Gain and Levels

Gain Staging Helper

Gain staging is the least glamorous part of mixing and the one that decides whether the rest of the session goes well. Two failure modes hide behind it. The first is a chain that quietly stacks six decibels of trim, so the channel is already at zero dBFS before the bus compressor sees it, which is why the mix sounds glued and cannot be automated around. The second is a chain that is too conservative, so every source arrives at the bus far too quiet and the master compressor ends up doing the work of a leveler. This helper walks a chain stage by stage, shows the cumulative level after each plugin as a vertical meter, compares the result against the target window for the instrument you are tracking, and works out how much trim a parallel or NY style branch needs before it starts pushing the bus over the ceiling.

Instrument or bus

Sets the target window the chain is judged against

Input level at the first plugin

dBFS

The level arriving at the top of the chain

Plugins in the chain

How many processing stages sit between the source and the bus

Trim per stage

Gain added or removed at each active plugin

Plugins switched off

Counted as bypassed at the end of the chain

Parallel split

How far the signal is trimmed when it is copied into a parallel branch

Result

How to use it

STEP 1

How to use this helper: start with the level arriving at the first plugin, not the level you think the source should have. That number decides everything downstream, and it is the one people guess at. Then count the plugins in the chain and the trim on each stage. If every stage is trimmed the same way, enter that value once. The table then shows the level after every single plugin, which is where you find the stage that is quietly running hot.

STEP 2

Next set the number of bypassed plugins. These are counted as switched off at the end of the chain, which is the usual situation when you are comparing a chain with and without its last stage. The chain length row tells you how many stages are actually contributing, and the stages adding gain row tells you how much of the chain is purely makeup. If every active stage adds gain and there are three or more of them, that is the finding to act on.

STEP 3

Read the target window for the instrument you selected. Kick, snare, bass, vocal, guitar, stem bus and master each get their own working window, and the verdict row compares the end of the chain against it. A final level above the window is not a disaster, but it means the next stage has less headroom than you think and the bus compressor will be doing gain reduction that you cannot automate around.

STEP 4

Finally enter the parallel split. This is how far the signal is trimmed when it is copied into a second path. The branch output, the wet level relative to the dry and the recommended branch trim are calculated from it, along with a power sum estimate of the dry and wet together. In an NY style compressor the parallel path is the control path, so the recommended trim is the makeup gain you need on that compressor.

Three common mistakes

Common mistake: counting the fader move as part of the chain. A fader does not change what the plugins see, it only changes what the bus receives. If you have already pulled the fader down three decibels and you also enter three decibels of trim inside the chain, the tool is right and your headroom estimate is wrong. Enter the level at the top of the chain, let the helper tell you where the chain ends, and then decide what the fader is for.
Watch out for identical trim on every stage. If you type two decibels per stage and the chain has three active plugins, the model stacks six decibels and flags it. Real chains rarely have identical trims, which is exactly why this is a mistake worth catching. A chain that needs different trim on each stage usually means someone reached for the plugin gain control to solve a level problem that belongs on the fader.
Key point that gets missed: the summed dry and wet path can clip even when the dry path alone is fine. Two uncorrelated sources of the same level sum to about three decibels louder, so a branch that is only four decibels under the dry will push the bus over the ceiling when it comes back in. The helper shows that estimate separately from the dry level because the two numbers need different actions: one is fixed with trim, the other with automation.
Easy to get wrong: treating the target windows as rules for the final master. A kick sitting at minus twelve dBFS is a working level for a drum bus in a dense arrangement, not a rule for how loud the finished track should be. The windows exist to keep headroom predictable and to stop the bus compressor from becoming a leveler. Loudness is decided at the end, on the master bus, with a limiter and a meter.

Why stacked trim ruins a mix before it ruins a level

Every stage that adds gain is a stage that reduces the headroom left for the stage after it. Three plugins trimmed up by two decibels each is six decibels of gain added inside one channel, which sounds like an innocent convenience and behaves like a hidden gain stage. The problems appear downstream: the next plugin receives a hotter signal, the bus compressor sees more level than the arrangement needs, and the automated parts of the mix no longer sit where you left them.

More than +6 dB of stacked gain inside a single channel is flagged, because that is where it starts to fight the bus compressor.
Between +4 and +6 dB is workable but tight, especially with a limiter already on the bus.
Mixing hot and then automating the fader down is not the same as staging properly, because the plugins downstream of the hot point behave differently.
A fader move after the chain is free. A gain trim inside the chain is not, and it is the one that changes how the plugins sound.

