To master one finished stereo track in Sound Forge Pro 2026, work through five checkpoints: measure the untouched mix, correct broad tonal balance, control dynamics with Wave Hammer, verify loudness and peaks, then save and reopen the exported file. Keep every comparison level-matched, and if a processor doesn't solve a problem you can hear, remove it.
This page is the short execution workflow. The full Sound Forge mastering guide covers album sequencing, multiple deliverables and wider platform decisions. Here the job is narrower: finish one stereo file without relying on a preset or a universal loudness number.
Quick answer: run Tools → Statistics, make small gain-matched corrections in Modern Equalizer, use the compressor and Volume Maximizer inside Wave Hammer as two separate decisions, rescan the complete file, convert bit depth only when the delivery format requires it, then reopen and check the saved master.
The five-step Sound Forge mastering workflow
- Measure the untouched source. Save the whole-track, both-channel Statistics baseline before processing.
- Balance the spectrum. Use Modern Equalizer only for problems that survive a level-matched bypass check.
- Control dynamics. Set Wave Hammer compression first, then decide how much final level the Volume Maximizer can add cleanly.
- Verify the result. Save and reopen the processed file with no extra chain applied, then rescan loudness and sample/true peaks. Make any final integer conversion after the remaining processing.
- Export and reopen. Inspect and listen to the file that will actually be delivered.
In the decision map below, the Wave Hammer box contains its two internal processors; the bit-depth branch is conditional rather than a required mastering effect.
Before processing: protect the source and know when to stop
Start from the finished stereo mix at its native sample rate and bit depth, and save a copy before touching the waveform. This stereo-file chain changes the combined mix. It does not give you the original vocal, snare and bass faders back; when the multitrack session is available, I'd fix those balances there.
Return to the mix when any of these problems is obvious:
- the source already clips or has audible distortion;
- one instrument or vocal needs a level change that affects only that element;
- the stereo bus is already limited so heavily that transients no longer recover;
- a narrow resonance moves with the melody and a static master-bus cut damages other notes;
- the requested loudness needs more limiting than the mix can take cleanly.
A mastering chain should improve a finished mix. It shouldn't stand in for a mix revision that's still possible.
Step 1: Measure the untouched file with Statistics
Select the complete track and both stereo channels, then choose Tools → Statistics. The current Sound Forge 2026 Statistics help documents the values below. Use Copy to clipboard to keep a baseline report with the filename and selection length. The Statistics guide explains the scan fields if you need to inspect a suspicious value.
| Readout | What it tells you | How to use it |
|---|---|---|
| Minimum and maximum sample values | Negative and positive sample extremes | Use the greatest absolute amplitude for sample headroom; one extreme alone can miss an overload |
| Maximum true peak | An oversampled peak estimate | Check peaks that can appear between stored samples |
| Filtered true peak | The larger filtered or unfiltered true-peak result | Catch readings affected by asymmetry or DC offset |
| Integrated LUFS | Weighted, gated program loudness over the selected track | Compare the source, brief and final render |
| Maximum short-term | The loudest three-second integration window | Locate sections that dominate perceived loudness |
| Maximum momentary | The loudest 400 ms window | Inspect brief loud passages without treating them as the whole song |
| LRA | Distribution of loudness through the program | Track broad dynamic change, not a quality score |
These numbers describe the selected audio, not an uncommitted effects preview. A peak below full scale does not prove that an earlier clipping distortion has disappeared. Sound Forge labels its true-peak fields dB FS in the help; distinguish those fields from sample peak when comparing with a dBTP delivery limit. For material shorter than a minute, EBU R128 does not recommend LRA because too few three-second values are available. None of these readings prescribes a universal music-master target. A quieter track can be healthy, and a louder one can still have useful dynamics. Note what the music is doing before deciding what has to change.
Step 2: Correct broad balance with Modern Equalizer
Open View → Plug-In Chain, click Add Plug-Ins to Chain, select Modern Equalizer, click Add, then OK. Loop a dense and a quieter passage to compare settings. The current Plug-In Chain help distinguishes real-time preview from the processing applied when you save the audio. Keep the untouched source separately; do not overwrite it while testing. The current Modern Equalizer documentation describes an eight-band parametric EQ with high-pass, low-pass, shelf, peaking and band-pass filters, Left/Right/Mid/Side processing, Output Gain, A/B memories, Bypass and oversampling up to 8×.
