How much of the finished picture will be invented?

About sixty-eight percent, for an ordinary phone portrait. That figure is fixed by geometry rather than by anything you choose: a 9:16 photograph is a little over three times taller than it is wide relative to a 16:9 frame, so roughly two pixels are built for every one you brought. The percentage is printed under the canvas the whole time you work, and it is the number that tells you whether the shape was a reasonable thing to ask of this particular photograph.

Is the blur going to look cheap?

It looks like what people have been shown a hundred thousand times, which is a different thing. Every video platform and half the social apps pad vertical footage this way, so the visual language is established and reads as a frame rather than as a fault. What looks cheap is a blur at too small a radius, where recognisable shapes survive behind the picture and the whole thing seems half-finished. Push the radius up until the backdrop is fields of colour rather than objects.

What if I need 16:9 for a video timeline?

Then a flat near-black is often the better choice, because your still will sit next to footage that is genuinely black-bordered and a blurred still will stand out from it. Two clicks: One colour, then the darkest swatch. The same applies in reverse for a bright white feed, where near-black bars read as a mistake and a blur reads as a decision.

Can I do this the other way, 16:9 back to 9:16?

Yes, and the rule is the same rule: the axis that is short grows. A wide photograph asked for 9:16 gains height above and below rather than losing its sides. It is a gentler conversion than people expect, because a landscape photo usually has sky at the top and ground at the bottom, and both of those extend more convincingly than the middle of a scene would.

Guide

9:16 to 16:9, and why the answer is usually a blur

A tall phone photo and a wide frame have almost nothing in common, which is why this conversion asks more of the fill than any other — and why one fill keeps winning.

The numbers first, because they explain everything else

An ordinary phone portrait is 1,080 across and 1,920 down. To be 16:9 at that height it has to be 3,413 across. So 2,333 columns arrive from somewhere — 1,166 on the left, 1,167 on the right, when the photograph is centred.

That is the whole story of this conversion. You are not adding a tasteful border; you are building something wider than the photograph itself, twice over, and then putting the photograph in the middle of it. The alternative is cutting 1,312 rows off a 1,920-row picture, which is usually a head and a pair of feet.

Common phone sizes taken to a 16:9 canvas
Photograph	Becomes	Built here
1,080 x 1,920	3,413 x 1,920	68%
1,284 x 2,778	4,939 x 2,778	74%
3,024 x 4,032	7,168 x 4,032	58%
1,170 x 1,560	2,773 x 1,560	58%

Notice the last two. A photograph that is already less extreme than 9:16 — a 3:4 camera frame, say — needs far less inventing, which is worth knowing if you have any choice about what you shoot for this.

Why the blur wins at this ratio

At smaller conversions the choice of fill is a matter of taste. At this one it is a matter of structure, because the fill is now the majority of the image and every weakness it has is magnified accordingly.

A mirror across 1,166 pixels is not one reflection; it is a reflection and then most of another, so the repeat is unmistakable and the picture reads as a pattern. An edge stretch across that distance turns a single column of pixels into a third of the frame, which is convincing on sky and nothing else. A flat colour is perfectly respectable and genuinely right for some destinations, but at this width it reads as a photograph in a box rather than a wide picture.

The blurred cover has one property the others lack: it makes no claim. There is no structure in it to repeat, no line to streak, no edge to argue with the photograph, so there is nothing for a viewer to catch on. It is also, and this matters, the thing people already expect to see behind a vertical photo in a wide frame.

What happens, and in what order

Open the photograph

Drop, choose or paste the vertical photo. It is read and turned upright by this page before anything is measured, and it is never sent anywhere.

Confirm the target is 16:9

Press 1, or click the 16:9 chip. The canvas is measured immediately: the height stays as it was and the width grows around the photograph.

Choose the fill

Press B for the blurred cover, which is the default and the right answer for most photographs. Press M for a mirror if the edges are texture rather than objects, or C for one flat colour.

Adjust radius and tone

Raise the blur radius until no shape behind the picture is recognisable, then take the tone down ten or fifteen percent so the sharp photograph separates from its own backdrop.

Place the subject

If your subject sits near an edge, nudge the photograph with the arrow keys, or pick a different anchor point, so the seam falls somewhere quiet.

Save it

Choose JPG for a photograph, then press the save button. The full-resolution file is painted by the same routine that drew the preview and handed straight to your browser's downloads.

Where this goes wrong

The commonest failure is a face near the edge of the original. The blurred backdrop is your photograph enlarged, so a face near the left edge becomes a large soft face behind and beside the sharp one, and the eye finds it immediately. The fix is placement: move the photograph off centre with the arrow keys, or pick a different anchor, until the enlarged copy puts that shape somewhere harmless.

The second is a radius left too low. A blur that still has recognisable objects in it reads as a rendering failure rather than a backdrop. Turn it up until you cannot name anything behind the picture.

The third is not a failure of the tool at all. Some photographs are simply not a 16:9 picture, and no fill rescues that — a tight portrait, a tall building, a document. Expanding those gives you a correctly shaped file containing a small photograph, which may be exactly what a platform requires and is never the same thing as a good wide picture. Knowing which of the two you have got is the whole skill here.

The awkward questions

How much of the finished picture will be invented?: About sixty-eight percent, for an ordinary phone portrait. That figure is fixed by geometry rather than by anything you choose: a 9:16 photograph is a little over three times taller than it is wide relative to a 16:9 frame, so roughly two pixels are built for every one you brought. The percentage is printed under the canvas the whole time you work, and it is the number that tells you whether the shape was a reasonable thing to ask of this particular photograph.
Is the blur going to look cheap?: It looks like what people have been shown a hundred thousand times, which is a different thing. Every video platform and half the social apps pad vertical footage this way, so the visual language is established and reads as a frame rather than as a fault. What looks cheap is a blur at too small a radius, where recognisable shapes survive behind the picture and the whole thing seems half-finished. Push the radius up until the backdrop is fields of colour rather than objects.
What if I need 16:9 for a video timeline?: Then a flat near-black is often the better choice, because your still will sit next to footage that is genuinely black-bordered and a blurred still will stand out from it. Two clicks: One colour, then the darkest swatch. The same applies in reverse for a bright white feed, where near-black bars read as a mistake and a blur reads as a decision.
Can I do this the other way, 16:9 back to 9:16?: Yes, and the rule is the same rule: the axis that is short grows. A wide photograph asked for 9:16 gains height above and below rather than losing its sides. It is a gentler conversion than people expect, because a landscape photo usually has sky at the top and ground at the bottom, and both of those extend more convincingly than the middle of a scene would.