There's a specific frustration that anyone who cares about photography or visual content has felt. You've taken a photo that looks genuinely great on your phone — crisp details, good colour, sharp edges. You spend ten minutes editing it. You upload it to Instagram. And then you look at it on the grid and something is wrong. It's not terrible. But it's noticeably softer than the version sitting in your camera roll. Like someone put a very thin layer of gauze over it.

For a long time I assumed this was a phone thing, or a connection thing, or maybe I was imagining it. Then I looked into it properly and found out this is completely deliberate. Every major social media platform recompresses your images after you upload them — reducing file size to manage the enormous storage and bandwidth costs of serving billions of images daily. The compression isn't random. Each platform uses specific algorithms, specific quality targets, and specific size thresholds that determine how much of your image quality survives the upload.

The good news: once you understand what each platform is doing, you can prepare your images to survive the compression with as little quality loss as possible. The fix isn't complicated. It mostly comes down to uploading images at the right dimensions and the right format — and understanding why the image you casually screenshot and upload looks so much worse than one you've properly prepared.

Why Social Platforms Compress Your Images at All

Before getting into the specific dimensions, it helps to understand why platforms do this — because understanding the reason makes the rules easier to remember.

Instagram alone hosts over 100 billion photos. Facebook serves somewhere north of 300 billion image views per day. At those scales, every unnecessary megabyte of storage and bandwidth is a real infrastructure cost. A raw photo from a modern smartphone is typically 4–12MB. Served at original quality to every person who scrolls past it, that adds up to infrastructure costs that are simply unviable for a platform serving billions of images daily.

So every platform compresses everything you upload. The compression aims to reduce file size dramatically — often to 200–500KB for a feed image — while keeping the visible quality degradation as minimal as possible. The algorithms they use are sophisticated, but they're working within hard constraints: the output file has to be small, and it has to happen automatically for every upload from every user in the world.

The problem is that these algorithms work better on some inputs than others. An image uploaded at the right dimensions, in the right format, with the right colour profile — the platform's compression touches it gently and the output still looks great. An image uploaded at the wrong dimensions — say, a 4000×4000px square photo when Instagram's feed displays at 1080×1080 — gets resized and recompressed, and each of those steps degrades quality. The platform isn't trying to make your images look bad. It's just that poorly prepared inputs produce worse compressed outputs.

What Compression Actually Does to Your Image

When Instagram or any platform compresses your image, a few specific things happen that explain the visual symptoms you see.

Downscaling

If your image is larger than the platform's maximum display size, it gets scaled down to fit. Downscaling done well is nearly invisible. Downscaling done by a fast, automated server-side algorithm is less careful. Details that exist at the original size — fine textures, hair strands, fabric patterns — can be blended together during the downscale and come out as a soft smear. This is the most common cause of the "blurry hair" or "fuzzy text" problem people notice after uploading.

JPEG recompression

Even if your image is already a JPEG, the platform recompresses it again at their target quality level. Recompressing a JPEG that's already compressed introduces what's called generation loss — the artefacts from the first compression get further distorted by the second. Edges that survived the first compression develop ringing. Areas of flat colour develop blockiness. The more times an image is JPEG-compressed, the worse it gets. Uploading a JPEG that's already been compressed twice before reaching the platform means you're starting from a worse baseline.

Colour profile stripping

Modern cameras and phone cameras often capture images in a wide colour space — Display P3 or AdobeRGB — that contains richer colour information than standard sRGB. Many platforms strip or ignore this colour profile and render the image as plain sRGB. If your editing app has been working in P3, the colours in the uploaded version can look duller or slightly shifted compared to what you see on your screen. This is especially noticeable in images with rich reds, deep blues, or vibrant greens.

Instagram: Exact Dimensions for Every Format in 2026

Instagram is the most compression-aggressive of the major platforms, and also the one with the most format variations — square, portrait, landscape, Stories, Reels covers, carousel — each with slightly different optimal dimensions. Here are the numbers that matter in 2026.

The Instagram quality setting that makes the biggest difference

Beyond dimensions, the single most effective thing you can do for Instagram image quality is upload as a JPEG at quality 85–95 rather than letting your phone's default export handle it. Most phones export sharing-optimised JPEGs at quality 70–80 to reduce file size for messaging. That quality level is fine for WhatsApp but is already at the bottom of what Instagram will preserve well. Start from quality 85+ and Instagram's recompression produces a noticeably better result.

The second thing: convert to sRGB before uploading. If you edit in Lightroom, Snapseed, or any app that works in a wide colour space, export specifically to sRGB. Instagram will convert anyway, but a controlled conversion during export is better than Instagram's automatic conversion, which can shift colours in unpredictable ways depending on the image content.

Facebook: Where the Rules Are Different Than You Think

Facebook's compression is generally less aggressive than Instagram's — partly because it serves a wider range of content types and device ages, which means it has to maintain broader compatibility. But it has its own quirks, and the right dimensions for Facebook are different from Instagram in ways that catch people out.

The PNG vs JPEG choice on Facebook

Facebook handles PNG and JPEG differently. For photographs — complex images with lots of colour variation — JPEG is fine and produces smaller, well-compressed output. For images that contain text, logos, sharp lines, or flat colour regions — promotional graphics, infographics, business card images — PNG produces dramatically better results on Facebook. JPEG's compression algorithm handles solid colours and sharp edges poorly, producing blocky artefacts around text that look terrible at small sizes. PNG's lossless compression preserves those sharp edges perfectly. The file is larger, but Facebook still displays it clearly and the text legibility is maintained.

