Video encoder tiers are preset quality levels that determine how hard CoreReflex works to compress your finished cut, trading render time and file size against visual fidelity. A draft tier gives you a fast, lightweight preview; a master tier produces a high-bitrate file built to survive the re-compression every social platform puts it through. Choosing the right tier is how you make each export the correct size and the correct look for wherever it's headed, without paying for quality the destination will throw away.
What video encoder tiers are
When a render finishes, the raw frames have to be encoded into a deliverable video file. Encoding is lossy compression: the encoder decides how much data to spend representing each frame. Spend more, and the file looks closer to the source but grows larger and takes longer to produce. Spend less, and it renders fast and downloads quickly, but soft gradients band and fast motion smears.
An encoder tier bundles those decisions into a named choice so you don't have to hand-tune bitrate, compression effort, and quality targets on every export. Instead of fiddling with codec parameters, you pick the tier that matches the job, draft, standard, high, and CoreReflex applies a coherent set of settings underneath, it is the same idea as "low / medium / high" in a game's graphics menu: one dial that moves several knobs together in a sensible way.
The trade-off every tier balances
Four things move when you change tiers, and they pull against each other:
- Visual quality, how faithfully the file represents the rendered frames.
- Bitrate, how much data per second the encoder is allowed to use; the main lever on quality and size.
- File size, bitrate times duration, roughly; what you actually upload and store.
- Render time and cost, higher tiers ask the encoder to work harder, which takes longer.
No tier wins on all four. A higher tier buys quality at the cost of size and time; a lower tier buys speed and small files at the cost of fidelity. The skill is matching the tier to the destination so you aren't rendering a 4K master to share an internal rough cut, or shipping a heavily compressed draft as a client's hero ad.
How tiers fit the CoreReflex render path
Tiers are not a cosmetic setting bolted on at the end, they are part of a deterministic render path. CoreReflex runs a real Remotion render-worker that operates as a queue: it claims a render job, renders the frames, encodes at the chosen tier, uploads the result to your storage, and writes an asset row pointing at it. That pipeline is hardened with automatic retries and a dead-letter queue, so a transient failure doesn't lose your render, and it can run on GPU for speed, the why behind that is covered in GPU video rendering.
Two details matter for tiers specifically:
- Faststart. CoreReflex writes files with the metadata moved to the front, so a video can begin playing before it has fully downloaded. That's what makes a delivered file feel instant in a browser instead of staring at a spinner, and it applies regardless of tier.
- Determinism. The same manifest rendered again produces the same cut. Tier is part of that manifest, so "render the master" tomorrow yields exactly what it did today.
Because all of this runs on an owned, end-to-end stack rather than a chain of third-party export services, encoder behavior stays consistent and predictable from draft to master.
Choosing a tier for the job
Match the tier to where the video is going to live:
| Tier | Best for | Optimizes for |
|---|---|---|
| Draft | Internal review, quick iterations, sharing a rough cut | Speed and small files |
| Standard | Most social and web delivery. Reels, Shorts, embeds | Balanced quality and size |
| High / master | Client deliverables, paid ads, archival, footage you'll re-edit | Maximum fidelity and headroom |
Drafts and internal review
When you're iterating on the edit, checking pacing, swapping a shot, getting a colleague's sign-off, quality is not the point; turnaround is. A draft tier renders fast and produces a small file that uploads and streams instantly. Use it for everything up until the cut is locked.
Social and web delivery
For most published social video, a standard tier is the sweet spot. Platforms re-encode whatever you upload anyway, so an enormous master is wasted effort, but a too-aggressive draft hands the platform's encoder bad source material and the result looks rough. A balanced tier gives the platform clean frames at a reasonable size. Pair the right tier with the right container and dimensions from our export formats guide and your uploads survive platform compression looking sharp.
High-bitrate masters and archival
When the file is the final product, a paid ad, a client handoff, or footage you intend to re-edit later, render the high tier. The extra bitrate gives downstream encoders headroom, preserves detail in gradients and motion, and means a re-compression doesn't visibly degrade the work. It's the tier you archive, because you can always make a smaller file from a master but never recover detail a draft already discarded.
Tiers and 4K/8K upscaling
Tier sets the encoding quality; resolution is a separate decision. CoreReflex generates natively and offers an own-tech super-resolution seam that can take a cut up to 4K and even 8K. The two compound: a high encoder tier on an upscaled master is what you want for a big-screen or premium placement, while a 1080p standard tier is plenty for a phone feed. Don't pay for 8K-at-master on a clip that will only ever play in a vertical mobile slot.
Why deterministic encoding matters
The reason tiers are trustworthy is that the render path is reproducible. Because the same manifest yields the same cut, you can render a quick draft to approve an edit, then render the exact same edit at the master tier for delivery and know nothing else changed, only the encode. There's no "final final v3" drift between what got approved and what shipped. That determinism, plus the retry and dead-letter safety net, is what makes tier selection a simple, low-stakes choice rather than a gamble you re-run all afternoon. You can browse the rest of our how CoreReflex works explainers for the same treatment of the render queue and the quality gate.
Frequently asked questions
What are video encoder tiers?
They're preset quality levels that bundle bitrate, compression effort, and quality targets into a single named choice, typically draft, standard, and high. Instead of hand-tuning codec parameters, you pick the tier that matches your destination and CoreReflex applies a coherent set of encode settings underneath.
How do I choose a video quality tier?
Match it to where the video will live. Use a draft tier for internal review and iteration, a standard tier for most social and web delivery, and a high or master tier for client deliverables, paid ads, and anything you'll archive or re-edit. The rule of thumb: don't render more quality than the destination can show, and don't ship a draft as a master.
Does a higher tier mean a bigger file?
Generally, yes. Higher tiers allow a higher bitrate, which spends more data per second to preserve detail, so the file is larger and takes longer to render. That's the trade you're making: fidelity and headroom in exchange for size and time, which is exactly why a master is overkill for a quick internal preview.
Does the tier change my edit?
No. The tier only affects how the finished frames are encoded into a file. Because CoreReflex renders deterministically, the same manifest produces the same cut, you can re-render an approved edit at a higher tier for delivery and be certain only the encode changed, not the content.
Can I render a draft and then a master of the same cut?
Yes, and that's the recommended workflow. Approve the edit on a fast draft tier, then re-render the identical manifest at the master tier for the final deliverable. Nothing but the encoder settings differs between the two.
Render at the right quality, every time
Encoder tiers are a small choice with an outsized payoff: fast iteration while you're still editing, clean delivery when you publish, and full fidelity when the file is the product, all on a deterministic render path that gives you the same cut every time. Start free, no credit card, and export your first cut at the tier the job actually needs. For the full encoding and render reference, see the developer docs.