To reassemble processed frames with the source timing in ComfyUI, connect the processed IMAGE batch to Create Video, connect Get Video Components’s fps output to its fps input, and connect its audio output if the new video should include the source soundtrack. Then compare the output’s picture duration with the expected duration from frame count divided by frame rate, and check separately that an audio track is present.
Capture the source values
Start with the source video connected to Get Video Components. The node provides the extracted frames as images, the audio track as audio, and the source frame rate as fps. Record the frame count in the extracted image batch and the displayed or reported fps value before processing. Keep the audio output available in the workflow; do not assume it will be included automatically when creating a new video.
The Get Video Components documentation identifies these as separate outputs. That separation is the key to preserving the parts you need: processed images can replace the original frames while the extracted timing and soundtrack remain available as inputs.
Reconnect the components in order
- Connect the
imagesoutput fromGet Video Componentsto the frame-processing workflow. Make sure the result you intend to assemble is anIMAGEbatch, not aVIDEOoutput. - Connect that processed image batch to
Create Video’s requiredimagesinput. - Connect the original
fpsoutput toCreate Video’s requiredfpsinput. Check that a wire reaches the input; otherwise the node may use its documented default of 30 fps. - If the delivery should retain the source soundtrack, connect
Get Video Components’saudiooutput to the optionalaudioinput onCreate Video. If you leave it unconnected, do not treat the resulting video as containing the source audio. - Run the workflow and inspect the resulting video’s picture duration and audio track separately.
Create Video expects images and an explicit frame rate; audio is optional. Its documentation lists a default fps of 30.0 and an accepted range of 1.0–120.0. See the Create Video documentation for the inputs and output.
Calculate the expected picture duration
For a sequence containing a known number of frames played at a constant frame rate, estimate its duration as frame count ÷ fps. For example, in a hypothetical source with 240 frames and a reported rate of 24 fps, the estimated picture duration is 10 seconds. If those same 240 frames are instead assembled at 30 fps, the estimate is 8 seconds: the frames are presented faster, not restored to the source timing.
This calculation assumes a uniform-rate sequence and that the count represents all frames to be assembled. It is a diagnostic estimate, not a substitute for checking the rendered result. If frames were added, removed, or duplicated during processing, use the processed batch’s actual frame count in the calculation. If the output differs from the estimate, first verify the count and the fps value actually connected to Create Video.
Use a reassembly check sheet
The numbers below are a hypothetical example, not a measured ComfyUI export. Use the same columns for your own workflow so a duration mismatch can be traced to the frame batch, rate input, or audio connection.
| Check | Hypothetical record | What to verify |
|---|---|---|
| Source frames | 240 | Count frames intended for assembly; update this if processing changed the batch. |
Extracted fps | 24 | Record the value emitted by Get Video Components. |
Create Video fps input | Connected to extracted fps (24) | Confirm the input is wired, rather than relying on the 30 fps default. |
| Expected picture duration | 240 ÷ 24 = 10 seconds | Compare with the resulting video’s picture duration. |
| Audio connection | Connected, if source sound is required | Play the result or inspect it in a player that reports audio tracks. |
Treat picture timing and sound as separate acceptance checks. A duration close to the estimate does not prove that audio is present; an audible soundtrack does not prove that the picture uses the intended frame rate.
Troubleshoot mismatches and limits
If the output is unexpectedly short or fast, count the processed images and inspect the value at Create Video’s fps input. A common hypothetical cause is leaving that input at its 30 fps default when the extracted source rate is 24 fps. For 240 frames, that changes the estimated picture duration from 10 seconds to 8 seconds. This example illustrates the arithmetic; it is not a claim about a tested export.
If the picture duration is plausible but the soundtrack is missing, check whether the audio output was connected to Create Video’s optional audio input. If the sound is present but ends before or after the picture, compare its timing with the assembled video; the cited node documentation does not specify how mismatched audio and picture durations are reconciled. Do not assume the node will stretch, trim, or otherwise repair the audio automatically.
This workflow preserves the extracted rate as an input and gives you a way to check the result; it does not restore frames removed during processing or guarantee every source property. The documented Create Video inputs and outputs here do not establish that every container, codec, or playback application will report timing identically. If the frame count or rate falls outside the documented input range, or the output differs from your intended delivery, verify the actual values and consult the settings and format requirements of the next tool in your pipeline.
Does `Create Video` preserve the source fps automatically?
Do not rely on that. The node has its own fps input and documents a 30 fps default. Connect the extracted fps output when the goal is to reuse the source rate.
Will the audio be included if I only connect images and fps?
The audio input is optional, so image and frame-rate connections alone do not provide the extracted soundtrack. Connect the audio output when the result must include it, then verify that the exported video actually has an audio track.
Why can the duration still differ after connecting the original fps?
The processed frame count may differ from the original, or the value at the Create Video input may not be the one you recorded. Recalculate using the number of frames actually assembled and check the output duration. If those match but the soundtrack’s endpoint differs, investigate audio timing separately rather than changing the picture frame rate.
Sources and next step
- GetVideoComponents - ComfyUI Built-in Node Documentation
- CreateVideo - ComfyUI Built-in Node Documentation
This guide interprets public documentation and uses explicitly hypothetical acceptance examples, not application tests. Controls and availability can change; verify the cited manual against your installed version.
For work created in Imagild, carry the input requirements and acceptance checklist into the film workspace. The external software controls discussed above belong to their respective products.