The many types/stages of audio files in a DAW session

Here are a few definitions so we can all talk about the same thing!

Raw recording: you just named the track, pressed record, generated audio file(s)

Takes: on the same track, you created new “playlists” and recorded over the same section on the timeline alternate “takes”

Edited track: you removed noises, you created fades, you adjusted levels (clip gain)

Compilation (comp): you selected different takes and gathered them on the same “playlist” on a track

Edited and processed track: see edited, you then added processing (EQ, compression, delay, reverb, autotune, amplification modeling etc.) – note that these processing are “live” as in using real-time plug-ins

Consolidated track: this applies only to the clips being rendered as a new audio file (ideally this is done on edited track so there are no noises, fades will be rendered, clip gains will be rendered and all parts appear on the timeline where they are supposed to be). It is a good practice to consolidate your track from the very beginning of the session even if the actual sounds only appear later in the arrangement. This way a consolidated track (i.e. audio file) can be imported in any workstation, lined up at Bar 1 and provided the tempo of the session is set properly, that track will play “in sync” with others. The consolidated track still maintains “live” processing.

Committed track: this is similar to the consolidated track with the addition of “printing” or “committing” the processing to a new audio file. This is done before the mixing stage.

Bounce, stack, blend, pre-mix: would refer to a mixed down version of several tracks usually belonging to a group of instruments. A good example is a vocal “stack”. When recording many individual tracks to create harmonies it is sometimes a good idea to “commit” to a blend of these tracks “into” a mono or stereo audio file. That blend ideally is made of cleaned (edited tracks) possibly Edited and processed tracks. The new audio file should ideally be “printed” starting at Bar 1 and be clearly labeled as a “blend” or “bnc” of the original tracks.

Mix (bounce or recorded): a Mix track is the result of either a “bounce” if working “in the box” or a recording via an interface of a blend of all tracks of a project. Ideally the audio file should start at Bar 1 and not necessarily contain a fade-out (since the fade out can be worked out during the Mastering stage).

Stem: this term is now overused and misused. Sometime confused with the “committed track”. In my opinion, a stem is created after the final mix of a song has been approved and “printed” (see Mix definition above).

A stem is an audio file (ideally starting from Bar 1) that is an isolated “layer” of the Mix. The stem represents the isolated “layer” the way it “sits” in the mix. That means that the stem (always a stereo file by the way!) is not necessarily an “optimized” version of said layer in terms of loudness/volume but simply the isolated layer the way it “sounds” in the mix. That means that the stems are the result of all the processes applied to that track (or group of tracks) such as equalization, compression, reverbs, delays etc. Including “where” that track (or group of tracks” fits in the balance of all tracks.

Rule 1 for stems: stems cannot be “printed” through a Master buss containing processing. This is especially true of “dynamic” processing. The master buss processing reacts to the sum of all tracks. An individual stem will not contain the same amount of energy as the whole Mix (sum of all tracks and effects) making the buss processing react differently and therefore changing the sound of that stem.  It is imperative to “bypass” all Master buss processing before “printing” stems through that Master buss. If the Master buss processing is essential to the sound of the Mix it could be “re-applied” later to the sum of all the stems.

Rule 2 for stems: stems audio files should all be printed from the same timeline starting point (I call it Bar 1 as a simplification). This is essential as the only compatible type of file between all DAW is the “wav” audio file. Printing every stem from the same timeline starting point is the only way to easily ensure that all the stems once imported in any DAW will play “in sync” with each other.

Rule 3 for stems: you must check the stems to ensure that each one only contains what they are supposed to contain. If for example a “Vocal” stem also contains a bass guitar part, it totally voids the usefulness of that “Vocal” stem. It becomes unusable and is then a complete waste of time and resources.

Stem A + Stem B + Stem C + …. + Stem Z = Mix

Bonus:

Dry stem: a dry stem is a stem for which you bypass all the “time-based” processors (Delay, reverb). This can be useful for remixes into other audio formats (surround, immersive) as you may need to apply different time-based process specific to these other formats and might not want to be “locked-in” to the sound of the reverbs and delays applied during the original mix.

Printing dry stems is time consuming and should be discussed with the artist/producer beforehand to make sure you are not “wasting” time printing audio file nobody really wants. Although there is something to be said about “over delivering” and also ensuring you have “every” possible option available for the future uses of the song.

Sample: usually refers to a rather short audio file meant to be “uploaded” into a Sampler and trigger by MIDI notes. It might be useful to “extract” samples from the Stems or Mix Session for future use in remixes or live show production. The musical director will thank you!

Notes

The recommended audio file format for audio production is Broadcast Wave File (.wav)

https://tech.ebu.ch/docs/tech/tech3285.pdf