RF is HandBrake's Rate Factor — the constant-quality control that decides how much detail the encoder keeps, and it is the same setting that gets abbreviated CRF. HandBrake's documentation puts the two names side by side: most official presets use "a Constant Rate Factor, sometimes abbreviated CRF or simply RF."
What most tutorials get wrong is the next part: RF is not a percentage, not a bitrate, and not a universal scale. The right number depends on which encoder you picked, and one encoder reads it in the opposite direction from all the others.
What RF actually controls
HandBrake can encode video in two modes. With average bitrate you control the size of the output file and give up control over quality. Constant quality does the opposite — you name a quality level and HandBrake varies the bitrate to reach it. HandBrake's advice on which to use is unambiguous: "Always use constant quality unless you have a specific reason not to."
The trade is that output size becomes unpredictable. Because it takes a different bitrate to hit a given quality level in different parts of a video, the size varies from source to source. Grainy footage comes out especially large, since all that grain is detail the encoder is being asked to preserve.
So RF answers "how good should this look," never "how big should this be." If you came here looking for a number that produces a 25 MB file, no RF value will do that reliably.
On the Video tab, "increase video quality by adjusting the control toward the right, and reduce video quality by adjusting the control toward the left." HandBrake adds that you should "begin by making small adjustments; plus or minus a few steps is usually noticeable."
The recommended RF ranges — read the table for your encoder
Here is the part that gets scrambled everywhere else. HandBrake publishes two separate sets of recommendations, and they are not interchangeable.
Recommended settings for the x264 and x265 encoders:
| Source resolution | Recommended RF |
|---|---|
| 480p/576p Standard Definition | RF 18–22 |
| 720p High Definition | RF 19–23 |
| 1080p Full High Definition | RF 20–24 |
| 2160p 4K Ultra High Definition | RF 22–28 |
Recommended settings for the SVT-AV1 encoder:
| Source resolution | Recommended RF |
|---|---|
| 480p/576p Standard Definition | RF 22–32 |
| 720p High Definition | RF 25–35 |
| 1080p Full High Definition | RF 25–35 |
| 2160p 4K Ultra High Definition | RF 25–40 |
For anything else — the hardware encoders, VP9, MPEG-4 — HandBrake gives no numbers at all. Its only guidance is: "For other encoders, start with a value in the middle of the scale, and adjust in small increments as desired."
That means copying the x264 numbers onto an NVENC, QSV, VCN or VideoToolbox encode is not supported by the documentation. An article that prints one RF table and labels it "HandBrake settings" has merged two tables that HandBrake deliberately kept apart.
If you are unsure which row applies, HandBrake's footnotes define the tiers: SD is 720x480 or fewer pixels (720x576 for PAL), 720p HD is 1280x720, 1080p HD is 1920x1080, and 2160p UHD is 3840x2160.
Why 4K gets a higher RF than SD
The table looks upside down the first time you read it — worse-looking numbers for better source material. HandBrake explains the logic:
"These values may seem counterintuitive at first, but the reasoning is simple and easy to demonstrate in practice. When encoding higher resolution content, you have more information to work with. Therefore, you can get away with using lower quality values than you can with lower resolution content, as any loss is a smaller percent of the whole, and ultimately less perceptible."
Low-resolution video is the fragile case, not the forgiving one. HandBrake adds that imperfections are more noticeable at larger display sizes and closer viewing distances, and that this is especially true below 720p, where the video is typically scaled up to fill the display and even minor quality problems get magnified.
The exception: Apple VideoToolbox runs the scale backwards
Every "lower RF = better quality" guide is incomplete without this, and HandBrake states it directly:
"For these encoders, a lower RF number produces higher quality video, and a higher RF number produces lower quality video. This is reversed for other encoders, such as Apple VideoToolbox, where a higher number produces higher quality. In all cases, adjusting HandBrake's quality control to the right always increases quality, and adjusting the control to the left always decreases quality."
Two things follow. First, if you are on a Mac using an Apple VideoToolbox encoder, typing "18" because a blog told you to gets you the opposite of what you wanted. Second, the slider is the safe interface: right is always better, regardless of encoder. Trust the direction, not the digit.
