Stems are an essential part of a final delivery package in post-production. They are a selection of grouped exports of all the main elements that make up a mix, such as VO, Dialogue, SFX, Atmos, Music, and more. They are typically delivered as stereo tracks - unless you're working in a multichannel format such as 5.1 or atmos.
Split the Difference
‘Splits’ is another term that gets used for more or less the same purpose – although some may disagree with this exact terminology.
Another term that occasionally gets used – although more in music than in post - is ‘multitracks’. These are generally understood to mean the individual track exports of all the elements of a mix e.g. Dial 1, Dial 2 etc. rather than the summed groups.
Why do we need them?
If you are a production company or editor working on a project, it’s always good to make sure you get the final mix stems from the mixer/sound designer. The main reason for this is that if there is a small amend needed, like a change to a VO line, removing an SFX or similar, it’s easy for the editor to do this without needing to go back to the mixer.
You are therefore buying yourself some flexibility and assurance that you’re going to be able to handle any last-minute issues or client amends that may arise. This can save everyone a lot of headaches, particularly when deadlines are looming!
M&E
Stems serve another valuable purpose. This is when you are delivering various audio versions for a project, such as an M&E version. M&E stands for Music & Effects, meaning any dialogue and voice-over elements are excluded. The primary use case for this is when clients require alternate language versions of a film.
Additionally, it is common to encounter other combinations of exports, like DME (Dialogue, Music & Effects). This allows for the seamless replacement of the original VO with a recorded version in another language, minimising any complications involved in the process.
To Dip, or Not to Dip…
One point to note is that there may be two different music stems included in a package. The first would be ducked/dipped which means the volume has been automated up and down to allow more space for the VO and possibly other elements. The other option is undipped meaning the music volume is left flat. Undipped will provide greater flexibility and allow for the music to be dipped manually at a later date. With a re-recorded VO for example.
A word of caution
One thing that is really important to emphasise and was a key part in my thinking behind this blog post, is that if the editor does do a final mix from the stems, they must ensure that a limiter is added to the master output channel at -1dB. Maybe even at -1.5/-2dB to be on the safe side.
The reason for this is that when the stems are exported, they will have been mixed through a limiter that will catch any peaks. However, when the stems are summed back together again, the build-up of amplitude from the stems playing on top of each other could nudge the master output of 0dB. Going over 0dB in the digital audio world is not good and can cause nasty distortion/clipping and is definitely to be avoided.
Similarly, it is likely the stems have been mixed to reach the desired loudness (LUFS) target. It would therefore be sensible not to adjust the volume of the stems too much. Particularly the dialogue track.
Extra, extra…
A final thought, remember that you can have a stem for anything! Therefore, if there is a specific element of a mix you want on its own, that’s absolutely doable. For example, you may want a separate stem of diegetic music, or for a specific dialogue track.
In summary
So, there you have it. A simple topic, but a good one to have a grasp of if you are working across post-production. Delivering projects is stressful enough and stems can be a lifesaver, so make sure you always ask for them. You’ll sleep better for it..!
コメント