Loudness & dynamics: choose the right tool

Describe the problem before choosing a processor

Loop the passage that bothers you and write a precise observation. Perhaps the final words of a sentence disappear, a snare hit dominates the bus, or a vocal becomes too forward in the chorus. “Make it louder” is too broad to tell you which control to reach for. Decide whether the issue happens within a phrase, between sections, or only at a different listening volume.

Check the source first. A muted clip, a sudden edit, an unintended automation move, or an overly dense arrangement can resemble a dynamics problem. Correct those causes before asking a processor to compensate. Keep an unprocessed version available so you can return to the original behavior during the comparison.

Understand the measurements and the listening goal

A sustained pad and a short drum hit can reach similar peaks while feeling very different in level. Look at meters to understand the signal, then listen to its role in the arrangement. Delivery specifications apply to the finished deliverable; they are not a reason to force every individual instrument to the same loudness target.

Scroll across the table to compare every column on a small screen.

Match the tool to the task

APU describes target-based leveling and equal-loudness compensation as distinct workflows. Ayaic provides WRITE and READ automation modes. Antelope identifies Opto-2A as an optical compressor/limiter and COMP-4K-BUS as a stereo bus compressor. These categories guide a useful audition; they do not establish that one product is the right purchase for every source.

Scroll across the table to compare every column on a small screen.

Product references: APU Loudness Leveler · Ayaic Dragon Fader · APU TrueGain · Antelope Opto-2A · Antelope COMP-4K-BUS

Prepare a comparison you can trust

1. Choose a representative passage

   Use 20–30 seconds containing a quiet phrase, an ordinary phrase, and a louder moment. Include a pause or a sustained tail. A single loud word will not tell you how a processor behaves across a performance.

2. Save the current balance

   Keep a copy of the track or a clearly labeled unprocessed version. Check its fader and automation. Start with enough headroom that you can compare processing without unintentionally clipping the output or another stage.

3. Keep monitoring consistent

   Use a comfortable listening level and keep the monitoring controls steady for the comparison. If you change the listening level, note that change and repeat the check. A louder playback can influence a judgment even when the musical balance has not improved.

4. Choose one processor and one goal

   Begin with the simplest chain that can address your observation. Do not add a leveler, compressor, limiter, and tonal compensation at once. You need to know which stage caused the change you are hearing.

Practice: make a vocal phrase easier to follow

1. Listen for the words that disappear

   Mark two or three moments that need attention. Listen in the mix as well as in solo. A phrase may be understandable on its own but masked by another instrument when the arrangement returns.

2. Make an initial manual correction

   Use clip gain or a simple automation move for an obvious large difference. Preserve the intended emphasis of the performance. This gives the processor a more manageable input and helps you recognize whether it is solving the remaining problem.

3. Start a leveler near the source’s current level

   For a practice audition, choose a target close to the ordinary passage rather than demanding a large increase. Adjust the tolerance and response so the important words are easier to follow. Treat this as a workflow exercise, not a release loudness specification.

4. Listen to pauses and quiet details

   Check whether breaths, room sound, and tails rise too far. Reduce correction limits or change the response if the processor is doing more than the musical task requires. A steadier meter alone is not a sufficient reason to keep a setting.

5. Compare with bypass at a similar level

   Adjust the processed output so the ordinary passage is close in perceived level to bypass. Switch a few times, with short pauses. Ask whether the words are more consistent without losing the phrasing you liked.

6. Check the loudest part of the song

   Move beyond the original loop. Listen to the chorus, a strong consonant, and the end of a sustained phrase. A setting that helps the verse may overreact elsewhere. Revise it before deciding the tool belongs across the entire performance.

7. Save a result and a reason

   Save the setting with a short note such as “verse words remain clear; breaths stay natural.” Keep the original comparison. If you cannot describe the improvement in the mix, return to the simpler version and reassess the arrangement.

Use compression for a dynamics decision

Compression changes the relationship between louder and quieter parts of a signal. The timing of that change affects how attacks and sustained sounds feel. On a processor with adjustable timing, compare a small change in attack or release while keeping the other controls steady. Listen to the rhythm and envelope rather than choosing a setting only because a meter shows more gain reduction.

With Opto-2A, the visible Gain and Peak Reduction controls offer a simpler operating approach; the manufacturer describes a program-dependent response. Do not look for the same adjustable attack and release controls on every compressor. Start with modest reduction, compare at a similar output level, and check that the result supports the performance.

For a stereo bus, listen to the whole group. Notice whether the kick makes the rest of the bus move too much, whether the groove becomes less distinct, or whether the balance improves. Avoid using a bus compressor to hide a single badly balanced track. Correct the individual source first, then repeat the bus comparison.

Product references: Antelope Opto-2A · Antelope COMP-4K-BUS

Choose automation when you want editable moves

Automation is useful when a specific word, note, or section needs a deliberate level decision. A visible curve lets you return to that moment and change the move. Before recording generated fader automation, save the existing curve and confirm which parameter or lane the host will write.

Dragon Fader’s publisher describes WRITE and READ modes for this kind of workflow. Test a short passage before applying it to the song. Play back the recorded moves, check a boundary between sections, and edit anything that draws attention. Follow the current publisher and host instructions for automation recording, since writing to the wrong lane can replace a move you intended to keep.

Product references: Ayaic Dragon Fader

Treat listening-level compensation as its own task

TrueGain links gain changes with equal-loudness frequency compensation. Its calibration connects the model to a monitoring setup. The publisher’s Translation Check mode approximates level-dependent hearing; it is not a recreation of another room, speaker, or headphone.

Begin with the publisher’s calibration instructions and a monitoring setup you understand. Save a normal reference, then evaluate a small compensated change. Ask whether the model helps you judge balance at the intended listening level. Do not use the compensation result as proof that a mix will translate to every playback system; ordinary reference listening remains part of that decision.

Product references: APU TrueGain

When the result sounds worse

Scroll across the table to compare every column on a small screen.

Finish the audition before buying

Work through the list here, or print the guide for your next session.

• I can describe the source problem in one sentence.
• I have compared processed and bypassed audio at a similar perceived level.
• The setting works on quiet phrases, loud moments, pauses, and the full arrangement.
• I understand whether I need leveling, automation, compression, or compensation.
• I have checked the exact plugin format, license platform, and authorization requirements.
• I have saved the original and can explain why I prefer the processed result.