Share video with a client without YouTube compression.
Four steps, and one setting YouTube does not have.

Read those four together and the problem with an unlisted screener stops being about softness. YouTube states that it re-encodes every upload. It states that your video will be converted to the highest resolution available to ensure successful playback on different devices and networks. And it publishes no figure at all for what that conversion produces.

So when a client writes back that the sky looks blocky, you have nothing to say. You cannot tell them the bitrate they received. You cannot tell them whether they watched the finished processing pass or the first one. You cannot tell them whether their connection dropped them to a lower rendition halfway through. Every one of those is a plausible explanation and none of them is checkable, which means the conversation becomes about your grade instead of about the pipe.

The timing failure is the one that actually costs money. YouTube says that when you upload a video it will initially be processed in low quality, and that higher qualities may seem to be missing for several hours. Its own example is a 60-minute 4K 30 fps video taking up to 4 hours. The natural thing to do, upload and send the link the moment the progress bar finishes, is exactly the thing that hands your client the low-quality pass.

One more reason not to lean on a visibility setting you do not control. On 23 July 2021 YouTube retroactively made older unlisted videos private unless the owner opted out, covering everything unlisted that had been uploaded before 1 January 2017. That was a defensible security decision and it is also a demonstration: the meaning of the setting your client’s link depends on is Google’s to change. A password you set, that you can change or remove without touching the URL, is a smaller and more predictable promise.

This is the part filmmakers try hardest to engineer away, and it is closed. The modestbranding parameter is, in Google’s own documentation, deprecated and has no effect. The player now determines the appropriate branding treatment based on a combination of factors, including player size, other API parameters and additional signals. You must not display overlays, frames, or other visual elements in front of any part of a YouTube embedded player, including player controls. Related videos cannot be switched off in an embed at all; setting rel=0 only restricts them to the same channel. Hiding the controls with controls=0 is the one piece of chrome the API still lets you remove.

None of that matters for a music video with two million views. It matters a great deal when a brand’s marketing director opens your cut and the platform decides what sits next to it.

The version of this that works in practice is not a migration. It is a split. The approval copy goes out from a page you control, streaming the exact export, with a password on it and the master downloadable beside it. Once the client signs off on what they genuinely saw, the finished film goes to YouTube for everyone else. The two do different jobs and there is no reason to make them fight.

The practical cost of the split is small. Passwords, the custom link, the unbranded player and the original-file download toggle are all on the free Starter plan, which carries 5 GB and one film at a time. That is enough to run a single screener at a time and decide whether the workflow is worth keeping before you pay anything.