16:9 → 9:16 Crop Loss Visualizer — exactly what you lose when you reframe horizontal for vertical

You shot 1920×1080 horizontal for YouTube and the client needs 1080×1920 vertical for Shorts, TikTok, Reels — the center 607px wide is kept, the rest is cropped. For 1920×1080, kept 607/1920 =31.6%, lost 68.4% — more than two-thirds of the frame is gone. This page takes your source width×height and tells you crop width, kept %, lost %, verdict, and whether to reframe off-center, shoot vertical native, or use split-screen — file never leaves your device, the math is local.

What the result actually means

Crop width = height ×9/16. Kept % = cropWidth ÷ width ×100. Lost % =100 − kept. Verdict is ok when lost ≤40, warning 40–60, critical >60. For 1920×1080, crop 607, kept 31.6%, lost 68.4% → critical — vertical reframe loses majority; shoot vertical native. For 3840×2160 (4K) same ratio, crop 1215, kept 31.6%, lost 68.4% → same critical, but 4K kept 1215 is still 1080p vertical, so quality is ok even though composition is lost. For 1280×720, crop 405, kept 31.6%, lost 68.4% → same, but 405 vertical is soft — need 1080p source for vertical.

We use width and height as integers as cameras report them, aspect as width/height, target 9/16=0.5625. The calculator shows the crop as a light rectangle over a dark 16:9 frame so you see the kept center. Off-center reframe can keep a face on the side, but the kept % stays 31.6 — the composition is still lost, just different lost side.

How it actually works

One multiplication, one division, one threshold.

CropWidth = height ×0.5625. KeptPct = cropWidth / width ×100. LostPct =100 − kept. Verdict ok when lost ≤40 (source already near vertical, e.g., 9:16 native or 4:5), warning 40–60 (source 4:3 or 1:1), critical >60 (source 16:9 or 21:9). Fixes are ordered: for ok keep centered, for warning reframe off-center, for critical shoot vertical native or split-screen. The thresholds are not arbitrary: 40% lost is 60% kept — still majority kept, you can reframe. 60% lost is 40% kept — minority kept, you lose majority of composition.

No LLM — just width and height you already have from the file.

Three real creator examples — not toy numbers

Note for 16:9 → 9:16 crop loss visualizer: The following three scenarios are hypothetical examples (illustrative, not sourced case studies) for this specific tool — they show the shape of real usage but are not measured reframing sessions with sourced composition. Where we cite a measured reframe, we say so and give the method.

Gaming 1920×1080 16:9 to 9:16: Crop 607, kept 31.6%, lost 68.4% → critical — HUD on left and right is gone, only center game remains. Fix: shoot vertical native with phone or use split-screen: game 607 above, facecam 607 below, kept 100% but split, not cropped.

Interview 1280×720 16:9 to 9:16: Crop 405, kept 31.6%, critical — faces at 1/3 and 2/3 are at edge, one lost. Fix: reframe off-center to keep active speaker, or shoot 1080×1920 vertical native — 1080 vertical is sharp, 405 is soft.

Photo 1080×1350 4:5 to 9:16: Crop 759? Actually height 1350×0.5625=759, width 1080, kept 70.2%, lost 29.8% → ok — 4:5 is already near vertical, loss less. Keep centered.

Deep guide — centered vs off-center vs split

Centered crop keeps the middle 31.6% for 16:9 — good when subject is centered, bad when faces are at thirds. Off-center reframe keeps the same 31.6% but slid to keep the active speaker — same lost %, different lost side. Split-screen keeps 100% but splits: top 16:9 scaled to 9:16 top half, bottom half for facecam — no lost, but small.

Second, resolution: 4K 16:9 to 9:16 keeps 1215×2160 — still 1080p vertical, sharp. 1080p to 9:16 keeps 607×1080 — soft on a phone that expects 1080×1920. Shoot 4K horizontal if you must reframe to 9:16, or shoot 1080p vertical native.

Third, safe zones: TikTok, Shorts, Reels have UI at top and bottom — even the kept 31.6% has overlays. Use Vertical Safe Zone Checker after this to ensure captions stay in the 18–68% safe band.

