Resolve: exporting clips with the right frame rate, effects and audio

Agree on a finished program or media for further editing

Before choosing a render mode, ask what the recipient needs: the finished presentation, excerpts ready to share, or graded shots to rebuild the edit. Record the format, resolution, frame rate, audio, timecode and effects that must be baked in. Keep the project and source media; preparing files for an editor means deciding what will remain editable in the next application.

The Resolve 20.3 manual, February 2026, pp. 4064–4065 distinguishes Single Clip, which produces one file for the selected range, from Individual Clips, which produces media for each clip in that range. Individual Clips is described for interchange and reconforming. Do not choose it just because you want several finished pieces; first decide what each file must contain.

For a finished presentation or segment, test Single Clip with the agreed range and audio. For a return to the editor, test individual clips with the required conform information and effects. The delivery guide helps record who receives the files, what is included and what has been accepted.

Individual Clips retains source frame rates: also check resolution and timecode

Resolve 20.3, pp. 4064–4065 documents the project frame rate for Single Clip and each source frame rate for mixed material in Individual Clips. It describes source timecode for Individual Clips and timeline timecode for Single Clip. Inspect the rendered file and any additional controls: that general description does not verify the timecode after baking in effects on a particular system.

Resolution has a separate control. Render at Source Resolution, pp. 4040 and 4065 lets you select original resolution in Individual Clips mode. Otherwise, output/project resolution and scaling apply. For source interpretation and mixed frame rates, see the XML guide.

Speed changes are missing: check Render Timeline Effects

Individual Clips, manual 20.3, p. 4065 allows timeline effects to be ignored or included using Render Timeline Effects. The Delivery Effects Processing chapter, pp. 4039–4041 explains why some processing can travel as XML/AAF data during a roundtrip so the editor can reapply it. Decide whether you are delivering that data or the baked-in effect, then compare the returned edit for processing applied twice.

Render a sample from the project copy with the settings recorded. Compare duration, in/out points and motion against the reference, then check timecode and audio separately. If the shot still differs, keep the sample and settings for support. Check the effects on each shot before processing the batch.

Record framing and effects settings together

For sizing adjustments, Disable sizing and blanking output, p. 4041 determines whether sizing and blanking are baked into the render. The manual states that this option does not disable transformations in Fusion or Open FX/Resolve FX plug-ins. Do not describe a package as effect-free without checking the controls and processing that produced it.

Record these settings before adding the render job. If the recipient needs to adjust framing or speed, agree on what is baked into the files and what depends on interchange data. Keep a reference export of the edit so you can compare a reconstructed sequence before rendering the entire batch.

Export Audio alone does not verify the mix, effects or timecode

The audio mix, baked-in effects and timecode are separate requirements. Define all three before exporting: enabling Export Audio does not demonstrate that the file contains the bus, processing and timecode the recipient needs.

The Audio Panel, manual 20.3, pp. 4073–4074 documents Export audio, source or synced audio, and Output Track for choosing Main/Submix. In Single Clip, All Timeline Tracks delivers separate tracks with their channels; it does not mean a single mix of everything. Agree on a mixed bus, separate voices or source audio, and record that selection.

For a finished excerpt with its mix, compare Single Clip over the agreed range with the required audio output. If you need camera timecode together with processing, test individual clips and read the file timecode as well as listening to the audio. Verify the timecode after rendering with the exact settings and installed version. If it differs, document the difference and agree on an alternative delivery with the editor before approving the batch.

Check a sample and record which batch files you reviewed

Include one source at each frame rate, a shot with speed or framing changes, and a spoken segment with processed audio. Compare the received files against the same edit reference, recording resolution, frame rate, timecode, duration, speech and included effects.

For multiple clips, use a new destination and check filenames and file count before delivery. The File section, pp. 4075–4076 documents unique filenames and separate folders to prevent outputs from the same source overwriting each other. Keep a record linking each output to its original. Do not rename linked originals to fix a render filename collision.