And the values never translate across encoders:
"Even where the quality scale is similarly labeled, quality values are not comparable between encodes across multiple encoders. You should not expect an x264 encode at RF 25 to be similar to an x265 encode at RF 25 or an SVT-AV1 encode at RF 25, as each encoder's quality scale corresponds only to its own internal rate control method, and these are three separate encoders corresponding to three different video coding standards."
How to pick a value without guessing
- Start from the preset's default. HandBrake's default preset is Fast 1080p30, and its presets ship with a quality setting already chosen. HandBrake's fallback advice: "When in doubt, stick to the recommended range or use the default setting for the Preset you selected."
- Move in small steps. A few steps either way is usually visible; there is no need to jump five points.
- Test before committing. You can test settings "by encoding and watching short clips using HandBrake's Preview feature" — the single biggest time-saver on a long source.
- Nudge quality up for animation. Raising quality by a few steps "may produce better results when encoding animated Sources (anime, cartoons)."
- Adjust for the screen. HandBrake suggests slightly higher quality for displays 50 inches / 125 cm diagonal or greater, or closer-than-average viewing; reduced quality is acceptable on small screens or where storage is tight.
- Do not go past the recommended range. "Using higher than recommended quality settings can lead to extremely large files that may not be compatible with your devices."
If all you need is a smaller file for a chat app or an email, installing a transcoder and learning a rate-control scale is more machinery than the job needs:
Try it right here
Runs in your browser. Your file is not uploaded anywhere.
That runs in your browser tab — the video is read from disk, re-encoded locally, and handed back, with no upload step. For the reasoning behind bitrate and resolution choices in general, see how to compress a video without losing quality, or go straight to the video compressor if you already know what you want.
Two things that quietly change the right RF
Grain. Grainy sources need a higher bitrate to hold all that extra detail, so the same RF produces a much larger file. HandBrake's documented responses are turning on the Denoise filter or reducing the RF value a few points.
The encoder preset. HandBrake exposes a preset and tune system on the Video tab for some encoders, and warns that "changes to the preset may also require changes to the quality slider or bitrate for optimium results" — RF and encoder preset are not independent dials. HandBrake's own page on presets and tunes is marked as a draft and does not list the available tunes, so check the official documentation rather than a third-party list.
Also worth knowing: your output can end up larger than your input. RF 0 applies no compression at all — it is lossless — and pushing quality far up the scale can produce a file bigger than the source while looking no better. That is the same surprise that catches people with lossless OBS recordings.
FAQ
Is HandBrake's RF the same as CRF? Yes, inside HandBrake — the documentation calls the setting "a Constant Rate Factor, sometimes abbreviated CRF or simply RF." What does not carry over is the number itself between different encoders.
What RF should I use for 1080p? RF 20–24 with x264 or x265. RF 25–35 with SVT-AV1. Those are two different HandBrake recommendations for the same resolution, and picking the wrong one is the most common mistake with this setting.
Is a lower RF always better quality? For x264, x265 and SVT-AV1, yes. For Apple VideoToolbox it is reversed — a higher number means higher quality. Moving the slider right always increases quality whatever the encoder, so use the direction rather than memorising numbers.
What RF should I use for NVENC, QSV or VCN? HandBrake does not publish recommended values for hardware encoders. Its only instruction is to start in the middle of the scale and adjust in small increments. Do not reuse the x264 table here.
Can I use RF to hit a target file size? No. Constant quality fixes quality and lets the bitrate move, so size depends on the source. HandBrake does offer an average bitrate mode but recommends against aiming for a target file size. For a hard cap like a chat upload limit, a compressor with a size target beats RF hunting.
Why is my HandBrake output bigger than the original file? Because you asked for more quality than the source was encoded with. RF 0 is lossless and applies no compression, and high-quality settings on an already-compressed source spend bits re-describing detail that is already gone. Back off toward the recommended range.
Want the rest of the toolkit instead of a transcoder? Browse all Clapr tools, or read more in the guides hub. RF behaviour and recommended ranges cited from HandBrake's official documentation — verify against handbrake.fr before relying on them, since HandBrake has changed preset and encoder behaviour between versions.
