A whole episode can meet a loudness target while one guest remains difficult to hear. The final number describes a program; it does not prove that each speaker is balanced. Work on the relationship between voices before the last delivery check.

Choose a representative exchange

Find a passage where the host and guest speak normally for several turns. Avoid using only a laugh or a quiet introduction as the reference. Listen without changing the headphone volume between speakers.

Adjust the quieter voice or the relevant regions until the exchange is comfortable to follow. Keep enough space below clipping when tracks play together. A louder background hiss after a gain increase may reveal a source problem that should be handled separately.

Do not copy the same gain number to both tracks and expect the result to match. Different microphones and speaking distances create different starting levels.

Check the difficult moments

Review a soft answer, an excited answer and an overlap. If the two voices are balanced during normal speech but a laugh jumps out sharply, make a focused adjustment rather than changing the whole track again.

After the edit and voice balance work, assess the complete episode against the delivery destination's current audio guidance. Keep peak level and perceived loudness as separate checks. A peak adjustment alone does not make every conversation equally easy to hear.

Export a sample and test it on ordinary listening devices at a comfortable volume. If you repeatedly reach for the volume control between speakers, return to the exchange and improve the balance before calling the episode finished.

Explore the related Podbento workflow.