A lot of people blame YouTube the second their footage looks different after upload. Sometimes that is part of it, but a lot of the time the issue starts earlier. The viewer is not set up right, the color pipeline is doing something unexpected, or the export tags are off. By the time the file hits YouTube, the image is already on shaky ground.
If you are on a Mac, this gets more noticeable. Resolve, QuickTime, Safari, and YouTube do not always show the image the same way unless the workflow is set up properly. What usually throws people off is the midtones. The shot feels brighter, thinner, or just a little off compared to what they saw while grading.
This is the setup I would use if the goal is a cleaner, more predictable YouTube upload. It fixes the viewer side first, then the CST workflow, then the export settings so the file holds together better once it leaves Resolve.
Why Resolve exports look different on YouTube
What is really happening here is that macOS uses ColorSync to interpret video through the display profile. On a wide gamut display like an iMac, MacBook Pro, or many newer monitors, Rec.709 can shift visually if Resolve is not accounting for that behavior.
That is why people leaned on Rec.709-A for so long. It was a workaround for older Apple display behavior. In Resolve 20.2.2, Blackmagic added viewer controls that do a better job of lining Resolve up with how macOS apps display video. So now the cleaner path is to fix the viewer properly instead of forcing Rec.709-A into the workflow.
Fix the Mac viewer before you export anything
The first setting I would check is Use Mac display color profile for viewers. For most Mac users, this should stay on. It lets Resolve hand the viewer off to ColorSync so what you see in Resolve is closer to what you will see in QuickTime and Safari.
The second one is Viewers match QuickTime Player when using Rec.709 Scene. Even in a node-based workflow, this helps clear up the old mismatch people kept running into. Once those two settings are in place, the image usually starts to behave more predictably across apps.
The only time I would turn the Mac display profile setting off is if I were grading through a true reference monitor with a DeckLink or UltraStudio setup and wanted ColorSync completely out of the chain.
Project setup for a cleaner CST workflow
I prefer a non-color-managed DaVinci YRGB workflow here with CST nodes doing the heavy lifting. The reason this works well is that it keeps the transforms visible. You know where the input conversion happens, where the output conversion happens, and that makes it a lot easier to troubleshoot when something feels off.
In Project Settings under Color Management, set:
- Color Science: DaVinci YRGB
- Timeline Color Space: DaVinci Wide Gamut Intermediate
- Output Color Space: Rec.709 Scene
That project output setting is not your final delivery transform. It is there so Resolve’s tools and scopes behave more predictably inside the project. The actual display-referred transform still happens in the final CST node.
Output CST settings that hold together better
Do your grading in the working space, then finish with an output CST on the last node. This is usually where the workflow starts to feel dialed in because you are controlling the display transform directly instead of letting it stay vague.
Set the final CST like this:
- Output Color Space: Rec.709
- Output Gamma: Gamma 2.4
- Use Forward OOTF: On
The reason this works is that it gives you a proper display-referred SDR output. That tends to hold together better for YouTube, QuickTime, and Safari once the viewer settings are also correct.
Rec.709 tags and metadata to leave alone
A lot of people miss this part because the file still exports either way. The problem is that the wrong tags can cause another app to interpret the file differently, and that is where the image starts to drift again.
For standard Rec.709 SDR delivery, the correct NCLC tag combination is 1-1-1. In Resolve, the safest move is to leave Color Space Tag and Gamma Tag set to Same as Project on the Deliver page. That gives Resolve the best chance of embedding the right Rec.709 metadata without adding another variable.
Why exporting in 4K still helps HD footage
If your source footage is 1080p, exporting in 4K can still give you a cleaner result on YouTube. The reason for that is YouTube tends to assign better transcodes and more bitrate to 4K uploads. So even when the source is HD, the final 1080p playback often looks cleaner than a straight 1080p upload.
If you are in Resolve Studio, you can take this a step further with AI Super Scale. For 1080p footage going to 4K, I would set it to 2x Enhanced and leave sharpness and noise reduction at default unless the shot clearly needs more work. It is GPU heavy, so I would save that for the end of the edit.
Best DaVinci Resolve export settings for YouTube 4K
For a direct export out of Resolve, this is the setup I would use:
- Format: QuickTime
- Codec: H.264
- Resolution: 3840 x 2160
- Quality: Restrict to 80,000 Kb/s
- Encoding Profile: High
- Entropy Mode: CABAC
- Key Frames: Auto
- Frame Reordering: Off
- Data Levels: Auto
- Color Space Tag: Same as Project
- Gamma Tag: Same as Project
- Force sizing to highest quality: On
- Force debayer to highest quality: On
This gives you a solid YouTube file without pushing the bitrate into bloated file sizes for no real gain.
Clean the footage before export
If the footage is noisy, compression gets ugly fast. Random noise forces the encoder to spend more data describing movement that is not helping the shot. Once that gets uploaded and transcoded again by YouTube, the image can fall apart even faster.
So if a shot needs noise reduction, do that before export. Cleaner footage usually compresses better, and that shows up in smoother gradients, cleaner shadows, and fewer ugly artifacts around moving details.
The cleaner pro workflow: ProRes first, then HandBrake
The direct H.264 export is fine for most uploads. If I want a cleaner final result, especially for footage with motion, texture, or a lot of fine detail, I would export a ProRes 422 HQ master first and then make the H.264 delivery file in HandBrake.
The reason for that is HandBrake’s x264 encoder tends to do a better job preserving detail and avoiding blocky compression than a straight H.264 export out of Resolve. It is an extra step, but it usually holds together better.
For the HandBrake encode, use:
- Codec: H.264
- Constant Quality: RF 18
- Preset: Slow or Slower
- Resolution: 4K
What I’d actually use day to day
If I need speed and consistency, I would export straight out of Resolve using the 4K H.264 settings above. If the project is more important and I want the cleanest file I can get for YouTube, I would go ProRes 422 HQ first, then HandBrake.
Either way, the bigger issue is not just the bitrate. It is whether the whole chain is set up properly. Once the viewer, CST workflow, tags, and export settings all agree with each other, the image usually stops shifting around so much between Resolve and the final upload.
That is where this starts to feel better, not just different. You can trust what you are seeing while grading, and that makes every export after that a lot less frustrating.