Burnish Vocal Chain 1.0.3 can analyze either an acapella or full mix locally, then propose editable settings across eight vocal stages. It measures tonal balance, dynamics, sibilance, space, and tuning style before turning those observations into settings you can accept, change, or throw away. That last part matters.
The wider Vitric Audio Studio 4 plugin launch introduced Burnish as an eight-stage vocal processor, but the reference system deserves closer attention. It does not promise to copy another singer or reproduce a finished record. It uses the reference to make a first pass while keeping the actual chain exposed.
That makes local reference-driven vocal matching more useful than a simple preset hunt. You can see whether the suggestion wants more cleanup, compression, space, or a different tuning approach, then judge it against your take. Match Amount scales the fitted settings from 0 to 150 percent, letting you underplay or exaggerate the proposal without rebuilding the chain.
Each of the eight stations also has its own scope, so the proposed settings are not buried behind one global match score. That matters when the tonal fit looks sensible, but the de-essing or space estimate feels wrong. You can reject one stage without abandoning the whole reference fit.
The important limitation is right there in the design. Burnish says the goal is to help your vocal sit the way the reference sits, not make your vocal sound identical to it. Pitch movement, diction, microphone tone, room sound, and dynamics still differ, making editable proposals more credible than exact-duplication claims.
For you, that changes how aggressively the suggestions should be treated. If you have a clean acapella reference, the tonal and dynamics readings are based on a more direct vocal signal. With a full mix, the reduced confidence label is a reason to listen harder before pushing Match Amount or accepting every stage.
The analysis also stays on the machine. Vitric says the reference process runs offline, in the background, and fully locally, with no cloud account required for that analysis. That matters with unreleased material, client references, or rough mixes because the audio never needs to leave the local system.
Burnish goes a step beyond reference files with its factory chains. Its 66 factory chains carry measured target profiles that can adapt to the current voice instead of staying fixed. That gives you another route into the same basic idea, a starting condition shaped by measurements but still open to manual decisions.
That mode is aimed at recording through the chain, where delay in the headphones can become annoying fast. The Classic path is the hard-tune option, so you are trading for immediacy and a particular tuning character. Burnish keeps the tone, compression, color, doubling, space, riding, and limiting chain available while you record rather than forcing you into a dry monitoring workflow.
Mix mode allows the full-quality path, and the Natural tuning option reports 12.5 ms of latency. Natural is the formant-preserving correction path and includes the Throat control, so this is not merely the same algorithm running slower. You are choosing a different processing route when the session no longer depends on near-instant monitoring.
That distinction is easy to miss if you only look at the headline latency number. Four samples at 48 kHz is about 0.08 ms before rounding, while 12.5 ms can matter during live monitoring. Burnish exposes the current mode and reported latency in its header, so the tradeoff stays visible while you work.
The wider Vitric Audio Studio 4 plugin launch introduced Burnish as an eight-stage vocal processor, but the reference system deserves closer attention. It does not promise to copy another singer or reproduce a finished record. It uses the reference to make a first pass while keeping the actual chain exposed.
Reference analysis makes a first pass, not a clone
Burnish extracts a 40-band long-term tonal curve along with dynamics density, a sibilance profile, a space estimate, and a three-position tune style. Those measurements feed interpretable rules for the individual stages rather than disappearing into one mystery control. The plugin also shows confidence labels, so a proposed move comes with some indication of how much trust Burnish puts in the analysis.That makes local reference-driven vocal matching more useful than a simple preset hunt. You can see whether the suggestion wants more cleanup, compression, space, or a different tuning approach, then judge it against your take. Match Amount scales the fitted settings from 0 to 150 percent, letting you underplay or exaggerate the proposal without rebuilding the chain.
Each of the eight stations also has its own scope, so the proposed settings are not buried behind one global match score. That matters when the tonal fit looks sensible, but the de-essing or space estimate feels wrong. You can reject one stage without abandoning the whole reference fit.
The important limitation is right there in the design. Burnish says the goal is to help your vocal sit the way the reference sits, not make your vocal sound identical to it. Pitch movement, diction, microphone tone, room sound, and dynamics still differ, making editable proposals more credible than exact-duplication claims.
Acapellas give Burnish clearer evidence than full mixes
Burnish treats an acapella and a full commercial mix differently. An isolated vocal enters the reference analysis with full confidence, while a full mix is isolated and marked with reduced confidence. That honesty matters because the plugin doesn't treat a buried lead vocal as evidence equal the a stem.For you, that changes how aggressively the suggestions should be treated. If you have a clean acapella reference, the tonal and dynamics readings are based on a more direct vocal signal. With a full mix, the reduced confidence label is a reason to listen harder before pushing Match Amount or accepting every stage.
The analysis also stays on the machine. Vitric says the reference process runs offline, in the background, and fully locally, with no cloud account required for that analysis. That matters with unreleased material, client references, or rough mixes because the audio never needs to leave the local system.
Burnish goes a step beyond reference files with its factory chains. Its 66 factory chains carry measured target profiles that can adapt to the current voice instead of staying fixed. That gives you another route into the same basic idea, a starting condition shaped by measurements but still open to manual decisions.
Track and Mix modes solve different vocal problems
Reference matching is only useful if the chain still works in the part of the session where you need it. Burnish separates low-latency tracking from the heavier Mix path instead of pretending one configuration is ideal for both. Track mode uses zero-lookahead dynamics and the Classic tuning path with four samples of total latency at 48 kHz, reported as 0.1 ms.That mode is aimed at recording through the chain, where delay in the headphones can become annoying fast. The Classic path is the hard-tune option, so you are trading for immediacy and a particular tuning character. Burnish keeps the tone, compression, color, doubling, space, riding, and limiting chain available while you record rather than forcing you into a dry monitoring workflow.
Mix mode allows the full-quality path, and the Natural tuning option reports 12.5 ms of latency. Natural is the formant-preserving correction path and includes the Throat control, so this is not merely the same algorithm running slower. You are choosing a different processing route when the session no longer depends on near-instant monitoring.
That distinction is easy to miss if you only look at the headline latency number. Four samples at 48 kHz is about 0.08 ms before rounding, while 12.5 ms can matter during live monitoring. Burnish exposes the current mode and reported latency in its header, so the tradeoff stays visible while you work.
