Blender 3.0 for Production - 2D & 3D Tracking Workflow - 01 Preparing Footage in the VSE

In this tutorial series, we will talk about Blender as a solution for 3D Camera and Object Tracking. Step by Step you will learn how to use Blender for preparing your footage in the Video Sequence Editor (VSE) and getting used to the Movie Clip Editor. Then we will cover the important fundamentals of 2D-Tracking in Blender and also some use-cases for directly using these data. Then we build upon these 2D-Tracks and start solving a 3D-Camera and talk about good and bad tracks, nodal shots, tricks like offset-tracking, joining of tracks, and analyzing solve problems. In the final tutorials, we will finish with Object-tracking for props replacement. After this tutorial series, you have a solid understanding of Blender's Camera and Object Tracking and how to use the solved data inside of Blender or in other applications like Maya, Cinema, or Houdini.
If you have questions or topics, which should be added, please let me know.
If you want to support me for publishing more free tutorials you could join my Patreons:
👉 / pixeltrain3d
or give me a fast Coffee time Tip:
👉 PayPal: paypal.me/pixeltrain
Want to start learning Blender Fundamentals with Helge?
👉 • Publication Trailer - ...
These topics are covered in this tutorial:
Why Blender as a Tracking Solution?
Understanding the tutorial series concept
Reasons why you use image sequences
Blender’s Video Sequence Editor (VSE)
Working with the File Browser
Generating Bookmarks
Understanding the different VSE Editor Types
Adding strips with the Add Menu
Stacking strips in the timeline
Arranging your edit
Time vs. Frames
Efficiently navigating in the timeline
Project Settings, Framerate
Strip properties
Color Management & Color spaces
Splitting strips
Working with strip handles
Preview during transform setting
Working with Freeze Frames
Writing an image sequence
Load image sequences
Snapping in the timeline
Configuring the Blender VSE memory
Color Adjustments for later tracking
Adjustment Layer strips
Adding strip modifiers
Increasing contrast with gamma and gain
If you want to use the same footage, here is my Gumroad link:
pixeltrain3d.gumroad.com/l/GEdjT
Be aware, that these files are included in my Patreon ;-)
/ pixeltrain3d
This tutorial series „Blender for Production“ focuses on Blender as a full Production Suite for (Indie) 3D- & VFX Productions. In these tutorials, I will explain how to work with the different Modules of Blender and how they are connected. I will give you an introduction to the specific functionalities, but also the production context and workflows, which can help you to use Blender in your productions.
As a Senior 3D & VFX Trainer, Helge Maus teaches 3D- & VFX Applications for 20 years. His focus lies on VFX, but he also works with many studios and agencies from different visualization areas. He focuses on Houdini FX, Blender, and NUKE for his daily work and training.
If you have questions or topics, which should be added, please let me know.
If you want to support Helge Maus for publishing more free tutorials you could join his Patreons:
/ pixeltrain3d
Find more information on www.pixeltrain.net
KZbin: / pixeltrain3dvfxtrainings
Twitter: / pixeltrain
Mastodon: mastodon.art/web/@pixeltrain3d
Instagram: / pixeltrain3d
LinkedIn: / pixeltrain
Artstation: www.artstation.com/pixeltrain

[@johntnguyen1976]

A little over a year later...and I wanted to come back and say that this particular series is my go-to reference for Blender camera tracking...and I've come back to it several times now for a revisit. It's hands-down the best Blender tracking series anywhere (paid or free). Thank you, sir! 🙏

[@ganapathym3664]

Clear teaching

[@ppsystems4317]

Again a very well explained tutorial. I also found your channel via Andrew Price... Congratulations, keep the good work !

[@FarukAhmet]

You have a wonderfully clear and unrushed style of presenting (kinda reminded me of Shankar Vedantam from NPR's Hidden Brain haha), great video! Found your channel via Andrew Price's newsletter email, and I'm very glad I did.

[@pixeltrain3d]

Thanks a lot, happy that you like it :-)

[@antonios5572]

just discovered your channel! great work!

