I would agree what each stem contains and whether the stems are expected to reconstruct the approved mix. Export them with a common start and duration, then import them into a fresh session for a test. Separate files are not a usable stem package if their timing or processing assumptions are unclear.
On this page
Our recommendation
I favour stems with explicit content and processing definitions, common timing and a tested reference relationship. Preserve the approved mix alongside them so the intended result remains clear.
Define the package
List dialogue, music, effects or other required groups, with channel layout, sample rate and bit depth. State whether master-bus processing is included, excluded or supplied separately. Independently normalising every stem can change their intended balance, so do not do it merely to make each file look equally loud.
Use the same programme start and end, including intended silence, for files meant to line up automatically. Keep useful names and a manifest. A dialogue stem beginning at its first spoken word may not align with music beginning at programme zero unless that relationship is explicitly recorded.
Test the reconstruction
Import the stems into a clean session with the documented settings, align them at the common start and listen to the sum. Compare with the approved reference mix at matched level. Check that a source has not been included twice or omitted from every group.
Remember that nonlinear master processing can prevent a simple sum of separately processed stems from matching the full mix exactly. Agree the expected outcome and explain the processing rather than claiming a perfect reconstruction that was never tested.
A duplicate-source check
If identical dialogue is accidentally present in two stems and both are summed at unity, coherent duplication can raise that signal by about 6 dB. The package may clip or sound unbalanced even though each file plays correctly on its own. A clean-session recombination test exposes the duplication before the recipient mixes it.
Before you act
- Agree stem contents and master-processing treatment.
- Use common start, duration and channel specifications.
- Reimport and compare the sum with the reference.
Use the Audio file size calculator to apply the guide to your own setup.
Keep exploring
- Distinguish dual mono, stereo and a missing dialogue channel
- Check loudness gain against the true-peak limit before rendering
Method note: this is an original planning procedure with illustrative inputs. It does not report a physical product test. Check the exact product, edition and official instructions before acting.
Build on what you just learned
Your next useful step.
Check channels, loudness and peak headroom.
Resolve 21 subframe audio edits: solve the click without moving the picture
Use fine audio timing when a frame-sized move is too coarse for a click, fade or transient.
Read the guide →
Clean dialogue without making the room pump between words
I would begin with a clean recording and restrained processing, then compare at matched listening level.
Read the guide →