Cómo Eliminar Silencios en DaVinci Resolve (Versión Gratuita)

Can the free version of DaVinci Resolve remove silence?

It can, as long as you are on 20.2 or newer. A lot of tutorials still say Resolve has no automatic silence removal, and for years that was true. Blackmagic added Ripple Delete Silence in the 20.2 update, described in the official 20.2 New Features Guide as a way to "analyze and remove silences from a selected clip on the Edit or Cut page timeline, using the similar tool found in Fairlight." The guide does not mark it as a Studio feature, and it works in the free download. The current release line is Resolve 21 (21.1 shipped in September 2026), and the command is still in the same place.

What the free edition does not have is the transcription-based workflow. Audio transcription, including its "Remove Silent Portions" option, needs DaVinci Resolve Studio. The Fairlight page's own Remove Silence command is also listed as Studio-only by third-party guides; I could not find an official Blackmagic table confirming that, so check the clip's right-click menu in your own install before you plan around it.

Which silence removal methods exist in Resolve, and which are free?

There are four realistic routes. They differ in whether the timeline gets shorter and whether you can review each cut before it happens.

How do you use Ripple Delete Silence step by step?

1. Duplicate the timeline first. Right-click the timeline in the Media Pool and choose Duplicate Timeline. The command cuts and ripples in one go, and you will want the untouched version to compare against.
2. Check Sync Lock on the other tracks. The 20.2 guide explains that ripple edits on the Edit page now preserve sync across tracks by cutting clips on other tracks at the edit point. A music bed under your talking head will be chopped at every removed pause. Turn off Sync Lock on tracks you want left alone, or add music after the silence pass.
3. Select one clip on the timeline (the linked video and audio of your talking-head take).
4. Choose Clip > Audio Operations > Ripple Delete Silence. The dialog opens with Threshold, Pre Head, Post Tail, Minimum to Trip and a Cross fade audio checkbox.
5. Set the values from the table in the next section. The selected clip now shows vertical red stripes where audio will be removed.
6. Scrub the stripes. Look for stripes that land inside words, especially soft endings like "-s", "-th" and trailing vowels. If you see them, lower the threshold (more negative) or add Post Tail.
7. Click Remove. Everything under the red stripes is cut and the remaining pieces are ripple-deleted together.
8. Play the joins. Listen to five or six random cuts at full volume on headphones. Undo with Cmd+Z / Ctrl+Z and adjust if the pacing feels rushed.

Which Ripple Delete Silence settings keep natural pauses?

The official definitions are short. Threshold is "the maximum signal level of your audio (in dB) that will be recognized as silence," and higher values remove more. Minimum to Trip is the minimum amount of silence, in frames, that gets stripped; lower values make more cuts and, in Blackmagic's words, can leave audio that sounds "chopped up." Pre Head and Post Tail set how much silence stays before and after each audible section, so the pause left behind after a cut is Post Tail plus Pre Head.

AutoSilenceCut's presets were tuned for spoken word, so they make a useful translation table. Our Natural preset only touches gaps of 0.8 seconds or longer and leaves 0.32 seconds of pause, split evenly on both sides of the cut. Here is the same behaviour expressed in Resolve frames:

Two caveats. Blackmagic's guide gives Minimum to Trip in frames but does not state the unit for Pre Head and Post Tail; third-party walkthroughs show them in frames too, so confirm in your dialog before copying the numbers. And the thresholds are starting points, not absolutes: AutoSilenceCut measures RMS energy in 25 ms windows, and Resolve's detector is not documented in that detail, so the same number will not always flag the same gaps. Natural is what I would use for interviews, tutorials and anything with a guest. Fast suits short vertical clips where viewers expect speed.

Why does it cut words or find no silence in a noisy room?

A level detector only sees loudness. If your pauses sit at -42 dB because of a laptop fan and your quiet words dip to -40 dB, there is no threshold that separates them cleanly. You get one of two failures: a low threshold finds almost nothing, or a higher one starts eating word endings.

• Read the noise floor first. Play a pause and watch the audio meter. That number plus 6 to 10 dB is your threshold.
• Lower the noise before detection. A high-pass filter around 80 Hz removes rumble that inflates the level of every pause. Resolve's Voice Isolation and Neural Engine noise tools are Studio features, so on the free edition a browser cleanup pass like the noise and silence remover is one option: it filters first, then detects silence on the cleaner signal.
• Raise Minimum to Trip. Fan noise produces short dips that look like silence. Requiring 0.8 seconds or more ignores them.

How do you remove silence by hand in Resolve 20.1 or older?

The manual method still matters for older installs and for anyone who wants to decide every cut. It is slow (budget 10 to 15 minutes per 10 minutes of talking-head footage), but nothing gets removed that you did not choose.

