Variable Frame Rate in DaVinci Resolve: Diagnose and Convert to CFR

Quick answer: don't convert a whole project just because one phone clip stutters or slides out of sync. Open the suspect file in MediaInfo first, reproduce the problem with that single clip in a clean Resolve project, make one constant-frame-rate copy and run the exact same test again. Leave the original alone. The CFR copy earns its place only if its duration, audio channels, cadence and sync all hold up side by side.

Variable frame rate does cause real editing trouble, but it has turned into the default diagnosis for every awkward phone or screen recording, and that wastes time. A file can misbehave because of its codec, a wrong timeline rate, damaged frames, slow storage or clock drift from a separate audio recorder. So skip “Did this come from a phone?” The better question: do this file’s timestamps actually vary, and does regularizing them fix the same symptom?

VFR Is Not the Same as Mixed Frame Rates

A variable-frame-rate file changes the gap between frames inside one recording. A constant-frame-rate file puts every frame on a regular time grid, even when the picture itself contains repeated frames or uneven motion.

Mixed frame rates are a different animal. Say you have one 23.976 fps camera clip, one 25 fps clip and one 59.94 fps clip. Every one of them can still be constant. Resolve has to fit them into the timeline rate you picked, but that doesn't turn any of those sources into VFR.

It helps to keep four layers apart in your head:

• Source timing: the timestamps written into the media file itself.
• Clip interpretation: how Resolve reads that source once it's placed in a project.
• Timeline frame rate: the time grid you edit on.
• Playback performance: whether your hardware can decode and display frames in real time.

Changing the Timeline Frame Rate or Clip Attributes changes how media is interpreted. It doesn't write a new source file with evenly spaced timestamps. Blackmagic’s manual is specific about Clip Attributes: a new Video Frame Rate changes the clip’s duration and playback speed in Resolve while its audio stays as it was, so forcing a rate onto a VFR clip can make drift worse. If the real problem is broader project setup, start with the DaVinci Resolve frame-rate guide before you transcode anything.

When VFR Is a Reasonable Suspect

VFR deserves a closer look when the trouble follows one particular recording or source family and you see one or more of these symptoms:

• picture and embedded audio start together but drift apart later on;
• a long capture behaves worse near the end than near the beginning;
• scrubbing lands inconsistently around the same section;
• the waveform and the visible action don't stay aligned;
• the file behaves differently after a CFR conversion while every Resolve setting stays the same;
• a phone, conferencing app or screen recorder changed its recording cadence under load.

Those symptoms justify a test. They don't prove the cause. Blackmagic’s user forum has recurring reports of sync problems, audio breakup and visual artifacts with some phone and capture files, with CFR transcoding offered as the workaround. Read that discussion as community evidence of the workflow problem, not as an official statement that every such file is unsupported.

Look somewhere else first when:

• every H.264 or H.265 clip plays slowly, including known CFR camera media;
• the issue disappears when proxies or optimized media are used;
• audio came from a separate device and drifts against every camera file;
• multiple clips have different constant rates but behave consistently inside their own files;
• the source won't import at all;
• the media repeatedly goes offline because paths or drives change;
• the file contains obvious corruption or decoding errors.

Slow decoding has its own fix in the Resolve slow-playback workflow. If the file fails at import, check the MP4 import guide and the current supported codecs and formats overview. A frame-rate conversion can't solve a codec, edition or operating-system limitation, so don't ask it to.

Preserve the Original Before You Test Anything

Never overwrite your only copy of the source. A timing conversion rewrites the media itself and, depending on the application and preset, can also change stream layout, metadata, timecode or color tags.

Set up a small test folder with:

1. the untouched source file;
2. one CFR copy with an obvious suffix such as _CFR_TEST;
3. a note with the source duration, the reported frame-rate fields and the audio-track count;
4. a clean Resolve project used only for the comparison.

Pick a clip that actually shows the problem. If the failure only appears after 35 minutes, a 20-second sample from the beginning proves very little. Use a clip long enough to reach the failing region, or make a test copy that keeps the relevant section without changing its timing first.

Check the Source in MediaInfo

MediaInfo makes a fast first pass because it reads stream and container metadata without Resolve interpreting anything. Its official field reference documents frame-rate mode, nominal, minimum and maximum frame-rate fields.

Open the file and switch to a detailed view such as Tree or Text. In the video stream, look for:

• Frame rate mode — Constant or Variable;
• Frame rate — often an average or reported working value;
• Minimum frame rate;
• Maximum frame rate;
• Nominal frame rate, when present;
• duration and frame count;
• codec, bit depth, chroma subsampling and color information;
• audio stream count, channel layout and sample rate.