LinkedIn: The Platform That Punishes Wrong Aspect Ratios Hardest

LinkedIn is where I see the most consistently bad image quality in professional contexts, which is ironic given that it's the platform where visual presentation arguably matters most. The problem is that LinkedIn's feed is dominated by images that were designed for other platforms and pasted in without resizing — Instagram-sized photos dumped into a LinkedIn post, profile banners stretched to the wrong shape, thumbnails that get awkwardly cropped.

Twitter/X and WhatsApp: The Ones People Get Most Wrong

Twitter/X's automatic crop preview is the most notorious source of image frustration on the platform. When you post a portrait image, Twitter shows a cropped preview in the feed — and its automatic crop algorithm is famous for making poor choices about which part of an image to show. This is especially bad for group photos, where the algorithm frequently focuses on the wrong person, and for promotional images where it cuts off the most important text.

The solution is to design for landscape from the start if the image is going to Twitter, or to use the platform's image editor to manually set the crop preview before posting. A 1200×675px landscape image fills the feed without any automatic cropping — what you upload is what people see.

The WhatsApp Document trick is genuinely one of the most useful things in this entire article. Most people share photos through the standard photo attachment flow and accept the compression as inevitable. Switching to Document mode takes one extra tap and preserves your original file quality completely. If you're sharing product photos with a client, sending edited photos to a collaborator, or forwarding anything where quality matters — always use Document mode.

The Five Mistakes That Guarantee Blurry Results

Now that the platform-specific details are clear, here are the five most common mistakes that cause blurry uploads — and exactly what to do instead.

Mistake 1: Uploading screenshots

Screenshots are typically saved at screen resolution — on a phone, that's usually the device's native resolution, which might be 1080×2400 or similar. But screenshots are saved as PNG or JPEG at whatever quality your device chooses, and they're often already compressed. When you screenshot an image and re-upload it, you've added one compression step before the platform adds its own. The result is a doubly-compressed image that starts from a worse baseline than if you'd uploaded the original file.

The fix is simple: always upload from the original source file, not a screenshot of it. If you're working with an image someone sent you, ask for the original rather than screenshotting their message.

Mistake 2: Uploading the wrong dimensions and relying on the platform to resize

Uploading a 4000×3000px photo to Instagram and letting Instagram resize it to 1080×810px sounds fine in principle — the platform is built to handle this. In practice, the automatic resize happens server-side using a fast algorithm that prioritises processing speed over quality. The resize you do yourself, in a proper image editor or a quality image resizer, can use slower, higher-quality algorithms (like Lanczos resampling) that preserve more detail during the downscale.

Pre-resizing your images to the exact dimensions the platform expects means the platform's processing touches your image less — it may still recompress for file size, but it doesn't need to resize. That one fewer operation can make a visible difference.

Mistake 3: Uploading JPEG images that have been saved multiple times

Every time you save a JPEG, it gets recompressed. Edit an image, save it as JPEG, open it again, make a small change, save again — that's two compressions before the platform adds a third. Generation loss is real and cumulative. For images that need maximum quality, work in a lossless format (PNG or TIFF) throughout your editing process and only export to JPEG as the very last step, at the highest quality your workflow allows.

Mistake 4: Using the wrong colour profile

Export your images as sRGB. If your camera or editing app uses Display P3 or AdobeRGB and you upload without converting, the platform converts for you — and that automatic conversion doesn't always map colours accurately. Vivid reds can become orange-red. Rich blues can flatten. Greens can shift. Convert deliberately to sRGB before uploading and you control the colour mapping rather than leaving it to the platform's automated pipeline.

Mistake 5: Using Standard quality mode in Instagram's app settings

As mentioned earlier: go to Instagram Settings → Account → Data Usage → Upload Quality and switch to High. This is a one-time change. Many people have been uploading on Standard mode for years without knowing the High option exists.

The Pre-Upload Workflow That Actually Fixes This

Rather than going through this checklist every time you post, here's a simple repeatable workflow that handles all of it. Follow it once until it's a habit, and blurry uploads stop being something you have to think about.

• Start from the original file. Not a screenshot, not a download from a previous upload, not a compressed copy. The original photo or the original export from your editing app.
• Edit in the highest quality your app supports. In Lightroom: export at quality 85+. In Snapseed: use the export options to save at maximum quality. In VSCO: export at full resolution. Most editing apps default to "optimised for sharing" which means compressed — override this.
• Convert to sRGB before exporting. In most editing apps this is in the export settings under "Colour Space." Choose sRGB. If you don't see the option, you're probably already exporting sRGB, which is fine.
• Resize to the platform's exact dimensions using a proper image resizer — not the platform itself. For Instagram feed posts: 1080×1350px for portrait, 1080×1080px for square. For LinkedIn: 1200×627px. Use 21k.tools' image resizer to get exact dimensions in seconds with no quality loss from unnecessary extra steps.
• Save the final resized version as JPEG at quality 90. This gives Instagram and other platforms the best starting point for their recompression — close enough to their target quality that their algorithm has less work to do, which means less degradation.
• Enable High Quality upload in Instagram settings if you haven't already. One-time change, permanent benefit.