If combing appears only in some parts of a video, do not assume every frame needs the same treatment. First confirm that moving edges really show comb-like lines; then compare HandBrake’s Deinterlace path with Decomb when you want processing to depend on comb detection. Judge the result in motion as well as on paused frames. A TV or DVD origin alone is not enough to decide.
Diagnose the artifact and choose a path
HandBrake documents Deinterlace as removing comb artifacts. Its summary describes Yadif as a popular, fast deinterlacer and Bwdif as motion-adaptive. It describes Decomb as switching among multiple interpolation algorithms for speed and quality. Therefore, compare the available Deinterlace and Decomb controls on the same excerpts, following the controls and behavior documented for the HandBrake version in use. Do not infer from this summary that Decomb selects individual frames according to comb detection; any such workflow assumption must be verified against the version-specific control description and preview.
Before choosing, separate combing from lookalikes. Inspect a moving edge at full size: alternating horizontal “teeth” can indicate combing, while block-shaped patches may be compression artifacts, and streaking or softness may be motion blur. These can coexist. Check a quiet frame too; damage that remains on a stationary edge is not necessarily caused by interlacing. The official HandBrake filter summary distinguishes Deinterlace and Decomb from Deblock, which reduces blocks from low-quality compression.
Use this ordered comparison:
- In HandBrake, open three short preview points from the source: one with mostly still fine detail, one with a horizontal moving edge, and one with fast motion. Include both a section where you suspect combing and one where it seems absent.
- Note the visible issue at each point before changing a filter. If the evidence is compression blocks, blur, or another distortion rather than comb-like edges, do not treat Deinterlace or Decomb as a general repair.
- Compare a Deinterlace option with Decomb on the same excerpts. Keep other settings and viewing conditions constant. The filter summary names Default and Bob tunings, and lists EEDI2 as a slower, higher-quality interpolation option for Decomb that may suit difficult sources; treat these as candidates, not guaranteed fixes.
- Play each preview at normal speed as well as pausing on representative frames. Record motion continuity, detail changes, and processing time. On a difficult source, compare the documented slower EEDI2 option with a faster candidate only if the first comparison leaves an actual problem to resolve.
HandBrake also documents Bob as an attempt to preserve motion better, with a slight penalty to perceived resolution. The official filter details do not provide a universal quality ranking for every source, so decide from your own excerpts rather than treating a filter name as a promise.
A three-excerpt acceptance worksheet
The following observations are hypothetical, not test results or expected HandBrake output. They illustrate how a creator preparing a 12-minute archive clip could record a decision without inventing a detection threshold. The times are chosen excerpts, not a prescribed sampling interval.
| Excerpt (hypothetical) | Observation before processing | Comparison observation | Decision to record |
|---|---|---|---|
| 00:42, still embroidered jacket | Fine horizontal stitching; no obvious teeth | Decomb preview looks similar; Deinterlace appears softer to this viewer | Preserve this detail as a reference point; do not infer that the entire clip is interlaced |
| 03:18, arm moving across frame | Repeated teeth on the arm edge, wall appears stable | Decomb preview reduces visible teeth; a paused frame still needs motion review | Favor the candidate that gives a single readable edge during playback |
| 08:51, fast pan past a fence | Teeth are hard to distinguish from blur in the source | Both previews remain ambiguous at normal speed | Mark the cause uncertain; compare another clear motion excerpt before choosing |
Write down which filter and tuning produced each preview, whether the feature is a HandBrake control or merely your viewing note, and what changed. A clean still alone is not acceptance: the motion excerpt should remain temporally coherent, and the still-detail excerpt should not lose more visible information than the job allows.
Limits and what to do if it fails
If neither path gives a clear result, choose a different excerpt with unmistakable motion and compare again under the same viewing conditions. If teeth remain, check whether the observed pattern is actually combing and whether the source is being interpreted appropriately before escalating. For a genuinely difficult source, EEDI2 is the documented slower, higher-quality Decomb interpolation candidate; compare its speed and visible result rather than assuming it will restore missing detail. If the evidence points to blocks or blur, investigate those artifacts separately instead of increasing deinterlacing strength.
Three quick questions
Does a DVD always need Deinterlace?
No. The source label does not establish that every frame has combing. Inspect representative motion and choose based on visible evidence.
Is Decomb always the better choice when only some scenes show teeth?
Not automatically. It is one documented interpolation path to compare with Deinterlace, but the filter summary does not establish that it performs frame-by-frame comb detection or prescribe it for every mixed source.
Should I judge by a paused frame or playback?
Use both. Pausing helps inspect edge detail; normal-speed playback reveals whether motion remains continuous and whether the apparent improvement holds over time.
Sources and next step
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.