[Tutorial] How to get a VHS-like look in DaVinci Resolve

This is the source clip I'm going to be working with of my cat Deanna. I'm going to demo this in DaVinci Resolve 17, but this should work in almost any version.

Ideally you would start with a clip that is 59.94fps or 50.00fps, depending on what part of the world you live in. The good news is that you don't need a 1080p recording, 720p will work just fine. SD video was recorded interlaced, which breaks up a frame into two fields, one of all the even numbered lines, one of all the odd numbered lines, which would be broadcast one after the other. Instead of taking one frame and breaking it up into two fields camcorders would often capture each field directly, which would make a 29.97 FPS recording look like 59.94 FPS, but at the expense of resolution. This is known as the Soap Opera Effect because soap operas used video cameras instead of film to reduce costs and speed up production.

If you don't have a 59.94/50p source clip that's fine, you just won't get the Soap Opera Effect look, or any interlacing artifacts.

More information on interlacing: https://en.wikipedia.org/wiki/Interlaced_video

After that we're going to add a new Parallel Mixer the same way we created the Corrector nodes.

First create a new project in Resolve, and then open the Project Settings and choose either 720×480 NTSC or 720×576 PAL, depending on your region. If you have 59.94/50.00p source material make sure you select Enable Interlace Processing.

When you bring your source clip into the Media Pool you'll likely be prompted to change your project frame rate. DO NOT CHANGE IT.

Take the part of your clip that you want to crapify and add it to a timeline.

Select your clip in the timeline, and in the upper right hand corner open up the Inspector.

Scroll down to the Retime and Scaling options and change Scaling to Fill.

You may also wish to add a video noise filter to your video at this stage, to emulate VHS noise. Which filter and how aggressive to make it are a matter of personal choice and preference. It's important to add it at this stage, because you want the grains of noise to be relative to the 480p image, and not HD noise over an SD image. The process of crapification we're doing will also knock some of the "digital" edge off the noise grain.

Switch to the Delivery Pane and choose your output format. For Windows Users I recommend using the AVI Format with the Grass Valley HQ codec. For macOS I would select Quicktime and use DV/DVCPro. If you're working with 59.94/50.00p source material make sure you have Interlaced Rendering selected.

Add this to your render queue and export.

You may wish to check your output at this point. If you see these kinds of striations or horizontal lines in areas of fast movement, you have successfully interlaced your material. What you are seeing here are the two independently recorded fields being shown at the same time.

So now this is what our video looks like. Softer already, but we're not done yet.

Now we're going to crapify it further and bake in some interlacing effects. To do this I'm using Shutter Encoder, an open source tool that provides a user-friendly interface for the powerful video manipulation tool ffmpeg.

For maximum flexibility I'm selecting the Apple ProRes output function. 422 is plenty good for what we're doing. I'm also going to scale down the video to 480×360, though you may want to go down to 320×240 if you want to go even fuzzier.

I'm hoping to bake in some deinterlacing artifacts, so I'm forcing deinteracing. Experiment with different algorithms until you find one that produces the correct amount of artifacts for your preference. If you wish to retain the interlacing artifacts then do not deinterlace, and Force to Progressive. Note that keeping these artifacts can dramatically reduce the final quality of your video once it's compressed for streaming, for reasons explained in this video: https://www.youtube.com/watch?v=r6Rp-uo6HmI

You can get Shutter Encoder here: https://www.shutterencoder.com/

So now this is what we've got. Pretty crappy, but we're still not done.

We're now going back to Resolve and we're going to make a 1080p29.97 or 1080p25.00 project.

If your entire video is going to be 4:3 and not incorporate any widescreen elements, then you may wish to set a custom resolution at 1440×1080.

Now, again, we're going to bring the clip into a timeline, open the Inspector, and here we want to change the Resize Filter to Smoother, though you may wish to experiment with this for your own preferences.

