Define two checkpoints, not one green badge
This workflow is for a producer or audio editor with a finished lossless master. Mixing, restoration and broad loudness correction belong in the related mastering workflow. Keep the approved source and record the receiving system's actual brief. Specify input peak and decoded-delivery peak separately, together with loudness and format expectations. Do not inherit platform requirements from preset names.
Probe the source, then measure it. A file extension is not proof of codec, sample rate or channel layout. Silence, very short material or failed analysis can leave readings missing; those are unverified cases, not clean passes. Save the meter method and rounding policy so a later reviewer can interpret small differences between reports.
Compare the codec result without confusing stages
The preview tool encodes selected variants, decodes them to floating-point WAV internally and measures that reconstruction. Its package supplies encoded previews and reports; do not assume it includes the temporary decoded analysis files. Read the actual decoded peak, headroom and loudness shift. Its fixed grading thresholds do not replace your written acceptance limits.
Audition dense transients, quiet tails and stereo detail at a comparable listening level. A louder candidate is not automatically a better one, and a safe numerical peak does not prove the absence of audible codec artifacts. These previews test local encoders and selected bitrates, not a streaming provider's current encoder, normalization or future re-encoding.
Reserve headroom without hiding the cost
Leave an already acceptable master alone. If revision is justified, retain the original and create a separate lossless candidate. The true-peak limiter runs bounded attempts with a verification tolerance; a limited status or output file does not by itself prove the agreed ceiling was achieved. Inspect the final measured peak, including missing values and rounding, and independently remeasure questionable results.
Record the loudness change and listen for softened transients or pumping. Keep the approved sample rate before limiting where possible; resampling or channel changes afterward can alter the result. Repeat the same codec previews on the revised source. If additional revision is needed, return to an approved lossless source rather than accumulating lossy generations. Choose WAV output when continuing through the three WAV-input converters.
Verify the exact export, not a similar preview
Pick the MP3, AAC or Opus branch required by the brief and record explicit settings. The preview may have different sample-rate or channel behavior from the final converter. A later conversion therefore needs its own probe, loudness measurement and decoded true-peak check; do not re-encode a lossy delivery through the preview tool as a substitute for measuring that file.
The peak meter's clipping flag is relative to its configured ceiling, not exclusively to full scale. Use actual values and the agreed limit; a tighter-than-zero threshold may flag an otherwise non-overloaded reconstruction. Do not silently relax a limit to make the label green.
Open the exact final file in the target player, check the opening, ending and expected channel behavior, and save the source revision, settings, reports and listening exceptions. If a service later transcodes it, review that served result separately. Album sequencing and package metadata belong in the related album workflow; this page does not publish a release or confer platform approval.