[@pixeltrain3d]

Thanks, Antonio :-)

[@fakecubed]

PNG compression is lossless. The compression slider does not cause artifacts or lose detail in your image. It only affects how long it takes to encode/decode the PNG, and how much space it takes up on your drive. If you have a decently fast computer there's no reason not to use compression. The difference in file sizes are quite significant. For big projects with a lot of frames this is important. Somebody at Blender probably figured out that 15% is a sweet spot for most cases, at least whenever they made that the default; it might be higher now with modern CPUs. There are diminishing returns the higher you go, but you can get very small file sizes even for very high resolution images. On a recent 4K project I was saving tens of gigabytes on my hard drive by maxing out the compression, for the exact same image quality, identical pixel data. Blender just has to work a bit harder to save and display them. You don't have to take my word for it either. You can confirm that it's lossless yourself by outputting the same image twice at 0 compression and then at 100 compression. Then in the compositor bring both images in, and do a mix node with a subtraction. You'll get an absolutely black image, meaning no difference between the pixels in the two files. You are right about JPEG. JPEG is lossy, and should never be used in any intermediate step.

[@pixeltrain3d]

Yes, makes absolute sense. So, for fast playback it should be 0% so that it hasn't to be decoded. Thanks for the clarification.

[@fakecubed]

@@pixeltrain3d There is still encoding/decoding, as the base PNG format still stores palette and other data which needs to be converted to display colors from encoded values in the bitstream, but the algorithmic complexity is much more basic without the added compression layer. The compression basically just makes the encoded stream smaller by assigning shorter bit codes to more common colors, particularly if those colors are clustered in the image near each other or there’s repeating patterns. This is a bit of a simplification. Anyway, small compression values shouldn’t have a serious performance impact but might have significant storage savings. I’m sure Blender devs chose 15% as a good compromise.

[@stefanguiton]

Great Video!

[@pixeltrain3d]

Thanks a lot :-)

[@maximoremedios]

Excellent series of tutorials, thanks. You are tracking on footage with distortion I believe. Aren't you supposed to track on undistorted footage?

[@pixeltrain3d]

This is part 01 of 13 tutorials about tracking. Yes, in 3D tracking we only track undistorted plates. I explain more about that in Blender Tracking - 09 Advanced Camera Tracking Workflow

[@maximoremedios]

Excellent! Do you recommend tracking on linear footage (OpenEXR) or footage with a gamma applied? Or it doesn't matter?

[@pixeltrain3d]

Everything which helps you to get a good contrast for your trackers. So, you even can help the tracker by adjusting the contrast or colors, track thatand then use your original footage for your comp.

[@carlitosdemurcia]

Thanks for your Tutos. I have some doubt. ¿perhaps blender VFX tools (like tracking) works better with video footage in PNG. that mpeg, mp4, quiicktime....etc? i have see that sometimes makes strange things with some frames in video footage in video format. And then....¿is the PNG format the better option in performance and quality to work most efficiently Blenderr tracking? In your experience.

[@pixeltrain3d]

Hi Carlos, In my experience you should always work with image sequences for tracking. There are several reasons: - If you exchange the data later with other applications like NUKE, Fusion, AEX there are really often framerate mistakes. Hard to find. In an image sequence you are forces to set the framerate per hand and so you can make sure, it's the same everywhere. - Depending on the software and their caching the decompression of the frames take a lot of time and performance. While working on your tracking, I dont want to have this delay. - Really often you prepare a separate footgate for tracking, in which you adjust the colors, contrast or so for better feature detection. Then you use these camera data for the original shot. So you need to do a separate sequence - Same for getting rid of the lens distortions in a wider shot. We will talk about that later in this series when it come to camera tracking Hope that helps, Helge

[@hatsunemikuchannel2023]

This is "How can do motion tracking with more than one scene", isn't it ? 🤔🤔😯😯 If yes, well, i needed this tutorial for my MikuMikuDance videos on Blender. 😯😯🤗🤗