--time_sample=5 not giving expected results

[PetervanLunteren]

I'm having trouble with the --time_sample argument. I'm probably doing something wrong... See the commands below. See my email if you want to test on the exact same video.

--time_sample=5

Expected: process only one frame every 5 seconds
Result: processes all 385 frames

'/Applications/.EcoAssist_files/miniforge/envs/ecoassistcondaenv-base/bin//python' '/Applications/.EcoAssist_files/cameratraps/megadetector/detection/process_video.py' '--json_confidence_threshold=0.01' '--recursive' '--time_sample=5' '--frame_folder=/Users/peter/Downloads/temp' '--keep_extracted_frames' '--max_width=1280' '--quality=85' '--output_json_file=/Users/peter/Downloads/single-sp/video_recognition_file.json' '/Applications/.EcoAssist_files/models/det/MegaDetector 5a/[md_v5a.0.0.pt](http://md_v5a.0.0.pt/)' '/Users/peter/Downloads/single-sp'

--time_sample=0.5

Expected: process only one frame every 0.5 seconds
Result: processes all 385 frames

'/Applications/.EcoAssist_files/miniforge/envs/ecoassistcondaenv-base/bin//python' '/Applications/.EcoAssist_files/cameratraps/megadetector/detection/process_video.py' '--json_confidence_threshold=0.01' '--recursive' '--time_sample=0.5' '--frame_folder=/Users/peter/Downloads/temp' '--keep_extracted_frames' '--max_width=1280' '--quality=85' '--output_json_file=/Users/peter/Downloads/single-sp/video_recognition_file.json' '/Applications/.EcoAssist_files/models/det/MegaDetector 5a/[md_v5a.0.0.pt](http://md_v5a.0.0.pt/)' '/Users/peter/Downloads/single-sp'

--frame_sample=150

Expected: process one frame every 150 frames
Result: processes 3 frames (0, 150, 300) as expected

'/Applications/.EcoAssist_files/miniforge/envs/ecoassistcondaenv-base/bin//python' '/Applications/.EcoAssist_files/cameratraps/megadetector/detection/process_video.py' '--json_confidence_threshold=0.01' '--recursive' '--frame_sample=150' '--frame_folder=/Users/peter/Downloads/temp' '--keep_extracted_frames' '--max_width=1280' '--quality=85' '--output_json_file=/Users/peter/Downloads/single-sp/video_recognition_file.json' '/Applications/.EcoAssist_files/models/det/MegaDetector 5a/[md_v5a.0.0.pt](http://md_v5a.0.0.pt/)' '/Users/peter/Downloads/single-sp'

I started diving into process_video.py and video_utils.py, but figured that it is probably something stupid on my side... So I thought I'd ask you if you see something wrong with my commands?

[agentmorris]

Thanks for reporting this, and for providing such a complete description. It turns out it was an esoteric bug that only manifested when --keep_extracted_frames was set. Oops! Fixed now, will follow up by email to see whether you also need me to update the Python package. Let me know when you are able to verify, but I will close the issue, because I'm 98% sure it's fixed.

Thanks!