Fill in the Resolve export worksheet with completed tests, not just planned checks. An accepted sample does not certify the whole batch: list the files and ranges reviewed and those still pending. To plan capture and delivery, connect this record to the session multicamera recording plan.

Several excerpts produced only one file: check names and destinations

The following controls are documented in manual 20.3. Also identify your build and source format.

In File, manual 20.3, pp. 4075–4076 and 4078, Use unique filenames adds a track and clip-position identifier to distinguish outputs. It is documented for Individual Clips with Filename uses set to Custom name. To preserve source names, check Place clips in separate folders: it separates excerpts into folders so they do not share both a path and filename. Agree on the method with the editor before generating the batch.

The manual describes session-based identifiers and naming variables; it does not establish a counter that scans all files already in the destination.

Export two excerpts from the same source to a new destination. Record each full filename and path, open both files and compare their ranges with the reference. If only one remains, keep the settings and stop acceptance of the batch. A fixed File suffix alone does not distinguish two excerpts assigned the same name.

Keep the relationship between output, source and cut. If the editor needs XML/AAF to rebuild the sequence, agree on that interchange too and check the return using the XML guide. Avoid renaming files after generating an interchange file that references them; verify the links in a copy before delivery. The delivery guide helps document the package the editor will receive.

Extra frames with zero handles: distinguish margins, full extents and retiming

Identify the Resolve version and build, source and project frame rates, and the first and last frames used. This separates a conform margin from an actual change to the cut.

The 20.3 manual, p. 4072 distinguishes Add X frame handles, which adds frames before and after a clip for adjusting cuts, from Render full extents, which exports the entire source in Individual Clips. Record both controls. A longer file may reflect an agreed delivery choice; compare the first and last frames used in the edit as well as the total media duration.

Another documented condition appears in the note on p. 4040: Resolve adds three frames of handles to retimed clips and renders whose frame rate differs from the project, to assist reconforming. Setting the manual handle value to zero therefore does not guarantee identical boundaries for every individual-clip render. Check this behavior in your installed version.

Compare an excerpt without speed changes and another with retiming, keeping the frame rates and settings recorded. Compare duration, timecode and used frames with the same edit reference, then check reconstruction in the destination editor. Record any missing frame-rate or interchange information before attributing a difference to handles or a software error. Do not change the project frame rate to hide the symptom.

The filenames and handles worksheet records requested margins, rendered media and reconstructed cuts. If the recipient needs finished pieces with exact boundaries, agree on that separately from a conform package. Keep the originals and project; verify the reconstructed cuts before accepting delivery.

Playback prefers originals: also check media selection in Deliver

If artifacts appear in the render, identify the camera, format, operating system and Resolve version. Compare the media being used before generating optimized media: creating another derivative does not establish the cause.

The 20.3 manual, pp. 217–218 separates Proxy Handling for playback from Use proxy media in Deliver. Prefer Camera Originals can display a proxy when the original is missing; choosing that mode does not establish which media the render used. Timeline Proxy Resolution is another function: p. 199 describes a reduced working resolution, independent of creating Proxy Media files.

For output, separately record Use optimized media, p. 4070, and Use proxy media and Use render cached images, p. 4071. The manufacturer allows these derivatives to save processing when available. If they were created at lower quality than the delivery requires, the manual says to disable the relevant control. Disabling proxies alone leaves the other two options unchecked.

Export a sample containing fine text, motion or gradients, keeping the output resolution, format and edit unchanged. Record a baseline state and change one control at a time if comparison is needed. Inspect the exported file as well as the viewer, and record which derivatives existed for each source. Keep the media and original settings: this check does not require deleting the cache.

Excluding derivatives from the selection does not establish a cause or fix every artifact. If the symptom persists, record its exact location and continue with file and playback comparison. Do not change the GPU or processing mode without checking compatibility and saving the previous configuration.

Originals are missing but proxies exist: define the delivery limits

Distinguish missing originals from files that are present but fail during playback. Check which proxy belongs to each shot before deciding whether it meets the agreed delivery requirements.