Make one broad correction at a time: match the processed level with Output Gain, switch Bypass, and listen again. The focused Modern Equalizer guide covers the band and channel controls. A broad 1 dB move can be consequential on a stereo mix, and there's no reliable recipe that says every master needs a cut at 200 Hz or a boost at 8 kHz.
Use the analyzer to locate energy, then use references and hearing to decide whether that energy is a problem. If the EQ curve becomes a collection of narrow repairs, stop and ask whether the mix or an earlier stem is the better place to work.
Step 3: Treat Wave Hammer as two decisions inside one plug-in
The official Wave Hammer help is explicit: Wave Hammer is one mastering plug-in containing a classic compressor and a Volume Maximizer. This means the classic Wave Hammer, not Wave Hammer 2.0, which uses up to four frequency bands. Add the classic effect after the EQ in the same chain. Its compressor and maximizer solve different problems; neither must work hard just because it is available. The classic Wave Hammer guide separates their controls and the legacy-interface caveats.
First decision: compressor
Evaluate the compressor without added maximization first. Use the effect’s ? help to confirm the stage controls in your build; the current web overview does not document every button. Adjust compression while listening to the loudest sections, transitions and sustained notes. The goal is controlled movement, not a particular ratio or attack time.
Reduce the processing when the groove becomes smaller, cymbal decay jumps forward, consonants spit or the low end pumps between kicks, then match the bypass level. If the compressed version wins only while it's louder, the comparison hasn't proved an improvement.
Second decision: Volume Maximizer
After the compressor decision is stable, adjust the Volume Maximizer. Increase its contribution gradually, stopping at the required level or backing off as soon as the sound suffers. Listen for flattened transients, rough vocal edges, narrowed depth and low-frequency pumping. The clean stopping point depends on the mix and delivery brief, so fixed gain and ceiling settings are poor substitutes for a test render.
A sample-based ceiling inside the effect is not a promise that the encoded or reconstructed signal will return the same dBTP value. Save and reopen a test, then inspect both true-peak fields in Statistics. A short excerpt can expose artifacts, but only the complete delivery file can pass the final loudness check. If one controlled maximizing pass cannot reach the requested level without damage, revise the mix or agree a different target with the recipient; do not silently change the brief.
Step 4: Verify loudness and make the bit-depth decision
Use File → Save As to make a processed high-precision working audio copy without final delivery dither or bit-depth reduction; for a Sound Forge project, use File → Render As. Reopen that output, check its own Plug-In Chain, and bypass any effects already baked into the file before auditioning or saving it again. Select the complete file and both channels, then run Statistics. Compare Integrated LUFS, LRA where applicable, both sample extremes and both true-peak fields with the baseline and brief. The Plug-In Chain workflow explains why preview, applied processing and saving are different checkpoints.
Don't treat −14 LUFS as a universal music-master submission target. Spotify’s own mastering advice recommends −14 Integrated LUFS with true peak below −1 dBTP, or below −2 dBTP for louder masters. That is destination-specific guidance; its Normal playback reference is not every service’s submission rule. Playback normalization varies by service and user setting, while genre, dynamics, client requirements and the loudest clean result all affect the master. If a broadcaster or client gives a formal loudness specification, I'd follow that document rather than a generic streaming chart.
If the brief calls for loudness normalization and the available method matches it, use Process → Normalize → Loudness on the processed working copy before final dither. This is conditional processing, not an automatic extra mastering step. The current Normalize help documents EBU R128 and ITU-R BS.1770/1771 modes, integrated loudness, tolerance, maximum true peak and loudness-range handling. Peak normalization is a different operation: it applies constant gain, so it also changes integrated loudness. The normalization guide separates the modes. For formal delivery, also confirm the required measurement method in the loudness-meter workflow; a preset name is not a compliance certificate.
Finish processing before the final integer conversion
If a 24-bit or 32-bit-float source must become a 16-bit deliverable, commit the approved EQ, compression, maximization and gain changes to an undithered high-precision working render first. Reopen that render and bypass the already-applied chain before continuing. Do not dither the unprocessed waveform and then apply the real-time mastering chain on save. If the delivery needs a different sample rate, complete that conversion before final dither; the sample-rate conversion guide covers the required-rate derivative. Then choose Process → Bit Depth → Bit-Depth Converter. The Sound Forge 2026 Bit-Depth Converter help lists Half Rectangular, Rectangular, Triangular, Highpass Triangular, Gaussian and POW-r dither options.