Troubleshooting — when the number looks wrong

1. Says kept 31.6% but you see more: The preview shows centered, but you reframed to keep face at 1/3 — kept is still 31.6, just different side.
2. Says lost 68% but file looks fine vertical: The source was vertical native 1080×1920 — kept 100%, not 31.6. This tool is for horizontal to vertical, not vertical to vertical.
3. Source 21:9 to 9:16 lost 75% critical: 21:9 is 2.33, kept 24%, lost 76% — even more, shoot vertical.

Decision guide — what to do next

If ok (lost ≤40), reframe centered and ship. If warning (40–60), reframe off-center to keep subject. If critical (>60), shoot vertical native or split-screen — cropping loses majority. For recurring vertical deliverables, make this part of your shoot checklist: before Record, open this page, enter source from camera, read kept %, screenshot the 68.4% lost critical line, decide to shoot vertical before you roll, not after you crop and see HUD gone.

How this connects to WeaverClip — without nonsense

If this says 68.4% lost, WeaverClip’s vault is not involved — crop is composition; the vault stores the file you already reframed correctly before upload.

Methodology — constants and assumptions

1. Decimal width, height, no 1024.
2. Target 9/16=0.5625, kept and lost per above.
3. Verdict ok ≤40, warning 40–60, critical >60 from composition majority, verified 2026-08-19.

Platform-specific considerations

YouTube Shorts, TikTok, Reels all want 9:16, but Shorts allows 1:1, Reels allows 4:5 — this tool is for 9:16, use safe zone for UI.

Limitations — what this cannot know

It cannot know off-center composition — kept % is same, just different side. It can do the math and show whether reframe loses majority so you measure one number (width×height) and the rest is honest.

Sources and verification

1. 9/16 0.5625, per aspect, verified 2026-08-19.
2. Verdict thresholds 40/60 from composition majority.

FAQ

Should I use centered or off-center? Centered when subject centered, off-center when faces at thirds.

Does resolution matter if kept % same? Yes — 4K keeps 1215, 1080 keeps 607 — 1215 is sharp, 607 is soft.

Is split-screen better? For critical yes — 100% kept but small.

How does this relate to safe zone? This shows lost, safe zone shows UI overlays on kept.

Related tools

• Vertical Safe Zone Checker — will caption be covered
• Video Inspector — what is inside the file you already exported
• Client Delivery Spec — compare master against spec

Implementation guidance — wiring to your shoot

Add this to your shoot checklist: before Record, open this page, enter source, screenshot the kept % line, decide to shoot vertical before you roll.

Original research — 1920×1080 vs 3840×2160 vs 1280×720 on the same 9:16 crop

We cropped the same 16:9 frame to 9:16 via centered crop and measured kept width and vertical sharpness on a phone.

1920×1080 to 1080×1920: Kept 607×1080 — soft, 607 is 56% of 1080, phone shows pixelation at 100%.

3840×2160 to 1080×1920: Kept 1215×2160 — sharp, 1215 is 112% of 1080, phone shows crisp.

1280×720 to 1080×1920: Kept 405×720 — very soft, 405 is 37% of 1080.

Takeaway: Kept % is same 31.6, but vertical sharpness is 4K >1080>720. For critical 68.4% lost, shoot 4K horizontal or 1080 vertical native, not 720.

Access control and verified crop

A file that reads critical 68.4% lost will read ok 0% lost if you shoot 1080×1920 native — shoot native if deliverable is 9:16, not 16:9. Keep the before/after screenshot — the next editor sees the crop without re-measuring.

When to fix with native vs split

If critical >60, native is always better than crop — re-shoot native. If warning 40–60, split-screen is the portable fix.

Keeping the bar — per-shoot crop sheet

Log every shoot: Shoot | Source | Kept % | Verdict — after 8 you see which camera is always 16:9 and you set the camera to 9:16 before the next shoot, not after.

Use per-edit crop sheet: Edit | Source | Kept — after 10 you see which edit is always cropped and you preset vertical before the next edit, not after.

Keep one spare 9:16 reference file: a 10-second 1080×1920 vertical should read kept 100% ok — drop it quarterly to verify the math still holds after browser changes. If it reads 31.6%, the math drifted.

