Batch Resize Photos for the Web Without Losing Quality

Published 19 October 2026 · by Tekiba

One longest edge for the whole batch

A batch from a phone or camera mixes landscape and portrait frames. Set a width and the portrait ones come out taller than the rest; set a longest edge and both fit the same box. At 1600, a 4032×3024 photo comes out at 1600×1200 and a 3024×4032 one at 1200×1600. Both edges scale by the same factor, so nothing is stretched, and nothing is ever enlarged: a photo already smaller than the number stays as it is, because a bigger copy of a small photo only adds blur.

Pick the number from where the photo will be shown, not from the camera. A photo much larger than its place on the page is extra download for nothing; one much smaller looks soft on a sharp screen. Some sites name their size: Shopify suggests 2048×2048 for square product photos, and WordPress makes a -scaled copy of anything over its threshold. Our bulk resize and rename guide covers the other modes and the naming patterns.

What the quality setting does

JPEG and WebP, as most photos use them, are lossy: they keep a file small by throwing away detail the eye is least likely to miss. The quality setting tells the encoder how much it may throw away. MDN's reference for the browser call that SignetDrop uses describes it as

A Number between 0 and 1 indicating the image quality to be used when creating images using file formats that support lossy compression

The number is not a percentage of the detail kept, and the same number gives different results in different encoders and browsers. In SignetDrop the slider runs from 30 to 100 and starts at 86, which keeps photos clean at a modest size. Lower settings shrink the file further and first show their cost in smooth areas: a sky turning into blocks, a gradient into bands, fine edges ringed with noise. The way to choose is to try one photo at two settings and compare them at full size, not as thumbnails.

JPEG, WebP or PNG

MDN's guide to image file types puts the general rule this way:

Photographs and other continuous-tone images can usually tolerate lossy compression, because the discarded detail is harder to see.

JPEG is lossy, so it is often a good fallback format for photographs

Google's WebP page, read on the publication date above, gives its own comparison:

WebP lossless images are 26% smaller than PNGs and lossy images are 25-34% smaller than comparable JPEGs at equivalent quality.

Those are Google's own figures for its own format; your photos will vary. In practice:

• JPEG for anything that leaves your hands: a marketplace upload, an email, a client. Every system takes it.
• WebP for photos on your own site, where you know the pages load it. SignetDrop offers WebP only in a browser that can save it, and says so in the list where it cannot.
• PNG for screenshots and graphics with sharp text, not for photos: it keeps every pixel, so a photo comes out much larger than a good JPEG. It has no quality setting, and the slider is hidden when you choose it.

Where quality is really lost: copies of copies

A lossy save throws detail away, and the next save of that result throws away more, on top of what was already gone. One save at a sensible setting rarely shows; a photo that has been through a messaging app, downloaded from a website and then resized again often does. So:

• Start from the original from the camera or phone, not from a copy you saved off a site or received in a chat.
• Do the resizing, the watermark and the format change in one pass. SignetDrop draws each photo once and saves it once, so the three cost one save, not three.
• Keep the originals. A web copy is for showing; if you need a different size next month, make it from the original again.

Resizing down also discards pixels, but in a controlled way: the browser averages them with its high-quality smoothing, which is why a well-made small copy looks sharp. The loss that shows is mostly the repeated compression, not the resizing.

Resizing a batch in the browser

1. Drop the originals in. Open SignetDrop and add the photos. They are read in the page and never uploaded.

2. Set the size. Under Size and format, set Resize to Longest edge and type the number, 1600 for instance. The free version stops at 2048: at that setting a 4032×3024 photo becomes 2048×1536.

3. Pick the format and quality. JPEG or WebP, and leave the quality at 86 to start. Try one photo at a lower setting if you need smaller files.

4. Add a watermark or a naming pattern if you want them, in the same pass.

5. Download the ZIP. The copies are new files written from the canvas, so they carry no EXIF: no GPS position, no phone model. That is usually what you want on the web; our guide to whether resizing removes EXIF covers the paid option that keeps it, JPEG to JPEG.

What this does not do

A smaller copy holds less detail than the original, and no setting changes that. SignetDrop does not sharpen, enlarge, or recover detail a compressed copy has already lost, and it does not convert HEIC where the browser cannot read it. It does not crop either; on the paid version it can pad a photo to a shape. Output above 2048px and larger batches are in the paid version, $29 once, with the full split on the pricing panel, and SignetDrop never writes its own name onto your photos.

Questions