Recording Compression — Operator Guide

Applicable from VRS 14.12.0. Compression for mixed call recordings is configured in docker/config.env.

Background

Previously, mixed call recordings used a fixed compression of 4000 Hz and 16-bit. That gave about 50% size reduction compared with an uncompressed mix, but it was not configurable. Changing the level required a code change.

Compression is now controlled from docker/config.env. You can keep the original ~50% behavior, use a stronger PCM option (~75%), or use GSM (~90%). Only new recordings after a mixer restart use the new settings. Older files stay in whatever format they were mixed with.

Important rules

Set options in docker/config.env.
After editing, recreate/restart the mixer so the new values apply.
Only one profile is active at a time.
If OUTPUT_CODEC=gsm is set → GSM is used. If OUTPUT_CODEC=gsm is commented → PCM compression from COMPRESSION_* is used.
For PCM profiles (~50% / ~75%), OUTPUT_CODEC=gsm must be commented.
For GSM (~90%), COMPRESSION_SAMPLE_RATE and COMPRESSION_BIT_DEPTH do not affect the mix (they are ignored).

Profile 1 — about 50% compression (previous default)

Use this when you want the same behavior as before, but configured in env.

COMPRESSION_ENABLED=true
COMPRESSION_SAMPLE_RATE=4000
COMPRESSION_BIT_DEPTH=16
#OUTPUT_CODEC=gsm

Setting	What to do
`COMPRESSION_ENABLED`	Set to `true`
`COMPRESSION_SAMPLE_RATE`	Set to `4000`
`COMPRESSION_BIT_DEPTH`	Set to `16`
`OUTPUT_CODEC=gsm`	Comment out (`#OUTPUT_CODEC=gsm`)

Profile 2 — about 75% compression (stronger PCM)

Use this when you need more compression, but still want PCM (not GSM).

COMPRESSION_ENABLED=true
COMPRESSION_SAMPLE_RATE=4000
COMPRESSION_BIT_DEPTH=8
#OUTPUT_CODEC=gsm

Setting	What to do
`COMPRESSION_ENABLED`	Set to `true`
`COMPRESSION_SAMPLE_RATE`	Set to `4000`
`COMPRESSION_BIT_DEPTH`	Set to `8`
`OUTPUT_CODEC=gsm`	Comment out (`#OUTPUT_CODEC=gsm`)

Profile 3 — about 90% compression (GSM)

Use this for the strongest size reduction (telephony-style GSM storage).

OUTPUT_CODEC=gsm
COMPRESSION_ENABLED=false

Setting	What to do
`OUTPUT_CODEC=gsm`	Uncomment / set (required)
`COMPRESSION_ENABLED`	Recommended `false` (see note below)
`COMPRESSION_SAMPLE_RATE`	No effect — may leave as-is
`COMPRESSION_BIT_DEPTH`	No effect — may leave as-is

Is COMPRESSION_ENABLED=false required for Profile 3?

No — not required for GSM to work.

When OUTPUT_CODEC=gsm is set, the mixer uses the GSM path and ignores compression rate/bit-depth settings.

Setting COMPRESSION_ENABLED=false is still recommended so the config is clear: it shows that PCM compression is not the active mode, and avoids confusion if someone later comments out OUTPUT_CODEC=gsm and accidentally leaves compression enabled.

Quick reference

Goal	Active settings	Comment OUTPUT_CODEC=gsm?	*Do COMPRESSION_ matter?**	Approx. savings
~50%	4000 Hz / 16-bit, compression on	Yes	Yes	~50%
~75%	4000 Hz / 8-bit, compression on	Yes	Yes	~75%
~90%	`OUTPUT_CODEC=gsm`	No (keep enabled)	No (ignored)	~90%

Apply a profile (checklist)

Edit docker/config.env using one profile above.
Confirm GSM is commented for ~50% / ~75%, or enabled for ~90%.
Recreate/restart the mixer.
Verify env inside the container (optional): docker exec mixer printenv | grep -E 'OUTPUT_CODEC|COMPRESSION'
Make a new test call and confirm size/quality.
Confirm UI play still works; for GSM, confirm download size stays small.