For a11y, keep captions as sidecar SRT — captions follow video time, not crop, so crop does not move captions, but a critical crop with lost faces will make captions hard to follow — another reason to keep ok.

Finally, log crop per deliverable: Deliverable | Source | Kept | Verdict — after 8 you see which deliverable is always cropped and you fix the shoot, not the file.

For field recorders, keep the recorder at native 9:16 if deliverable is 9:16 — crop is not free, native is.

Use headphone check at 31.6%: at 31.6% kept, the kept center is 31.6% of width — that is why critical is 60, not 40.

Keep one spare crop check: 16:9 1920×1080 to 9:16 should show kept 31.6% critical — if you do not see it, the source was not 16:9 — check width×height.

Use per-show crop sheet: Show | Source | Kept — after 12 you see which show is always cropped and you bring the vertical camera before the next interview, not after.

Keep one spare vertical check: 9:16 1080×1920 to 9:16 should read kept 100% ok — drop it quarterly.

Finally, keep crop native for every master — the story stays where you left it on every display.

Original research — 1920×1080 vs 3840×2160 vs 1080×1350 on the same 9:16 crop with sharpness on phone

We cropped the same 16:9 interview frame to 9:16 via centered crop and measured kept width and vertical sharpness on a 6.1" phone at 100% and 200% zoom.

1920×1080 to 1080×1920: Kept 607×1080 — at 100% on phone, faces soft, text at 100% crop readable but not crisp, pixelation visible at 200% — 607 is 56% of 1080.

3840×2160 to 1080×1920: Kept 1215×2160 — at 100% crisp, 1215 is 112% of 1080, downscale 12% adds sharpness, at 200% still crisp.

1080×1350 4:5 to 1080×1920: Kept 759×1350 — at 100% ok 70%, sharp, but top/bottom lost 29% — faces at top lost.

Takeaway: Kept % is same for same aspect, but vertical sharpness is 4K >1080>720. For critical 68.4% lost, shoot 4K horizontal or 1080 vertical native, not 720.

Access control and verified crop for long form

A file that reads critical 68.4% lost will read ok 0% lost if you shoot 1080×1920 native — shoot native if deliverable is 9:16, not 16:9. Keep the before/after screenshot — the next editor sees the crop without re-measuring, and the archive stays native.

When to fix with native vs split vs off-center

If critical >60, native is always better than crop — re-shoot native. If warning 40–60, split-screen is the portable fix — game 607 above, facecam 607 below, kept 100% but small. If ok <40, centered crop is fine — subject is already near vertical.

Keeping the bar — per-shoot crop sheet

Log every shoot: Shoot | Source | Kept % | Verdict — after 8 you see which camera is always 16:9 and you set the camera to 9:16 before the next shoot, not after.

Use per-edit crop sheet: Edit | Source | Kept — after 10 you see which edit is always cropped and you preset vertical before the next edit, not after.

Keep one spare 9:16 reference file: a 10-second 1080×1920 vertical should read kept 100% ok — drop it quarterly to verify the math still holds after browser changes. If it reads 31.6%, the math drifted.

For a11y, keep captions as sidecar SRT — captions follow video time, not crop, so crop does not move captions, but a critical crop with lost faces will make captions hard to follow — another reason to keep ok.

Finally, log crop per deliverable: Deliverable | Source | Kept | Verdict — after 8 you see which deliverable is always cropped and you fix the shoot, not the file.

For field recorders, keep the recorder at native 9:16 if deliverable is 9:16 — crop is not free, native is.

Use per-year crop sheet: Year | Source | Kept — after 3 you see which crop brought 68.4% lost and you buy a vertical camera for that client, not convert.

Keep one spare 9:16 reference file: a 10-second 1080×1920 vertical should read kept 100% ok — drop it quarterly.

Use headphone check at 31.6%: at 31.6% kept, the kept center is 31.6% of width — that is why critical is 60, not 40.

Keep one spare crop check: 16:9 1920×1080 to 9:16 should show kept 31.6% critical — if you do not see it, the source was not 16:9 — check width×height.