A reported Variable mode is strong evidence that the source has irregular timing, and a real gap between minimum and maximum frame rate backs that up. A single average such as 29.97 isn't enough on its own: a VFR file can still report an average sitting right next to a familiar broadcast rate.

Metadata can also be incomplete or misleading. Some files leave fields out; some containers expose a nominal value that says little about the actual frame intervals. Treat MediaInfo as the first layer of evidence, then back it up with a controlled playback and sync comparison.

If the footage matters, save the MediaInfo output with your project notes. That gives you a factual record of the original before any transcode reshapes it.

Build a Controlled Resolve Test

A useful test changes one thing at a time. Comparing the original inside a heavy production timeline against the CFR copy in an empty new project proves nothing, however convincing the result looks.

Set the test up like this:

1. Create a new project.
2. Choose the intended timeline rate before you import the test media.
3. Import only the original suspect clip.
4. Place it at the start of a new timeline.
5. Add visible sync markers at the beginning, the middle and near the end. A hand clap, a sharp spoken consonant, a mouse click, a slate or any clear visual action works better than vague lip movement.
6. Write down the exact sections where drift, stutter, audio breakup or frame instability shows up.
7. Close the project or duplicate the timeline before importing the CFR copy.
8. Place the CFR copy at the same start position and repeat the same checks.

Keep render cache, effects, scaling and speed changes out of this stage. You're comparing source timing here, not grade performance or Fusion load.

If both files behave the same, VFR probably isn't the useful diagnosis. If the CFR copy fixes the same repeatable symptom while duration and audio stay correct, you have a solid reason to use it.

Convert One Sample in HandBrake

HandBrake is the friendliest GUI route to a CFR working copy. Its frame-rate documentation distinguishes variable and constant output: VFR keeps the source timing pattern, while CFR produces one constant output rate. The same page names import into an NLE as one of the few cases where CFR is the right choice, which is exactly the job here.

For the first sample:

1. Open the untouched source in HandBrake.
2. Choose a container and video encoder that Resolve decodes reliably on your system.
3. Open the Video tab.
4. Set Framerate (FPS) to Same as Source.
5. Select Constant Framerate.
6. Check the Audio tab and include the tracks and channel layouts the edit actually needs.
7. Send the output to the test folder with the _CFR_TEST suffix.
8. Leave the original file in place.

HandBrake documents that Same as Source + Constant Framerate detects the source rate and makes any variable portions constant at that rate. That's a sensible first comparison when you want to regularize the recording rather than impose a new creative frame rate. One trap: HandBrake’s built-in presets use Peak Framerate, which leaves anything below the cap untouched, so check that Constant Framerate is still selected after you load a preset.

Don't pick 24, 25 or 30 just because it looks familiar. Set an explicit rate only when you know the target: a production deliberately built around 25 fps, say, or a source meant to be 30000/1001 where the conversion restores that known cadence. A wrong explicit rate can add extra dropped or duplicated frames before Resolve ever sees the file.

Frame-rate mode is only one part of a transcode. Check the video encoder, quality setting, dimensions, color metadata and audio choices separately. A CFR output doesn't guarantee that every source track or tag made it through.

Optional FFmpeg Diagnosis and Conversion

FFmpeg earns its place when you want a repeatable command, batch control or a second diagnostic signal. Its official vfrdet filter detects variable timing. A diagnostic command looks like this:

ffmpeg -i input.mp4 -vf vfrdet -an -f null -

Read the filter output as evidence, not a verdict on the whole edit. It counts frames with variable and constant timestamp gaps; it can't tell you whether your particular Resolve symptom comes from that variation.

FFmpeg’s documented fps filter creates a specified constant frame rate by dropping or duplicating frames as needed. A simplified conversion looks like this:

ffmpeg -i input.mp4 -map 0:v:0 -map 0:a? \
  -vf "fps=fps=30000/1001:round=near" \
  -c:v libx264 -crf 18 -preset medium \
  -c:a aac -b:a 320k output-cfr.mp4

30000/1001 is an example, not a universal target. Replace it only after you've identified the intended constant rate. The codec and audio options are workflow decisions too. This example maps the first video stream and any optional audio streams, but a source with multiple languages, unusual channel layouts, subtitles, data streams or production timecode needs a deliberate mapping plan.

For long-term editing media you may prefer an intraframe mezzanine codec to H.264. Choose it for the project’s quality, storage and platform requirements, not because VFR itself demands one specific codec.

