I am assuming that you have market research, and I am curious...@Klitz @Ml22 @TesseractOrion This has nothing to do with whether the phones can process the video - it’s about whether they can process video in real-time. The only reason we recommend iPhone SE 2nd gen to Android folks is because that is the cheapest phone on the market that can do real-time processing with SwingVision (less than $200 on Amazon). We have dozens of folks switch to iOS every week just for SwingVision so we try to recommend the most economical option. Of course anything newer will be even better.
We have tens of thousands of customers and >95% of them use A13 or newer chips. Those folks have a 10x better experience than folks on older devices that can’t do real-time processing. It’s night and day. Plus they get access to real-time features like line challenges and audio feedback.
As such, we will not go to Android until there are a few flagship devices that are capable of real-time processing.
There’s really no other agenda here - we’d love to serve as many people as possible but will only do so if it’s a good experience. Geekbench ML results are publicly available - there have been no Android devices that can match iPhone 11 in NPU (ie AI performance) prior to Pixel 6. You’re welcome to see the results yourself here:https://browser.geekbench.com/ml-benchmarks
I know this might be hard to accept but Apple‘a Neural Engine really is several years ahead of the industry. I was pretty shocked to see this given my experience at Tesla Autopilot. Fortunately Google’s doing some amazing work with their Tensor chips and Qualcomm seems to have caught up this year.
Android for SwingVision is on the way, assuming this macro environment doesnt cause us any setback with hiring/layoffs![]()
Do tennis players, when compared to general population, use IPhones and Androids at a higher, lower, or the same distribution percentages?