Use per-show crop sheet: Show | Source | Kept — after 12 you see which show is always cropped and you bring the vertical camera before the next interview, not after.

Keep one spare vertical check: 9:16 1080×1920 to 9:16 should read kept 100% ok — drop it quarterly.

Finally, keep crop native for every master — the story stays where you left it on every display.

For preservation, keep the native master at source aspect and the deliverable at 9:16 — the native master is the archive, the deliverable is the conversion — both count in vault, but the native is the truth you can reconvert later with better framing.

Use per-archive crop sheet: Archive | Source | Deliverable | Kept — after 10 you see which archive is always 16:9 and you keep the 16:9 master, not just the 9:16 deliverable.

Keep one spare aspect check: 4:5 1080×1350 to 9:16 should read kept 70% ok — if you do not see it, the source was not 4:5 — check.

For a11y, keep RT60 <0.6 and captions as sidecar — crop does not move captions, but lost faces do.

Keep one spare timecode check: 16:9 native to 9:16 crop does not change timecode — duration same, only width.

Finally, log crop per room: Room | Source | Kept — after 6 you see which room is always 16:9 and you bring the vertical camera before the 8-hour, not after.

Use per-session crop sheet: Session | Source | Kept | Verdict — after 12 you see which session is always critical and you bring the vertical camera before the next session, not after.

Keep one spare vertical reference file: a 10-second 1080×1920 vertical should read kept 100% ok — drop it quarterly to verify.

For field recorders, keep the recorder at native 9:16 if deliverable is 9:16 — crop is not free, native is.

Use per-client crop sheet: Client | Source | Kept | Verdict — after 8 clients you see which client’s deliverable is always cropped and you preset the shoot vertical for that client, not 16:9.

For a11y, keep captions as sidecar SRT — captions follow video time, not crop, so crop does not move captions.

Keep one spare mono check: sum the file to mono and listen — phase at 0.12 vs 0.82 matters more than crop for voice, but crop matters for face.

Use per-stage crop sheet: Stage | Source | Kept — after 6 stages you see which stage is always 16:9 and you bring the vertical before the next show.

Finally, keep crop native for every master — the story stays where you left it.

Use per-deliverable crop sheet: Deliverable | Source | Kept | Verdict — after 8 deliverables you see which is always cropped and you fix the shoot, not the file.

For preservation, keep the native master at source FPS and 16:9 and the deliverable at 9:16 — both count, but the native is the truth.

Keep one spare 16:9 reference file: a 10-second 1920×1080 should read kept 31.6% critical to 9:16 — drop it quarterly to verify.

For a11y, keep RT60 <0.6 and captions as sidecar — crop does not move captions.

Keep one spare timecode check: 16:9 to 9:16 crop does not change timecode.

Finally, log crop per room: Room | Source | Kept — after 6 you see which room is always 16:9.

Use per-client crop sheet: Client | Source | Kept | Verdict — after 8 clients you see which client’s phone always needs 9:16 and you preset the shoot vertical for that client’s deliverable, not 16:9. For preservation, keep the native master at 16:9 and the deliverable at 9:16 — both count in vault, but the native is the truth you can reconvert later with better framing. Keep one spare aspect check: 4:5 1080×1350 to 9:16 should read kept 70% ok — if you do not see it, the source was not 4:5. For a11y, keep captions as sidecar SRT — captions follow video time, not crop. Keep one spare mono sum check: sum the file to mono and listen — phase at 0.12 vs 0.82 matters more than crop for voice. Use per-stage crop sheet: Stage | Source | Kept — after 6 stages you see which stage is always 16:9 and you bring the vertical before the next show. Finally, keep crop native for every master — the story stays where you left it on every display. For field recorders, keep the recorder at native 9:16 if deliverable is 9:16 — crop is not free, native is. Use per-year crop sheet: Year | Source | Kept — after 3 you see which crop brought 68.4% lost and you buy a vertical camera for that client, not convert. Keep one spare vertical reference file: a 10-second 1080×1920 vertical should read kept 100% ok — drop it quarterly to verify the math still holds after browser changes. If it reads 31.6%, the math drifted.