Using Proxy Files for Delivery, manual 20.3, p. 218 describes Use proxy media for output using proxies and for work without access to originals. Proxy Handling, p. 217 documents substitution for missing originals during playback. These are different functions: those sections do not establish an export process that detects corrupt files and automatically substitutes only those clips.

Before testing, list which originals are missing, which can be opened and which proxy is linked to each shot. Record the derivative resolution, format, duration and timecode, then check the shot and cuts. The documented purple line for missing originals helps locate them; it does not prove every original that is present is intact. For relinking and source comparison, see the Media Offline guide.

If a temporary proxy-based version is agreed, identify the shots that depend on proxies and their available quality. An output with more pixels cannot recover detail absent from the derivative. Keep originals, proxies and the project to review substitution; verify picture, speech and continuity in the generated file before accepting it.

The render source worksheet separates available media, control settings, the planned test and its actual result. Keep a delivery requiring originals pending until the necessary shots have been checked. Communicate that scope to the recipient through the event delivery guide.

Only audio or a trim changed: check when video can bypass re-encoding

Identify the Resolve version, build, Free or Studio edition, and operating system. Copying video without re-encoding also depends on the exact format and codec variant.

Bypass re-encode when possible, manual 20.3, p. 4070 allows video essence to be copied when output Format, Codec and Type match the source and the media is compatible. The manufacturer describes simple cuts of previously rendered material and replacing an audio mix. The result is still a new file: copying video does not remove the need to write and check the delivery.

The same section states that grading, Resolve FX, compositing, resizing, stabilization or Fusion effects require re-encoding the affected clip. Transitions require processing over their duration. An enabled checkbox does not demonstrate that the entire timeline can be copied; retain required effects and record where processing is needed.

For compatibility, distinguish Decode, Encode and Codec Passthrough in Resolve 20 Supported Formats and Codecs, July 2025. For example, the macOS table on p. 3 lists mov passthrough for Apple ProRes; H.264/H.265 have decoding and encoding entries but a dash under passthrough. Note (1), p. 14, requires no effects or color changes and a format Resolve can both read and render. That list applies to the stated version and platform; verify your build and exact format combination.

Export one sample with a simple cut and another with the required audio change or effect, using an identified project copy. Record format, variant, resolution, frame rate, processing, Bypass setting and render time. Check picture, mix, channels and file boundaries against the reference. Render time, file size or the absence of a preview alone does not prove the video essence was copied.

Speech is delayed only after export: compare output settings and the file

If speech matches the picture in the edit but arrives late in the exported file, first compare the same point in both. Then isolate one output setting at a time; changing two controls together prevents you from identifying which one changes the result.

The 20.3 manual, p. 4069 describes Frame Reordering for QuickTime H.264/H.265: it enables B-frames, and disabling it changes the tradeoff between encoding speed and visual quality. Bypass, p. 4070 concerns copying or re-encoding video. Neither is presented there as a universal sync repair.

Keep the project and initial export for comparison. Identify a visible and audible gesture or impact near the start, middle and end; compare those points in the source, edit reference and rendered file. Record whether the offset stays constant or grows, and where and in which player it occurs. Do not change frame rate or interpretation across the edit to compensate for a symptom without checking the files first.

If these controls need testing, use the project copy and change one at a time, where available for the selected format. Record the initial state, one sample with Bypass changed and another with Frame Reordering changed, leaving all other settings unchanged.

The Bypass and sync worksheet provides fields for compatibility, settings and comparison points. List the reviewed ranges and outputs; a short excerpt does not verify an entire presentation. If the symptom persists, use the file and playback comparison guide and list unresolved shots in the delivery record.

If the file also changes contrast or color during playback, separate that observation from sync using the levels and color guide. Record the file, frame and playback destination for each comparison.

If the recipient requests interlaced output or motion changes when resolution is converted, follow the fields and scaling guide. Record source, processing and output before accepting that delivery.