Target windows, and what they are really for

The windows in this helper are pre-master working levels, the level each source should be delivering into the next stage of the mix. A kick around minus eighteen to minus twelve dBFS leaves room for a bus compressor to work without crushing it. A bass a little lower, a vocal a little higher, and a stem bus somewhere in between. They are not loudness targets and they are not rules about how the finished track should sound on a streaming platform.

The window tells you whether the bus compressor is doing level matching or actual mixing.
A channel far below the window means the bus has to make up the difference, which drags the whole mix down and lifts the noise floor of that channel.
A channel far above the window means the next stage has no room left and any quiet passage on that instrument will not survive the bus compression.
Consistency across the arrangement matters more than the exact number. A mix with every source inside two decibels of its window will translate far better than one with a perfect kick and a chaotic rest of the arrangement.

Parallel branches, NY compression and the +3 dB rule

A parallel path copies the signal, usually attenuated, into a second chain. The dry signal stays untouched while the branch adds weight, glue or a different flavour. The one thing you cannot ignore is that the two paths add up. Two uncorrelated signals at the same level sum to roughly three decibels more than either one alone, and the helper uses that power sum to show where the bus ends up when the branch comes back in.

For a transparent blend, the wet path wants to sit about 6 dB under the dry. The recommended trim is calculated from your split and your stage trim to land there.
Between 3 and 6 dB under the dry, the branch becomes part of the sound rather than a safety net, which is sometimes exactly what you want.
In an NY style compressor the parallel branch is the control path that drives the gain reduction on the original. There the calculated trim is the makeup gain on that compressor, and the dry plus wet estimate still tells you what the bus receives.
Automate the branch return rather than riding the main fader, so the level relationship survives the arrangement.

FAQ

How do I decide whether a chain is genuinely too hot?

Look at two numbers rather than one. The final level tells you where the chain ends relative to the target window, and the headroom row tells you how much room is left before zero dBFS. A chain that ends inside the window but with only two decibels of headroom is tight but workable, especially with a bus compressor after it. A chain that ends outside the window is doing something you did not intend, usually because the trim was set to solve a level problem that belonged on the fader. In most mixes the answer is to take trim out of the chain rather than to lower the fader further.

Should every plugin in a chain really have the same trim?

No, and the fact that this helper assumes equal trim is a feature of the model, not a rule for your mix. It is there to make a hidden stack of gain visible. When you look at the table and see stage two sitting six decibels above stage one for no musical reason, that is the finding. Different plugins legitimately want different input levels, and a compressor wants a hot signal while an EQ or a reverb usually wants a moderate one. Use the table to find the stage that is out of line with the rest, then correct that one.

What is the difference between this and just watching the channel fader?

A fader shows you the output of the chain, and it does not tell you what any plugin inside the chain is receiving. If a compressor is set to three decibels of gain reduction and the makeup gain on the plugin is twelve decibels, the fader can be in a comfortable place while the compressor is being driven far harder than intended. The stage table here shows the level at every point in the chain, so you can see whether a plugin is being fed a workable signal. That is the number that determines how the plugin sounds.

Why is the parallel branch recommended six decibels under the dry?

Because that is where the branch stops being audible on its own and starts acting as glue. At six decibels under, the branch mostly thickens the low end and evens out the transients without adding its own voice. Closer than that, within about three decibels, and you start hearing the branch as a separate layer, which is fine when that is the intention and distracting when it is not. Much further under and the branch does nothing worth the routing. The helper still shows the numbers for any setting, so if you want the branch louder, take the recommendation as a starting point rather than a rule.

Does this tool make my mix loud enough for streaming platforms?

No, and it is not built to. Loudness lives on the master bus, after the mix is finished, and it is a mastering decision driven by a limiter, a target integrated level in LUFS and the density of the arrangement itself. The helper works entirely in pre-master working levels, because that is where headroom problems start. Once the mix is balanced and staged, the master chain handles the final level, and doing that too early is exactly what this tool is designed to help you avoid.

Limits of this tool

Limits: the model assumes one identical trim per stage and treats bypassed plugins as switched off at the end of the chain. Real chains with different per plugin trim are only approximated.
Not for: predicting what a specific plugin will do, or for setting a compressor ratio. This is a level and headroom tool, not a dynamics tool.
When this stops: the parallel estimate assumes two uncorrelated sources. A branch that is pitch shifted, time stretched or very similar to the dry will sum differently and the peak estimate becomes unreliable.
Scope: target windows are pre-master working levels for a dense club arrangement. A sparse acoustic record or a film score will want different numbers.