Struggling with automation workflow

I’m using ardour to edit some recordings I made and I am having some trouble figuring out an easy-to-use workflow for this. Now I am wondering if my workflow is structured the wrong way, I am missing some Ardour features that would make this easier or I should just accept that this is a lot of manual work and clicking.

The project I am working on is a multi-track recording of a backing track (karaoke recording of songs) and some singers with individual microphones. The recording contains a number of different songs with different people each time (so a single microphone is used by different people in different songs, not all microphones are used in all songs).

What I plan to do is create CD markers around each song so I can export songs individually later. Each song will need its own mix, for which it seems logical to use automation. For now, just controlling mutes and fader levels would be sufficient, at a later point I might also want to add per-song EQ and FX.

What would seem like a good workflow to me is to:

  1. first setup mutes for the song,
  2. then play the song a few times to find a good overall fader level for each channel.
  3. I might be done at that point, or maybe individual parts of the song need additional level changes.

I am mostly struggling with step 2. I think I know what the end result here should be (four automation points: One before the song starts with the previous fader level, one directly after with the fader level for the song, and then two more at the end of the song in reverse order), but I am wondering how to get there in an ergonomic way. On second thought, maybe just two points might be enough, since my songs are not back-to-back and I do not really care about the fader levels in between songs.

In my head, a sensible workflow would be to set up the perimeters of the song, add automation points there at the current level (to ensure making no changes to the outside of the song) then play the song and adjust the fader to a decent level. Then you would need some way to apply the fader level that you ended up with to the automation track (i.e. create the two automation point “inside” the song).

Some things I tried:

  • You can select part of the automation track (using range markers), then delete, which inserts two automation points at the start and end of the song at the current automated level. However, this only seems to work for a single channel at at time.
  • You can manually add points to the automation track, but that is cumbersome, especially because I have half a dozen channels.
  • If you manage to create points at the start and end, you can select the automation line and drag it (which moves the points left and right up and down). This might be closest to what I want to do, but it is not very ergonomic: it is dragging on the automation lanes instead of the main fader, cannot be (AFAICS) surface-controlled, and finding the right channel in the editor interface with automations visible is not easy. Maybe this could be an extra automation mode, though, where moving the fader does not insert points (like with write/touch/latch), but modifies the existing points before and after the current playhead position? Or one alternative could be to modify all points in the current song (i.e. selected/active range or something like that?).
  • I can set the mode to “write”, fiddle with the faders until they sound right, then delete all points up to the final fader level, but then I would still need to (manually?) create a point at the start of the song at the final level. Also, write mode is disabled whenever you stop, run backwards or change the playhead position, so that is annoying if you want to skip back and hear something again (also because there does not seem to be an easy way to set the automation mode of a bunch of different channels at the same time).
  • I can set the mode to manual, find the right fader levels, then switch to write mode and play the entire song, which ends up creating the right points/line. But this means having to play the entire song (even when not intending to make further changes) and again setting tracks to write mode is cumbersome. If there are no automation points beyond the current edit point, then playing a part of the song could also work: This produces an open-ended automation line that should keep the fader levels further automations are added later in the timeline, but might also end up with just a single automation point midway the song (where you stopped playing) instead of having one at the beginning and end.

A common theme is that setting automation modes to multiple channels seems to involve a lot of manual work. I know there is a Lua “Engage Automation” script, but that always enages all automation tracks (which I think I might not want) and still involves selecting the relevant channels first (but I guess that can be made easier by creating a selection-only group that I activate only when I need to select the channels in the group). I also know the ctrl-shift-click on the fader automation button in the mixer, but that always does all channels (which might be ok except for write mode) and only applies to the fader automation (which is probably the one you want to change most).

Some completely different approaches I considered:

  • Mixer snaphots per song. I think these will not play well with exports and are limited to 8, so that would not work.
  • Separate session per song. This would involve cutting up the single recording into a lot of session, which seems to be a lot of work up front, and then might run into a lot of duplicate setup in terms of plugins, FX, etc. later.
  • Separate tracks per song. I guess you could just duplicate each track for each song where it is used and then trim the regions/playlist in the track to just the range of the song. Then you would need no automation for changes between songs, but this would end up with tens if not hundreds of channels, which I think will not make things easier.

So maybe I am approaching this all wrong? Or am I missing some tools that are available (maybe some lua scripts to make this easier)? Any thoughts or suggestions? How are others approaching their mixing in this area?

Just a few thoughts Every point is a method on its own. They can be combined of course but i don’t mean to do everything pointed out. Some of the methods recreate the same outcome in different ways and therefore there is no point doing both.

Normalisation

Split the regions of the tracks that have highly alterning levels at each song then normalize the regions. So you have at least a more stable level from the start.

using VCA

choose one song as reference. Level everything to taste. Group the vocalstracks to a VCA. Then you can adjust the levels for each song without fiddling inside the voxtracks.

add tracks

split the vocals to new tracks for each singer to get more detailed control over each performance.

Just read your post again and the track method you mentioned yourself. I’m not sure how you’ll end up in 100s of tracks as you mentioned.

I hope i didn’t miss anything from your post. If so, sorry! Otherwise, hope it helps.

Thanks for your input!

That might indeed work to get more even levels (especially for the backing track, as I did not spend any time normalizing these beforehand and instead just used the fader for that during the live performance), but I still expect I would want to set the specific levels for each song (but this could remove one track from the equasion, i.e. use the region normalization to set a base level for the backing track and then mix the voices relative to that level).

For each song, I have 1 or 2 close-miked singers, one overhead channel to capture the original venue reverb/ambiance (really 4 mics, but joined together in one VCA), and a backing track. That gives me 3 or 4 channels per song. I have 20 or so songs in the recording, so that would be 60-80 channels (so not really hundreds, but more than I think would be convenient).

Gotcha. Question would be, do you need to adjust all tracks but i get your point. Just to be sure: You know you can attach a track to several VCAs? I think that might be useful to get control over various tracks instead of handling each track fader but adds another level of complexity in the process.

Why would you split each song into a separate set of tracks, rather than just having 3 of 4 channels, with different songs at different points of the time line?