Today we are excited to release Sighthound Video 6.0 Beta 2. Before I go into the list of changes, I would like to thank each and every one of our users who tried Beta 1, and especially those of you who submitted their clips to help us improve the detection engine. Every one of the clips you’ve submitted counts, and has contributed to improvements in the new version.
So, what’s new in Beta 2?
New detection models
With the help of the input you have provided, along with the continued progress of our computer vision scientists, we’ve been able to make marked improvements in our detection rates. This work is a process that never stops, so whenever you feel we could do better or have a specific problem -- please submit that clip to us. Together, we will make it as good as can be!
This functionality allows us to offload some of computer vision work from the main CPU onto the system’s GPU, thus freeing up CPU cycles for other things.
Now GPU offloading is supported for the systems with 6th to 8th generation Intel Core or Xeon processors with Iris® Pro graphics and Intel HD Graphics.
We know many of you have asked for Nvidia GPU support. I want to reiterate our commitment to providing that option, and while timing hasn’t been finalized yet, the hope is to finalize NVidia support later this year.
Performance improvements on MacOS
Beta 1 may have taken a step back in performance on MacOS, compared to SV 5.1. This happened due to the introduction of new computer vision functionality: while we had been able to keep the performance in check on Windows right away, MacOS implementation needed catch-up. Beta 2 accomplishes just that.
The memory leaks we’ve previously mentioned have now been resolved.
We hope the new version works well for you. As always, in case of problems don’t hesitate to reach out to us.