone else has gone this far down the rabbit hole, and/or feedback on what I’m doing.

Random thoughts:
1. For other bands’ SBD feeds, like Los Lobos, I often use music rebalance to untangle the feed and bring in what I want. Usually vocals, sometimes guitar or keys. I’ve found that SBD feeds of those instruments add texture to the tone. The ambient can sometimes have a one-dimensional feel, and lack the complexity that’s available. Then I go through the whole obsessive process above.
2. I feel that most of the time when my recordings sounds bad, they're too distant and muddy. So lately I’ve been using eq on the mud and also running music rebalance to pull out the vocals and mix them back in to supplement. Really brings presence. The 2 combined have worked wonders on my 90’s KM84i recordings.
3. Steely Dead hand me a usb stick with 24 tracks from the monitor mixer each night, but that would take too long to mix, especially when on tour with a show a night plus travel. I usually just go with my mp6 recording, unless I’ve blown the vocal mix and/or there’s something extra, like a guest musician my setup didn’t capture. They laughed evilly the first time they handed the usb stick to me, saying ‘we sure as shit don’t have time to mix these haha’

[grawk]

I don’t try to fix the mix in the venue, I just try and capture it as accurately as i can. 

[checht]

^ I get this.

And, choices of mics, downstream gear, mic technique, and position in the house all have substantial impact on the recording.

In all I'm doing, my goal is to find the best-sounding outcome. I don't think that folks are as interested in fidelity to the sound resulting from my underlying gear choices as they are interested in a great-sounding recording. Especially given that the majority weren't at the show to begin with.

Also, what is 'the mix' when I'm getting ambient plus SBD that isn't being heard directly by anyone at the show?

Hope this doesn't sound argue-y, just trying to 'splain my intent.

[hoserama]

It really depends on the soundboard mix.

I have found that the new stemming tools (I use Demucs) really help a lot with uneven soundboard mixes, which are usually vocal forward. So when mixing with an audience, you can adjust the soundboard levels of bass + vocals + drums + everything else along with adjusting the audience recording. 

[AbbyTaper]

It really depends on the soundboard mix.

I have found that the new stemming tools (I use Demucs)

Is there an option to save files in 24 bit with Demucs?

[morst]

if they were recorded on a single recorder, it's fairly straight forward.  Normalize each source individually, then align tracks by zooming in to match the start times to account for the speed of sound, then listen, adjusting the balance between soundboard and audience mics til you like the result.  Listen to any edits using as many different playback methods as possible so you can make sure the results sound good everywhere.

I absolutely do not utilize this workflow.

[AbbyTaper]

if they were recorded on a single recorder, it's fairly straight forward.  Normalize each source individually, then align tracks by zooming in to match the start times to account for the speed of sound, then listen, adjusting the balance between soundboard and audience mics til you like the result.  Listen to any edits using as many different playback methods as possible so you can make sure the results sound good everywhere.

I absolutely do not utilize this workflow.

I like the sources to be as close to the same level as possible before mixing.

[Gutbucket]

^Sort of a "pay me now or pay me later" type thing, mostly about workflow.  Can be helpful beforehand if some channels are significantly different in level. Wouldn't want to do that if trying to maintain some level relationship across channel pairs established while recording. I don't really care one way or another since I always end up adjusting levels by ear in the end anyway.. at least for everything except the four channels of an ambisonic mic where maintaining a truly accurate level match across the channel set is necessary for it to work correctly.

Good to hear of the practical applicability of "stemming tools" such as rebalance and demucs.  I've not used them yet myself but have followed their development and they now seem to be a really excellent tool for our use.

Kudo's to those getting it done night to night on the road, streamlining the process out of necessity.  A good effective working process is a thing of beauty.  I've plenty to learn on this account.

