16:9 to 9:16 Crop Loss
A 16:9 frame cropped to 9:16 keeps 31.64% of its pixels and loses 68.36%. The 1920 × 1080 original becomes 608 × 1080, which trims 656 pixels from the left edge and 656 from the right. The whole-pixel frame keeps 31.67% and loses 68.33%; the difference is the half pixel in 607.5.
Direct answer: 68.4% of the area is cut awayWhy the number is 68.36% and not something else
Both frames share one dimension and give up the other. A 16:9 source is wider than a 9:16 target, so the height survives at 1080 pixels and the width has to be rebuilt from the target ratio: 1080 × 9/16 = 607.5, rounded to 608 pixels.
That leaves 608 × 1080 = 656,640 pixels out of 1920 × 1080 = 2,073,600. The kept fraction is 656,640 / 2,073,600 = 0.316406, and 1 − 0.316406 = 0.683594.
The 0.36% gap between 68.36% and the tidy 68.4% is rounding. The exact width is 607.5 pixels, and a frame cannot hold half a pixel. Round up to 608 and the kept area rises to 656,640 pixels, or 31.67%, which is the 68.33% the calculator above reports. Round down to 607 and it falls to 655,560 pixels, or 31.61%. The rounding decision is worth a third of a percentage point, so the calculator always states the integer it used.
The general formula
Let the source ratio be rs and the target ratio be rt, both written as width ÷ height.
loss = 1 − min(rs, rt) ÷ max(rs, rt)
The kept fraction is min(rs, rt) ÷ max(rs, rt). Nothing else enters the formula: not resolution, not pixel count, not aspect of the subject.
Two consequences follow straight from that line. First, the loss is symmetric: 16:9 to 9:16 costs exactly what 9:16 to 16:9 costs. Second, the loss depends only on how far apart the two ratios are, so a 4:3 to 5:4 crop costs less than a 3:2 to 1:1 crop even though both look like small adjustments.
All ten ratios, all hundred pairs
Every cell is the percentage of image area removed when the source in the left column is cropped to the target in the top row. Percentages come from the formula above and were checked against a real pixel crop at 1920 wide; the two methods agree to within 0.06 of a percentage point, which is rounding.
| From \ To | 16:9 | 4:3 | 3:2 | 5:4 | 1:1 | 4:5 | 9:16 | 21:9 | 2.39:1 | 2:1 |
|---|---|---|---|---|---|---|---|---|---|---|
| 16:9 | 0 | 25.0% | 15.6% | 29.7% | 43.8% | 55.0% | 68.4% | 23.8% | 25.6% | 11.1% |
| 4:3 | 25.0% | 0 | 11.1% | 6.2% | 25.0% | 40.0% | 57.8% | 42.9% | 44.2% | 33.3% |
| 3:2 | 15.6% | 11.1% | 0 | 16.7% | 33.3% | 46.7% | 62.5% | 35.7% | 37.2% | 25.0% |
| 5:4 | 29.7% | 6.2% | 16.7% | 0 | 20.0% | 36.0% | 55.0% | 46.4% | 47.7% | 37.5% |
| 1:1 | 43.8% | 25.0% | 33.3% | 20.0% | 0 | 20.0% | 43.8% | 57.1% | 58.2% | 50.0% |
| 4:5 | 55.0% | 40.0% | 46.7% | 36.0% | 20.0% | 0 | 29.7% | 65.7% | 66.5% | 60.0% |
| 9:16 | 68.4% | 57.8% | 62.5% | 55.0% | 43.8% | 29.7% | 0 | 75.9% | 76.5% | 71.9% |
| 21:9 | 23.8% | 42.9% | 35.7% | 46.4% | 57.1% | 65.7% | 75.9% | 0 | 2.4% | 14.3% |
| 2.39:1 | 25.6% | 44.2% | 37.2% | 47.7% | 58.2% | 66.5% | 76.5% | 2.4% | 0 | 16.3% |
| 2:1 | 11.1% | 33.3% | 25.0% | 37.5% | 50.0% | 60.0% | 71.9% | 14.3% | 16.3% | 0 |
Read the 9:16 column when you are cutting a vertical version for short-form video. Read the 16:9 row when you are delivering to a landscape player. The diagonal is zero because a frame cropped to its own ratio loses nothing.
What this costs in practice
If the subject sits in the middle third of the frame, a 16:9 to 9:16 crop can work: the 608-pixel column covers the central 31.6% of the width, and that is usually where the face is. If the subject sits near an edge, the crop has to move off centre, and the total trim does not change. Moving the crop only decides which side pays.
The loss is not recoverable by upscaling. A 608 × 1080 crop from a 1920 × 1080 master still has 608 real pixels across, so a 1080 × 1920 vertical export from that crop is being stretched by 1.78× on the horizontal axis. Recomposition at capture, or a second camera angle, is the only way to get a true 1080-wide vertical frame.
One shortcut that does work: shoot or render wider than the delivery ratio on purpose. A 4:3 or 1:1 master crops to 9:16 at a lower cost than 16:9 does, because the source is closer to square.
Crop loss questions
How much of a 16:9 video survives a 9:16 crop?
31.64% of the pixel area. A 1920 × 1080 frame becomes 608 × 1080, so 68.36% of the picture is discarded.
Is cropping to 9:16 the same as squeezing the frame?
No. A crop keeps the pixels that fall inside the new frame and drops the rest. Squeezing keeps every pixel and changes its shape, which distorts faces and circles. This page calculates the crop, not the squeeze.
Which conversion loses the least?
Among the common pairs, 21:9 to 2.39:1 loses 2.4% and 4:3 to 5:4 loses 6.3%. 16:9 to 4:3 loses 25.0%, and 9:16 to 21:9 loses 75.9%, the worst in the table.
Does a higher resolution reduce the loss?
No. The percentage is a ratio of areas, so a 4K master and a 1080p master lose the same share. The absolute number of pixels lost changes; the fraction does not.
How was the table verified?
Two independent methods. The formula gives one number per pair; a second pass built the actual integer crop from a 1920-wide source and counted the surviving pixels. The largest disagreement across all hundred pairs was 0.06 of a percentage point.
Method note
No outside data is used on this page. Every figure is arithmetic on the definition of a centred crop, so you can reproduce any cell with a calculator. The formula, the verification pass, the calculator above and the table are original work for this site.
Related
Chained crop loss: why 4:3 → 16:9 → 9:16 costs 76.3%, not 93.4%