Verify the CFR Copy Before Relinking

A finished encode isn't an approved edit source yet. Compare the new file against the original and against what the project needs.

Duration

Compare the reported durations closely. A tiny rounding difference around frame boundaries can be normal; a growing gap or a large duration change needs investigation. Put both clips at the same timeline start and compare where they end.

Audio structure

Count the tracks that survived and check their sample rates, channel counts and channel order. Listen to every channel you need. A stereo preview track won't stand in for isolated microphones if the project relies on them.

Sync

Check the beginning, middle and end. A one-hour recording doesn't pass because its first ten seconds line up. Use sharp visual and audible events wherever you can.

Cadence and motion

Watch pans, scrolling text, game motion and cursor movement. A CFR conversion may have to drop or duplicate frames, so make sure the resulting cadence is acceptable at the chosen target rate.

Color and levels

Compare a few dark, bright and saturated frames. A transcode can touch more than timing when color tags, range handling, chroma format or bit depth change.

Metadata and timecode

If the edit depends on source timecode, reel names, creation dates or camera metadata, check those fields one by one. Never assume a consumer conversion preset kept them.

If anything the edit needs fails, revise the conversion path before you touch the project. The target was never “a file Resolve opens.” You want a working copy that keeps everything this edit depends on.

Replace Media Safely in Resolve

Do the first replacement in a duplicate project or timeline. Keep the original media online and keep a clear map between original and CFR filenames.

For a handful of clips, import the copies into their own bin and swap them in deliberately. For a bigger set, know how Resolve’s relink tools search. Blackmagic’s manual says Relink Selected Clips matches on file name and timecode and ignores the old path, while Change Source Folder expects the same folder structure in a new location. A test suffix or stripped timecode can leave a relink with nothing to find, and duplicate names across phones, cards or capture sessions can point Resolve at the wrong file.

After the swap, verify:

• clip identity;
• in and out frames;
• timeline duration;
• speed changes and freeze frames;
• audio channel mapping;
• sync at multiple points;
• any edit that lands near a dropped or duplicated frame.

If media goes offline along the way, stop and work through the Media Offline diagnosis guide instead of changing more variables at once.

If the CFR Copy Does Not Fix the Problem

That's still a useful result. It means VFR is no longer the leading explanation, and you can move on to the next branch.

Timeline or clip-rate mismatch

A known CFR source can still judder when the timeline cadence and source rate are a poor match. Recheck the project, timeline and clip interpretation settings.

Decode performance

If both files use similarly demanding H.264 or H.265 settings, CFR alone may not make playback any lighter. Try an editing codec, proxies or optimized media and compare again.

Separate-recorder clock drift

Two devices can run at slightly different clock speeds. Both files can be CFR and a long recording will still drift. That calls for audio resynchronization or time correction, not a blind video frame-rate conversion.

Damaged media

Corrupt packets or missing frames can cause local glitches that survive conversion, or lead the converter to hide the damage by duplicating nearby frames. Inspect the source right around the exact failure point.

Storage or transfer faults

A slow, disconnecting or damaged drive can cause dropped playback, offline events and read errors. Copy one file to a reliable local drive and repeat the clean-project test.

When to Convert the Whole Set

Batch conversion only makes sense after one representative sample passes and the source group is consistent enough to share one set of rules.

Convert the wider set when:

• MediaInfo shows the same timing pattern across the group;
• the clips share the same recording method and required streams;
• the sample CFR copy fixes a repeatable problem;
• the chosen output codec, frame rate and audio mapping suit the full set;
• storage and processing time are planned;
• the originals remain preserved.

Keep testing clip by clip when the source folder mixes devices, apps, rates, orientations, HDR/SDR modes or audio layouts. One preset can turn out technically valid files while quietly stripping something you need from half the set.

A Short VFR-to-CFR Checklist

Before conversion:

• preserve the original;
• inspect Frame rate mode plus minimum, nominal and maximum fields;
• separate VFR from mixed constant FPS;
• reproduce the symptom in a clean project;
• choose one representative sample.

After conversion:

• compare duration;
• verify every required audio track and channel;
• check sync at the beginning, middle and end;
• inspect motion cadence and critical edit points;
• compare color and metadata;
• replace media only in a duplicate project or timeline;
• expand to a batch only after the sample passes.

It takes longer than flipping one dropdown, but it saves you from a far more expensive mistake: converting an entire project for the wrong reason and finding the damage after the originals have been moved aside.

Jason Miller I run DaVinci Resolve Club as an independent publication: hands-on edits, color grading breakdowns, Fairlight sessions, Fusion tests, and honest notes on where Resolve gets in the way.