1. Make the audio waveform tall: open the timeline View Options and increase the track height so pauses are obvious.
2. Park the playhead at the start of a pause and split with Timeline > Razor (Ctrl+B on Windows, Cmd+B on Mac). Split again where speech resumes, leaving a few frames of breath on both sides.
3. Select the piece between the two splits and choose Edit > Ripple Delete, or right-click and pick Ripple Delete. The rest of the timeline slides left.
4. If you lifted pieces with a plain Delete and left holes behind, Edit > Delete Gaps closes all of them at once.
5. Add a short audio crossfade at joins that click. One or two frames is enough for dialogue.

How do you bring a browser cut list into Resolve with an EDL?

Resolve's built-in command is all-or-nothing: you see the stripes, but you cannot keep stripe 14 and remove the rest. If you want to hear each cut and restore the ones that hurt the delivery, run the detection in the video silence remover, review the cuts in the browser, and export the decisions as an EDL that Resolve can rebuild as a timeline. Nothing is uploaded; the file is decoded and analysed locally.

What the AutoSilenceCut EDL contains

These details come straight from the exporter source code, so you know exactly what Resolve will receive. Choose "Adobe Premiere / DaVinci (.edl)" in the export format menu and you get a file named yourclip-autosilencecut.edl with:

• A CMX 3600 header: TITLE: yourclip - AutoSilenceCut and FCM: NON-DROP FRAME, or FCM: DROP FRAME when you pick a drop-frame rate.
• One cut event per kept segment, on track AA/V (video plus both audio channels), with source in/out and record in/out timecodes.
• A reel name made from the first 8 letters and digits of the file name (other characters become underscores), plus a * FROM CLIP NAME: comment with the full name without extension.
• Timecode written at the frame rate you choose: 23.976, 24, 25, 29.97 drop-frame, 29.97 non-drop, 30, 50, 59.94 drop-frame, 59.94 non-drop or 60 fps. Drop-frame timecode uses the SMPTE ; separator (for example 00:01:00;02), and every cut point is rounded to a whole frame at that rate.
• A source start timecode, 00:00:00:00 by default, that you can change to match the camera file.

The frame rate setting decides whether the cuts land where you heard them. It defaults to 30 fps non-drop unless the browser can reliably detect the video's frame rate while loading it, in which case that rate is preselected; your last choice is remembered in the browser. Always confirm it against the clip's real rate in Resolve. Screen recordings and phone clips usually start at 00:00:00:00, but many cameras write time-of-day or 01:00:00:00 timecode. Set the source start timecode to the first frame's timecode shown in Resolve's metadata panel so the EDL source in and out points match the file.

Import steps in Resolve

1. Check the clip's frame rate and start timecode in the Media Pool (Clip Attributes or the metadata panel) and make sure the EDL was exported with the same values.
2. Import the original video into the Media Pool first.
3. Choose File > Import Timeline (Shift+Cmd+I / Shift+Ctrl+I) and select the .edl file.
4. In the import dialog, untick "Automatically import source clips into media pool" so Resolve links to the clip you already imported, and make sure the frame rate matches the one you exported.
5. If clips arrive offline, open Project Settings > General Options > Conform Options and try turning off "Assist using reel names". Forum users report this is the usual fix when the source file has no embedded reel name.
6. Spot-check the first cut and the last cut against the waveform. If the last one drifts, the frame rate does not match; if every cut is offset by the same amount, the source start timecode does not match.

The export menu also offers "Final Cut Pro / DaVinci (.fcpxml)", which Resolve can import too. It is a version 1.9 sequence with frame-accurate times at the chosen rate and a media reference that contains only the file name, since a browser cannot read folder paths. Expect the clip to appear offline after import; relink it to the original file (Resolve matches it by file name) and the cuts snap into place.

Common mistakes when removing silence in Resolve

• Pre Head and Post Tail at zero. Words collide and breaths vanish. Keep at least 0.1 seconds on each side for anything conversational.
• Running it with music on the timeline. Since 20.2, sync-preserving ripples cut the music bed at every removed pause. Add music afterwards or turn off its Sync Lock.
• Trusting the default threshold in a noisy room. Measure the pause level on the meter and set the threshold from that, not from a tutorial.
• Forgetting the picture. Every removed pause is also a jump cut. Our guide on jump cuts without jumpy audio covers punch-ins and J-cuts that hide them.
• Exporting the EDL at the wrong frame rate. A 30 fps EDL in a 24 or 25 fps project, or non-drop timecode for 29.97 drop-frame footage, puts the cuts in the wrong place. Pick the rate that matches the footage before you export.
• Not keeping the original timeline. Ripple Delete Silence is undoable only until you close the project. A duplicated timeline costs nothing.

If you mostly edit audio-only shows, the same logic applies in a DAW; see Audacity Truncate Silence settings and the broader silence detection settings guide. Premiere users can follow removing silence in Premiere Pro without a plugin.