OrynosMusicMixing & mastering guide
How to mix vocals with a beat online
Set the right stem roles, keep the beat intact and bring a lead vocal forward without making the whole instrumental pump.
Export stems from the same starting point
Export the instrumental, lead vocal and supporting parts from the same session start, even when a vocal enters later. Preserve the silence before that entrance. Files trimmed independently to their first sound can line up at the upload boundary while their performances land on the wrong beat.
Use the same sample rate for all stems and include the full tail of any printed effects. Listen to the files together before uploading. Avoid sending both a full beat and all of its individual instruments unless you intentionally want both, because that can double the backing and change the balance before any processing starts.
Tell the mixer what each file is
In Mix, upload your WAV stems and choose Check stem roles. Assign the instrumental to Music and the vocal files to their actual roles. File names can help a first guess, but a role you check yourself is more useful than expecting the engine to infer every production decision.
A double supports a lead; it is not a second competing lead by default. A genuine second performance has its own small timing and pitch differences. Making left and right copies of an identical recording does not create those performance differences. OrynosMusic can derive a stereo presentation from a general mono double, but that is an effect rather than a new take.
| Role | Use it for |
|---|---|
| Lead | The main vocal line that should carry the lyric. |
| Double | A supporting vocal, including a mono double needing a derived stereo presentation. |
| Double L / Double R | Separate supporting takes intended for the left and right sides. |
| Music | An already combined stereo instrumental or other musical backing. |
Further reading: iZotope: what double tracking does (opens in a new tab)
Get the dry vocal-to-beat balance right first
Choose your mix, select a model and start with Balanced. Galaxy is recommended for the Mixing workflow. Keep the initial comparison dry enough to hear the vocal clearly, then adjust Vocal balance if the lead needs to move forward or back. Open Mix details only when a specific problem needs more control.
Check a quiet line and a loud phrase. If individual words jump out, lowering the entire lead may hide the rest of the performance. Source-aware phrase levelling and controlled peaks can reduce that variation before you make a small final balance decision. Supporting takes should add size without making consonants blur or obscuring the main lyric.
Create space without ducking the whole beat
A pre-mixed instrumental may already contain compression, limiting, reverb and a deliberately wide image. Treat it as a finished relationship between instruments. Heavy full-band ducking can make the drums and atmosphere rise and fall with every vocal phrase even when only a narrow part of the spectrum is competing.
Selective dynamic correction can make room only in competing frequencies while the voice is active. OrynosMusic's contextual space stage targets qualified overlap in the Mid signal and preserves the stereo Side signal. Begin conservatively and listen to the beat between words as well as beneath the vocal. Reduce the amount if the backing loses body or movement.
Further reading: iZotope: when to use dynamic EQ (opens in a new tab)
Use pitch correction and effects for a reason
Pitch correction is optional. Enable the vocal's Autotune control when you want correction, then check the key and scale against the song. A wrong key can move an expressive note to the wrong destination. Listen to sustained notes, slides and consonants; a fast response is an audible style choice, not automatically a more accurate natural sound.
Add reverb and delay after the vocal level makes sense. Reverb can make a voice feel farther away, and a long delay can compete with the next lyric. Start with a small amount, check the busiest section and compare with the effect off. Use separate left and right doubles intentionally, and check that the lead still feels stable in mono.
Review the finished mix in context
The selected model finishes the combined mix, so listen to the render before downloading. Check the vocal in the chorus, the low end at moderate volume and the space at phrase endings. If a vocal is distorted in the source or the beat has already been heavily limited, go back to the available source rather than trying to hide every problem with more processing.
Change one thing at a time and compare at similar loudness. Keep notes such as 'the second verse needs less vocal' or 'the doubles cover the last word.' Those observations lead to a useful revision more directly than asking every control for more intensity. Save your downloaded WAV locally before the five-minute completed-result window expires.