In this lesson I explain the anamorphic workflow from shooting to post.
Please read Exploring Cheap Anamorphic Adapters for the Panasonic GH5, because it has the basics on the anamorphic squeeze and desqueeze, as well as some adapters you can use with mirrorless cameras.
Watch this video first:
Step One: Knowing what the Squeeze and Desqueeze factor is
Both are the same number. Depending on the lens you are using you will have a certain squeeze number. The more expensive and modern anamorphic lenses have a correct representation of the squeeze factor. E.g., Arri Master Anamorphics are 2x, so you can exactly stretch them by two times later in post. The cheaper anamorphic lenses might not be exactly 2x or whatever.
How do you find out the correct squeeze factor?
The simplest way is to just draw a circle or square in Photoshop and then shoot that (the computer monitor). You can even download perfect circles/squares from Google images and just align your camera perfectly (use your camera level) and monitor so they are parallel.
Later in post production you can stretch the ‘shot’ image only along the x-axis (zoom feature in Resolve) and can compare it with the circle (import it as another image and place it on top of the shot image with lower opacity) so you’ll get the exact desqueeze. Once you’ve done that this factor remains constant throughout, so you don’t have to do it for the same lens-camera-adapter combination. Simple!
Step Two: What setting to shoot in?
This entirely depends on what formats your camera supports. Here are some common formats:
- 16×9 UHD or HD – the most common.
- 4:3 UHD or 5K, like the Panasonic GH5, as I’ve explained in my review of v2.1 of the firmware update.
- 4K DCI, which is 4096:2160 or 1.9:1
To get a true anamorphic look, you need a sensor that’s exactly 6:5. Only the Arri Alexa (in Open Gate), Red Monstro/Dragon, Sony Venice 6K have sensors big enough to give you a true 6:5 frame.
The best you can do with mirrorless cameras is 4:3, so in the Panasonic GH5 you can shoot 5K which gives you the full size of the sensor and is the widest field of view.
I’ll use this as an example. This is how it works:
- Resolution: 4992 x 3744
- Aspect ratio of the sensor: 1.33
- What happens when you add a 2x anamorphic lens: Multiply by 2. Whatever the anamorphic factor is, multiply 1.33 by that. If it’s 2, then the answer is 2.66:1.
- What is the traditional anamorphic widescreen aspect ratio? It’s 2.39:1. To get an exact 2.39:1, you need an anamorphic lens with a squeeze factor of exactly 1.8x. Like I mentioned in the earlier article, only one lens+adapter combination meets that, so you’re either going to display the image as-is, or crop the sides.
- Your choices: If you crop the sides, you lose resolution.
- So, if you’re planning to display in 2.39:1, then it’s better to shoot in 5K so when you crop, you’re still left with 4K. Makes sense, right?
- Or, you could just show the image as-is. With the Internet, that’s not hard to do. But it’s not to everyone’s taste.
- Let’s assume for the sake of simplicity you pick 2.39 to stay true to the original anamorphic aspect ratio. This means you’ll need to crop your image. How much should you crop? Two ways:
- 2.66/2.39 is 1.116 or 11.6%. This gives you a final resolution of 4473 x 3744.
- You could just start with 6:5 (1.2:1), so the resolution is 4493 x 3744.
- These numbers shouldn’t be different, but the reason they are different is because we are rounding off the numbers. If you want to be accurate, stick to 4493 x 3744.
In simple words, you’ll be shooting 4992 x 3744, but cropping off the sides a bit to get 4493 x 3744 so you get a final aspect ratio of 2.39:1 once you apply a 2x desqueeze.
Side note: What does this do to the crop factor?
MFT has a crop factor of 2x. However, if you crop the sensor to 4493 the crop factor becomes 2.22x. E.g., a 25mm lens is 50mm (35mm equivalent) on MFT, but 55mm on the same camera and lens with a 2x anamorphic adapter designed for a final aspect ratio of 2.39:1.
If you decide to stick to 2.66:1, then there’s no additional crop factor.
Step Three: Should I turn on Anamorphic Desqueeze in-camera?
Unless you’re shooting with a true 2x lens and want to display 2.66:1, then it’s pointless. However, you could tape the sides to get a rough idea of 2.39:1 if you like.
Or, you could just learn to compose with the squeezed frame, it’s not that hard with some practice.
Sticking to the Panasonic GH5, there’s no 6:5 frame guide, so you’re stuck to looking at the image as 2.66:1. Then you could apply a bit of black tape to both sides so you see 2.39:1.
Expensive cameras like the Arri Alexa and so on allow you to create custom frame lines, but unfortunately low end mirrorless cameras don’t offer this level of customization.
Step Four: Post Production
There are two strategies you could adopt:
- Keep the horizontal length and squeeze the vertical. E.g., The horizontal resolution is 4992 and the video will continue to remain 4992 horizontally. The vertical height will change depending on your aspect ratio.
- Keep the vertical height and desqueeze the horizontal. E.g., the horizontal resolution is 4992 and the video is now stretched to become 9984px horizontally, or 9K! However, you don’t get the resolution of 9K, only 5K. I don’t prefer this solution because if someone in the future were really watching in 8K or higher, it would look poor compared to true 8K content.
- So let’s stick to varying the vertical height and continue with the Panasonic GH5 story: We decided to shoot 4493 x 3744 so we end up with 2.39:1. There are three ways we can go for final publication:
- One: We decide to output to YouTube UHD, which would be 3840 x 1600 (In this case I suggest shooting for 2.4:1 so the numbers round up nicely). So you would create a timeline of 3840 x 1600 and then drop your footage in. You would apply the desqueeze via vertical zoom and then scale the footage so it fits to 3840 horizontally. That’s it.
- Two: We decide to output to 4K DCI, which is 4096 × 1716. So you would create a timeline of 4096 × 1716 and then drop your footage in. You would apply the desqueeze via vertical zoom and then scale the footage so it fits to 4096 horizontally. That’s it.
- Three: You crop to either of the above. Since you have extra resolution you can always use that extra room to crop or pan/scan.
Most people shooting mirrorless will have their videos shown on the Internet. Even though YouTube accepts varying aspect ratios, for 4K, it recommends UHD or 3840 instead of 4096.
That’s it! There’s nothing more to this, really. If you have any questions please feel free to ask.