You might have already heard about our Video Coach competition through our socials or directly within the Carv app, and we're incredibly thankful for everyone participating. But why are these video sessions so crucial?

Here's the inside scoop on how your footage is revolutionizing All-Terrain coaching from the ground (or snow) up.

Let’s start with the basics - what is a video session?

A "Video Session" refers to footage recorded with Carv's in-app feature, Video Coach. In a video segment, your skiing is recorded and Carv overlays your Ski:IQ score to visually track your performance.

Introduced in 2022, the Video Coach feature was designed to offer a new perspective on learning from your Carv data by showing how technique adjustments affect your Ski:IQ in real-time.

The Evolution of Video Sessions

Video sessions have always been integral to improving our system—we're not just using them to unlock this next era of Carv; we've used them at every step along the way… but this process looks a little different outside of Video Coach mode.

To build our Ski:IQ system and continue improving it, Charlie and Sam, our data scientists, need to collect and categorize thousands of skiing segments by skier turn type, ability level, and terrain to build a reliable database.

How labeling works

Labeled data is a driving force behind almost all of Carv's machine learning in some way, and video sessions are the gold standard. This process involves our experts watching user videos, noting exactly what is happening, and matching these details to the data that Carv records.

For example, we label the type of turn that someone is making. This then trains the machine learning model to better understand the granular differences in our data between a carving, skidded, short or long turn.

Matt, our Sales and Membership Lead, has taken on this process over the past month and labeled a whopping 480 videos from our users. With his background as a ski instructor in Japan, Matt can pick out a variety of turn characteristics and techniques, as well as judging ability level in your videos. This will all go into a huge database that feeds into the algorithm, making it smarter and more sensitive to terrain type than ever before.

Taking this process to All-Terrain

Now, we're applying this methodology to all-terrain skiing. By analyzing video sessions across diverse mountain terrains, we aim to train our models to recognize various snow surfaces - such as fresh powder, bumps, and crud - and adapt our coaching techniques to these conditions.

This process began last summer, and as Carv users this season will know, we’ve already come a long way!

• With Terrain Adjustment in Ski:IQ Hintertux, Carv knows when you’re developing your off-piste skillset or shredding on steeper terrain, and adjusts your scoring accordingly to reward your hard work.
• Our ungroomed snow tips, which recognize the unique difficulties of powder and bumps, and suggest new ways to apply carving techniques to trickier terrain have also received a lot of love.

But we want to take this one step further...

Our current mission

Initial Video collections: behind the scenes

To get the ball rolling for the data science team, the Carv team set up shop in Axamer, Austria, at the ‘Skinnovation’ convention with the aim of capturing as many video coach sessions as humanly possible. This was a blast, and a huge success as well, as we got to meet Carv users of all abilities and really focus in on video analysis on a day that was mixed visibility at best.

The event brought home a respectable 300 videos, ranging across abilities and terrain difficulties, to our data science team.

Thanks to Beth, Imogen, Laura and Sam for all their hard work!

Our current objective

We’re running a challenge to collect as many videos from you guys as possible so we can speed up the journey to All-Terrain coaching.

We need videos of every ability level in all the non-piste conditions, be it powder, bumps, chopped-up snow, ice, crud or anything else.

We’ve set ourselves the challenge of getting 1000 videos by the end of April to fuel our data science team.

Your All-Terrain videos will help us with:

• Labeling and analyzing different skiing techniques in varied terrains.
• Enabling our professional instructors to create personalized coaching insights.
• Integrating these coaching insights into the Carv app for a comprehensive skiing experience.

To find out more about the All-Terrain Challenge, and to be in with a chance of winning Custom Carv Skis 🎿 click on the link below.