Now we're going to separate the Chroma and Luma channels so we can independently manipulate them. Video, and video tape is no exception, separates brightness information (Luma) from color information (Chroma) so they can be compressed separately. Through a process known as Chroma Subsampling color resolution is reduced to make the whole video stream easier to transmit or store. Formats like VHS and 8mm had exceptionally poor chroma resolution, and by separating the chroma from the luma we can emulate that more accurately.

More info:
Luma: https://en.wikipedia.org/wiki/Luma_(video)
Chroma: https://en.wikipedia.org/wiki/Chrominance
Chroma Subsampling: https://en.wikipedia.org/wiki/Chroma_subsampling

To do this we're going over into the Color pane, and right click on the clip in the node in the viewer, and select to Add Parallel node.

If you wish you can add Node Labels to keep things straight.

Select your Luma node and in the Wheels set the Saturation control to 0.

Full disclosure: after stumbling on how to make things work reliably, I wound up running in to this tutorial (https://youtu.be/GasdNpuI470?t=251) that said the secret for moving forward is to add two Corrector Nodes instead of doing everything in our Chroma node. You do that by right clicking into the empty area of the node map and choosing Add Node, then Corrector.

Connect the output of the Chroma Node (the yellow box on the right side of the node) to the input (the yellow arrow on the left side of the node) of each of the two Corrector Nodes you have just created. It's as simple as clicking and dragging from the yellow box to the yellow triangle.

Now, to cover our asses and make sure we're doing everything right we're going to right click on each of the new Corrector nodes and we're going to set the Color Space to Y'UV. This is commonly called Component Video, where Y is the Luma channel and U and V are the Chroma channels.

Now we need to isolate the Chroma channels, and we do this by right clicking on the nodes and disabling Channel 1 (Y) and Channel 2 (U) on one node, and Channel 1 (Y) and Channel 3 (V) on the other.

Now right click out in the open and add a Parallel Mixer node.

Now connect the outputs from the two Correction nodes to the Parallel Mixer

Right click on the original Parallel node we created way back at the start, and Morph into Layer Mixer Node.

Change the Composite Mode to Add.

Create a new Corrector node, and connect the output of the Chroma Parallel Mixer node to it.

Now connect the final Chroma corrector to the Layer Mixer.

Change the Corrector's Color Space to HSL

Select the node, and down at the bottom in the Curves window select the B channel, which in HSL mode actually represents L, or "Lightness" in the HSL color space. Grab the point at the upper right hand corner and drag it all the way down to the bottom. This isolates *just* the color.

In the upper right hand corner open the Open FX tab, and drag in a blur of your choice. I generally go with the Gaussian blur, but you may like the lens blur more.

Take the output of the Chroma Corrector, and connect it to the Blur node, and then the Blur node to the Layer Mixer.

If you select the blur and open the Open FX controls and manipulate the blur until it looks right for you. Historically analog tape formats had worse horizontal Chroma resolution than, so it may be worth it to ungang the two settings and adjust them independently.

And this is the end result. It could be better if I had gone to 320×240, but for this tutorial I didn't feel like going back all the way and starting over again. Noise and over-exposing the shot in the color grading tool would help complete the look, but my primary goal was to show how to effect the sharpness of the chroma independently of the luma for that color bleed that is quintessentially tape.

Now select the Luuma corrector, and down in the Curves selection choose the Y adjustment and bring the top right point in to the left to push the shot to *nearly* over-exposed. Analog cameras of the day had really crappy auto-exposure controls, so blown out shots were common.

To further this, you may want to create a point near the bottom left and lift the black level a smidge, so there's less contrast between light and dark. The amount to adjust is subjective. Add as many points as you feel are necessary to get the correct over/under-exposure look you desire.

If you wish to do more direct color grading on the footage, to give it a cast, change the white point, etc, right click on the Layer Mixer node and Add Serial node to add a corrector downstream from it. From there it's like grading footage in Resolve normally.

And now what we have looks more like this. It would seem a bit more VHS-like if I had added some noise, but I wanted to focus on the color aspect of this guide, and adding any additional noise would have made the GIFVs look blocky and distract from what I was trying to demonstrate.

Alternatively, just put the stuff on tape and capture it back, or record on an old camcorder. There's a ton out there.