Audition the quietest fade or tail and apply dither once. The official help says the best POW-r mode depends mainly on the signal, so there's no defensible type for every track. An unchanged file already at the required integer bit depth needs no second dither pass. A processed 24-bit source may leave a floating-point effects chain before a 24-bit integer export, so matching source/output labels alone does not settle quantization. Check the final processing and export path; avoid two dither stages. FabFilter’s explanation of final-stage dither also distinguishes fixed-point conversion from MP3 or AAC encoding. Keep an undithered high-precision working version when available, and derive each delivery from that version. Do not make a 16-bit intermediate solely to create an MP3.
Step 5: Export, reopen and check the deliverable
Export only the versions the brief needs. In Save As or Render As, choose the file type and template/custom settings deliberately. Use a filename that identifies the version, sample rate and PCM bit depth where applicable. Check required metadata, too: supported fields depend on the output format. Close and reopen the exported file, confirm no unwanted chain is active, and check it from start to finish.
Check the reopened file’s duration, channel count, sample rate and encoding properties; use Statistics for Integrated LUFS, LRA where applicable, both sample extremes and both true-peak readings. The save-time loudness log is measured before codec encoding, so it cannot replace this check of a lossy export. Listen to the first and last seconds, every fade, transitions around edits, the loudest chorus and the quietest tail. Check derived lossy files for clicks, pumping, harshness or codec-related distortion.
Finally, compare the export with the untouched source and a relevant reference at matched perceived level. The master should show a clear improvement once the loudness advantage is removed. If it doesn't, I'd return to the last checkpoint that changed the sound instead of adding another processor.
Stop signs and the next move
| What you hear or measure | Likely checkpoint | Next move |
|---|---|---|
| Low end pumps after loud hits | Wave Hammer compressor or maximizer | Reduce processing and retest the loudest section |
| The louder version loses depth when level-matched | Volume Maximizer | Back away from the first audible trade-off |
| True peak is higher than expected after saving | Render and verification | Adjust the chain, save a new test and rescan |
| The 16-bit fade sounds grainy | Bit-depth conversion | Compare dither on the quiet tail and apply it once |
| A vocal or instrument is still too loud | Source mix | Return to the mix or stem instead of forcing the stereo master |
Five-step checklist
- Measure: save the whole-track, both-channel Statistics baseline.
- Balance: use Modern Equalizer with level-matched A/B checks.
- Control: set Wave Hammer compression, then maximization.
- Verify: reopen the processed copy without an extra chain, rescan loudness and peaks, then make any required final integer conversion with one planned dither stage.
- Deliver: reopen the export, inspect it and listen against the untouched source.
Frequently Asked Questions
What order should I use for mastering in Sound Forge Pro?
Measure the complete source first, then correct tonal balance with Modern Equalizer, control dynamics with Wave Hammer, verify loudness and true peak, convert bit depth only when required, and reopen the exported file for final QA.
What is Wave Hammer in Sound Forge Pro 2026?
Classic Wave Hammer combines a compressor and Volume Maximizer in one plug-in. Wave Hammer 2.0 is a different multiband effect. Evaluate the classic stages separately and verify the saved result with Statistics.
Does a −1 dB ceiling guarantee −1 dBTP?
No. A sample-based plug-in ceiling does not guarantee the same reconstructed true-peak reading. Save a test and inspect maximum true peak and filtered true peak in Tools → Statistics.
Should I master every streaming track to −14 LUFS?
No. −14 LUFS is not a universal submission requirement. Use the destination brief, the music's dynamics and the loudest clean result, and remember that playback normalization varies by service and user setting.
Should I peak-normalize after Wave Hammer?
Only when you deliberately need the resulting constant gain change. Peak normalization changes integrated loudness along with the peak level, so rescan the complete file afterward rather than treating Normalize as an automatic final step.
Which POW-r dither type should I use?
There is no single type for every master. Sound Forge's current help says the best mode depends mainly on the signal. Compare the quietest fade or tail, and apply dither once when the final bit depth is being reduced.
Is Ozone included with Sound Forge Pro 2026?
Do not assume it is included: Ozone is not listed in the current Boris FX product comparison. You need a separately licensed, compatible Ozone build if you want to use it as a VST effect. Check that build’s requirements rather than relying on an old bundle description.