> ## Content Index > Fetch the complete content index at: https://izotoperx.net/llms.txt > Use this file to discover other available public pages before exploring further. # How to Use iZotope RX Loudness Optimize - URL: https://izotoperx.net/izotope-rx-loudness-optimize/ - Published: 2026-09-30T21:20:55.000Z - Updated: 2026-10-01T17:21:19.000Z - Description: Loudness Optimize lifts quiet material that falls below the relative LUFS gate so it counts toward integrated loudness. Learn from the whole file, compare playback, then set delivery LUFS and true peak separately with Loudness Control. - Author: Brandon Hayes - Tags: Workflow, Post Production **RX Loudness Optimize can help a track play louder after streaming normalization by raising quiet material that falls below the integrated-loudness gate.** It uses upward compression, so the benefit depends on what is actually quiet in the file. A soft musical passage may be worth lifting; a pause or room tail may be better left alone. The module does not set a delivery LUFS target or true-peak ceiling. In RX 12, Loudness Optimize is available in Standard and Advanced. Its [official guide explains how the gate, controls and Learn function work](https://docs.izotope.com/rx12/en/loudness-optimize.html?ref=izotoperx.net). Start by reading the trace, then decide whether raising the excluded material improves the sound. ## Read the gate and Measured % first Open Loudness Optimize with the full file selected. The loudness trace plots momentary loudness over time: blue sections sit above the dashed white LUFS gate and contribute to the integrated measurement; gray sections sit below it and do not. The gate moves with the material: RX places it 10 LU below the average momentary loudness, and you cannot set it to a chosen LUFS value. ![Loudness trace with blue measured sections above the dashed LUFS gate and gray unmeasured sections below it, next to Measured % and Integrated LUFS readouts](https://storage.ghost.io/c/2e/c1/2ec1682b-08cc-4997-b0d0-7361fc45beb8/content/images/2026/09/rx-loudness-optimize-gate-trace.webp) Blue sections count toward integrated loudness; gray sections below the gate do not. **Measured %** reports how much of the file sits above the gate. A low value suggests that some quiet content could be brought into the calculation. Near 100%, there is usually less room for this method to lower the integrated reading. The percentage does not judge the mix. Look at the gray passages and ask whether they contain music or speech that ought to be more audible. **Integrated LUFS** is the gated measurement for the file, not the level of each phrase. Oddly, lifting quiet material can lower that reading because the material now counts in the average. A streaming service may then apply more normalization gain. The point is to improve playback without upsetting the balance, so compare the sound as well as the number. ## What the four controls change **Boost** caps the gain applied to quiet content. **Threshold** determines where upward compression begins in dBFS; material above that threshold remains unaffected by this processor. **Ratio** sets how strongly below-threshold content is raised, subject to the Boost cap. **Speed** controls how quickly the boost begins and also sets lookahead: the manual notes that longer speeds are less prone to distortion. Those are processing controls, not a streaming-platform target. ![Loudness Optimize control map with Boost, Threshold, Ratio at 4:1, Speed at 120 ms and the Learn button](https://storage.ghost.io/c/2e/c1/2ec1682b-08cc-4997-b0d0-7361fc45beb8/content/images/2026/09/rx-loudness-optimize-controls.webp) The documented 4:1 Ratio and 120 ms Speed are starting points, not delivery targets. More gain is not automatically better. Upward compression can bring ambience forward along with a quiet voice or instrument. A faster Speed or stronger Ratio may change the measurement more, but can also change the feel of a transition. Keep the original balance in mind while you audition the result. ## Use Learn on the whole file, even for a partial render Choose Ratio and Speed first. RX’s default 4:1 and 120 ms are a documented starting point. Select the entire file and click **Learn**. It tries combinations of Boost and Threshold under your chosen Ratio and Speed, then selects the pair that gives the best reduction it can find in integrated loudness; it does not change Ratio or Speed. The [full-file Learn recommendation](https://docs.izotope.com/rx12/en/loudness-optimize.html?ref=izotoperx.net) still applies when you plan to process only one section: learn from the whole file, then select and render the intended section. ![Three-stage workflow: select the full file, run Learn with Ratio 4:1 and Speed 120 ms, then render only the chosen region](https://storage.ghost.io/c/2e/c1/2ec1682b-08cc-4997-b0d0-7361fc45beb8/content/images/2026/09/rx-loudness-optimize-learn-workflow.webp) Learn from the whole file; a smaller selection can then be rendered with the learned settings. Preview before rendering. Compare the transition into and out of the quiet section, especially if it contains reverb, background noise or a delicate ending. If the change helps, render the intended selection and check the finished file’s loudness and peaks. A better number alone is a poor reason to keep a worse sounding passage. Learn may report that no Boost and Threshold combination reduces integrated loudness at the selected Ratio and Speed. Check Measured % and the trace again. If most of the file already counts toward the measurement, the module may have little to offer. Leave the file alone unless a different dynamic choice also sounds better. ## Check playback; set delivery targets separately Optimize is useful when the relative gate leaves meaningful quiet content out of the measurement. It does not replace a specification check. If the destination asks for a defined integrated LUFS value and a true-peak maximum, use [Loudness Control for the final target and peak ceiling](https://izotoperx.net/izotope-rx-loudness-control/) after the balance is settled. That module uses fixed gain across the selection and a post-limiter when needed to satisfy the true-peak setting, as described in the [RX 12 Loudness Control documentation](https://docs.izotope.com/rx12/en/loudness-control.html?ref=izotoperx.net). For a streaming-oriented decision, [Streaming Preview simulates normalization and encoding with platform presets](https://docs.izotope.com/rx12/en/streaming-preview.html?ref=izotoperx.net). Use it to compare playback, allowing for platform settings that can change. Do not export the file after applying Streaming Preview: its documentation says that would bake the preview encoding into your audio. If the real problem is uneven dialogue from line to line, start with the [dialogue-cleanup workflow](https://izotoperx.net/clean-dialogue-rx/); the integrated reading cannot tell you which phrase needs a level ride.