Blocky uploads are almost always a second re-encode

Whatever you send, the platform decodes it and re-encodes it into its own formats at its own bitrates. You do not choose the codec, you do not choose the bitrate ladder, and no export setting persuades it to skip the step.

So the goal is not avoiding compression. The goal is handing the encoder the easiest possible job, and refusing to add a compression of your own on top of the one you cannot escape.

That reframe sorts every piece of advice you will read on this. Anything that reduces how hard the encoder has to work is worth doing. Anything that claims to bypass the encode is simply wrong.

It also explains why the same post looks different on two phones in the same room. Playback quality depends on which rung of the ladder a viewer's connection pulls down, and that is not yours to control either.

This is the biggest cause by a distance. Every time a video is decoded and re-encoded, detail is thrown away permanently, and the losses stack rather than cancel out. Editors call it generation loss and it is the reason a file can look worse each time it moves.

The common way to collect three generations without noticing: film on the phone, export from an editing app, send the file to yourself through a messaging app, then upload. The messaging app is usually the worst offender, because it is optimising hard for a small file rather than a good one.

Fix it by counting your encodes. Original take, one export, one upload. Move files with a cable, a shared album set to original quality, or a cloud drive that stores rather than transcodes. Never move a master through a chat thread.

Stop re-saving as well. Opening a finished export, trimming a second off it in a second app and exporting again is a whole extra generation spent on a change you could have made once.

Uploading over a weak or congested connection is a common cause of a soft post and the least obvious one, because nothing on screen tells you it happened. Apps handle a struggling upload by shipping a smaller version rather than failing outright.

Upload on wifi, with the app in the foreground, and let it finish. Backgrounding the app halfway through or switching from wifi to mobile data mid-upload is enough to change what actually gets sent.

Where the app offers a high quality upload option, turn it on and check it again after updates, because settings like this quietly reset. It does not defeat the platform encode. It changes the file the platform encode is given.

If a post looks worse than the file sitting on your phone, delete it, wait for a better connection, and post the same file again rather than hunting for a new export setting. The export was usually fine.

Hand over 1080 by 1920 at thirty frames a second. That is the shape the player draws without scaling anything, and scaling on the way in is softness you paid nothing to receive. Capture resolution is a separate decision with a different answer; this is about the file you deliver.

Some footage is simply harder to encode, and the difference is visible. Heavy grain from a dim room, confetti, rain, fast pans, tight patterned shirts and constant handheld shake all force the encoder to spend bits on noise instead of on your face.

So light the room. A brighter scene means less sensor noise, less noise means fewer wasted bits, and fewer wasted bits means a cleaner picture where it matters. Lighting is a compression setting that happens to also be a lighting setting.

Lock the camera down. A tripod, a stand or a stack of books removes the global motion that eats bitrate on every single frame of a talking take, and it costs nothing.

Screen recordings are their own case. Fine text and thin lines survive compression badly, so record at the highest resolution the device offers and zoom in while recording rather than expecting the viewer to squint later.

Adding text, filters, effects or trims in the platform's own editor after import means the file gets processed again on the way through. Finish the edit first and upload something that needs nothing done to it.

There is one sensible exception, and it is captions. Many people are better off using the built-in caption tool, for reasons that have nothing to do with quality and everything to do with viewers being able to switch them off.

Watch the small things too. A cover frame chosen from inside the app is fine. Re-cropping a video that is already the right shape is not, because it is a scale and a re-encode for no gain.

And here is the limit no setting removes: every platform recompresses everything. A perfect file and a merely good file both come out the other side as the platform's own encode, and past a certain point the extra care stops being visible to anybody. Get the four causes above right, then stop optimising and go and write a better opening line.

If your workflow currently runs phone, editing app, chat thread, platform, the cheapest quality win available to you is deleting a step rather than changing a setting.

Cutroom takes the take straight off your phone and returns one finished 1080 by 1920 MP4 at thirty frames a second, with the captions already burned in and the silence already trimmed, which is a single export instead of a chain of them. The batch is 100 credits and the export is 20 credits an output minute. The first three are free in your first seven days, with 600 credits and no card; Lite at $19.99 a month for 1,250 credits cancels in one click under Settings.

The honest limits: vertical only, three minutes of source at most, and the platform still re-encodes whatever you hand it. Nothing on this page changes that last part, and anything claiming to is selling you something.