Regarding intent, my goal is producing a listenable, enjoyable recording that mentally transports the listener to the time and place of the performance and allows them to experience the music in that live-performance context.  I'm not really sure what accuracy means in this context. Or rather, I sort think of accuracy in an inverse way.  Done right it sounds "most real" and especially truthful to the listener's expectation of "accurate", whether they were present at the performance or not.  Done well the resulting recording often sounds subjectively better than it did live to me.  If "accurate" in pursuit of how it sounded live were to sometimes require that the resulting recording sound less good, less convincing and make for a less enjoyable listening experience than I'm otherwise able to manage, then accuracy in itself is not sufficiently compelling and not what I'm trying to do.  I suppose I'm seeking accuracy as measured by the listener's mindset rather than by some measure imposed at the front end. The choices I'm making are all determined by shaping the end result.. by the listener's awareness, the last thing that occurs.  I'm not attempting to produce accurate acoustic measurements of bands in venues, I'm recording musical performances intended to be enjoyed.

I suppose the pitfall of this approach is the potential for all of one's recordings ending up sounding somewhat the same - good and enjoyable, yet somehow not translating essential differences between different acts performing in different venues on different days.  Fortunately I don't find this to be the case, yet it is something I think about and listen for.  I suppose that is inevitable to some extent, manifesting as the "house sound" of each taper, as reflected though one's choice of gear, setup, post mixing, and everything else.

[checht]

^Sort of a "pay me now or pay me later" type thing, mostly about workflow.  Can be helpful beforehand if some channels are significantly different in level. Wouldn't want to do that if trying to maintain some level relationship across channel pairs established while recording. I don't really care one way or another since I always end up adjusting levels by ear in the end anyway.. at least for everything except the four channels of an ambisonic mic where maintaining a truly accurate level match across the channel set is necessary for it to work correctly.

Good to hear of the practical applicability of "stemming tools" such as rebalance and demucs.  I've not used them yet myself but have followed their development and they now seem to be a really excellent tool for our use.

Kudo's to those getting it done night to night on the road, streamlining the process out of necessity.  A good effective working process is a thing of beauty.  I've plenty to learn on this account.

Regarding intent, my goal is producing a listenable, enjoyable recording that mentally transports the listener to the time and place of the performance and allows them to experience the music in that live-performance context.  I'm not really sure what accuracy means in this context. Or rather, I sort think of accuracy in an inverse way.  Done right it sounds "most real" and especially truthful to the listener's expectation of "accurate", whether they were present at the performance or not.  Done well the resulting recording often sounds subjectively better than it did live to me.  If "accurate" in pursuit of how it sounded live were to sometimes require that the resulting recording sound less good, less convincing and make for a less enjoyable listening experience than I'm otherwise able to manage, then accuracy in itself is not sufficiently compelling and not what I'm trying to do.  I suppose I'm seeking accuracy as measured by the listener's mindset rather than by some measure imposed at the front end. The choices I'm making are all determined by shaping the end result.. by the listener's awareness, the last thing that occurs.  I'm not attempting to produce accurate acoustic measurements of bands in venues, I'm recording musical performances intended to be enjoyed.

I suppose the pitfall of this approach is the potential for all of one's recordings ending up sounding somewhat the same - good and enjoyable, yet somehow not translating essential differences between different acts performing in different venues on different days.  Fortunately I don't find this to be the case, yet it is something I think about and listen for.  I suppose that is inevitable to some extent, manifesting as the "house sound" of each taper, as reflected though one's choice of gear, setup, post mixing, and everything else.

Very nicely articulated Mr. Bucket.

Regarding your final paragraph, your point is well taken and hits home. I work with the band to get the recordings to sound how they like/want, and they like consistency. Also, my ambient mics are all onstage, and the pair closest to the house faces upstage, away from the house. So that's limiting how much the venue impacts the sound right there. I like to normalize pre mixing, b/c these days I have a pretty good sense of what ratio of pair 1 to pair 2 to SBD channels to start with, and that's a timesaver. At the same time, I believe that while all of their reoordings may have a familial relationship, my other recordings are quite heterogeneous.

Cheers