The US visa and Diversity Visa photo specification, explained

The rhythm people remember is an entry window opening in early October, results the following spring. That is not what happened in the last cycle. The State Department deferred the DV-2027 entry period rather than opening it on the traditional schedule, and DV-2027 selection opened on 11 March 2026.

Separately, requirements that took effect on 10 April 2026 changed what submitting an entry involves. Entrants are required to upload the biographic page of a valid passport, and to pay a fee of one US dollar which is not waivable. Any walkthrough written before April 2026 describes a form that no longer matches the one you will be filling in.

As of late August 2026, no DV-2028 entry window has been announced. That is the complete and honest answer. Nobody outside the State Department knows when the next one opens, and any site that publishes a date is either guessing or trying to sell you something. Check travel.state.gov and dvprogram.state.gov, and treat every other source on timing as noise.

None of this touches the photo specification, which has been stable throughout. Preparing a compliant photo now is not wasted effort. It is the same photo an actual visa application will need later.

Here is the whole digital image requirement, with what each line is actually testing. If anything below ever drifts out of date, the version that counts is the one the State Department publishes on its digital image requirements page:

https://travel.state.gov/content/travel/en/us-visas/visa-information-resources/photos/digital-image-requirements.html

Read the list quickly and you conclude that the maximum size is the safe choice. More pixels, better photo, and 1200 x 1200 is explicitly allowed. That instinct is backwards, and the reason is the compression ratio.

A compression ratio compares the size of the finished file against the size the same image would occupy with no compression at all. At 24 bits per pixel, every pixel costs three bytes uncompressed. So an image of 600 x 600 is 360,000 pixels, or 1,080,000 bytes raw. A ratio of 20:1 means the JPEG may not be smaller than a twentieth of that, which is 54,000 bytes, or roughly 54 KB.

In other words the specification has a floor as well as a ceiling. Your file has to be under 240 KB and above the number your chosen dimensions imply. Run that arithmetic across the permitted range and the shape of the problem becomes obvious:

At the maximum dimensions the two limits nearly meet. You are trying to land a JPEG in a band about twenty-odd kilobytes wide, and the exact width of that band depends on whether the checker treats 240 KB as 240,000 bytes or 245,760. That is a needlessly tight target to hit by hand.

At 600 x 600 the same photo has a window of well over a hundred kilobytes, and a normally exported JPEG of a head-and-shoulders portrait lands inside it without any effort at all. The face is not meaningfully better identified at 1200 pixels than at 600. Nothing rewards you for using the top of the range.

So the default worth adopting is: crop square, resize to exactly 600 x 600, export as JPEG at a good quality setting, and check that the result is somewhere between about 60 KB and 240 KB. If it comes out at 300 KB, reduce quality slightly rather than reaching for the smallest file you can make.

The specification also accepts a scanned print, and states the terms precisely: a photo of 2 x 2 inches, which is 51 x 51 mm, scanned at 300 pixels per inch. Multiply it out and two inches at 300 ppi gives exactly 600 pixels, so a correctly scanned print lands precisely on the minimum digital size. The two halves of the rule are the same rule.

That also means scanning a 2 x 2 print at 150 ppi gives you 300 x 300, which is below the minimum and cannot be rescued. Set the scanner to 300 ppi or higher, scan in colour, and save as JPEG rather than PDF or TIFF.

What you must not do is enlarge a small image to reach 600 x 600. Upscaling invents pixels by averaging the ones already there. The file passes the dimension check and looks soft and smeared to anyone examining it, which is a slower and more frustrating way to fail. If the original is 400 x 400, the answer is a new photograph, not a resize.

Failures split into two groups. The technical ones are the ones you can verify yourself in under a minute, and they are the more common cause of an entry being thrown out because they are invisible to the person submitting.

The second group is everything about the photograph itself: how large your head appears in the frame and where it sits, your expression, the background behind you, whether you are wearing glasses, and the narrow circumstances in which a head covering is acceptable. These are real requirements and they reject plenty of entries.

I am not going to put numbers on any of them here, because guides that do are frequently wrong and the consequence of following a wrong number is losing an entry. The State Department publishes the composition rules alongside worked examples of acceptable and unacceptable photos, and that page is the only version worth trusting. Read it before you take the picture rather than after.

In general terms, the photo should be recent and should look like you do now, taken against a plain, light, unpatterned background with even lighting and no strong shadow behind your head. Face the camera squarely with a neutral expression or a natural smile, with both eyes open and your whole face visible. Glasses are generally not acceptable. Head coverings are accepted only for religious reasons and must not obscure the face. No filters, no beauty smoothing, no retouching, and no cropping somebody else out of a group photo.

Take the photo in daylight against a plain wall, with the camera at eye level and roughly an arm and a half away rather than at arm's length, which distorts faces. Have somebody else hold the camera if you can. Then work in this order, because the steps interfere with each other if you do not.

Crop to a square first, framing the head and shoulders as the official examples show. Resize that square to 600 x 600. Convert to sRGB if the source was a phone. Export as JPEG. Look at the resulting byte count and confirm it sits between roughly 60 KB and 240 KB, and only then adjust quality if it does not.

Doing it in that order matters. If you compress to hit 240 KB and then resize, the resize changes the file size and you have to start again. If you compress before cropping, you spend bytes on background you are about to delete anyway. Crop, resize, convert, export, check.