2025-04-25, 04:51 PM
(This post was last modified: 2025-04-25, 04:53 PM by Efficient_Good_5784. Edited 1 time in total.)
While you can make the trickplay task use a GPU, it will actually make the process slower than on a CPU.
What I have been told before is that unless your CPU is really low-powered and slow, it's better to just let the CPU handle the creation.
That said, trickplay creation has an option to use video keyframes, which is what you want to use if you want this to go by faster.
The reason keyframes makes things go faster is because normally, the trickplay task needs to process a video frame by frame. As it's going frame by frame, it's keeping track of the time and obtaining a frame at the desired interval that it will then convert to a jpg image for the trickplay file. The process of going through each frame takes time.
The reason you need to go through each frame in order is because modern video codecs have different types of frames. The majority of the frames present in an encoded video are not full images by themselves, but are the difference between the last few frames and itself at that point in time in the video.
Keyframes (also known as I-frames) are 100% full images that don't need any other frames to show the full picture. Depending on how the video was encoded, you will have a number of these all throughout the video. You can quickly pull these up since they are already complete by themselves.
By choosing to use keyframes in Jellyfin's trickplay task, what you're doing is finding the closest keyframe relative to each time period you set for the task and using that for the video scrubbing preview. You'll be trading accuracy of your time interval for images, but using keyframes is super fast compared to the traditional way to create them.
So tl;dr is to enable the keyframe option if you want this to go by fast.
What I have been told before is that unless your CPU is really low-powered and slow, it's better to just let the CPU handle the creation.
That said, trickplay creation has an option to use video keyframes, which is what you want to use if you want this to go by faster.
The reason keyframes makes things go faster is because normally, the trickplay task needs to process a video frame by frame. As it's going frame by frame, it's keeping track of the time and obtaining a frame at the desired interval that it will then convert to a jpg image for the trickplay file. The process of going through each frame takes time.
The reason you need to go through each frame in order is because modern video codecs have different types of frames. The majority of the frames present in an encoded video are not full images by themselves, but are the difference between the last few frames and itself at that point in time in the video.
Keyframes (also known as I-frames) are 100% full images that don't need any other frames to show the full picture. Depending on how the video was encoded, you will have a number of these all throughout the video. You can quickly pull these up since they are already complete by themselves.
By choosing to use keyframes in Jellyfin's trickplay task, what you're doing is finding the closest keyframe relative to each time period you set for the task and using that for the video scrubbing preview. You'll be trading accuracy of your time interval for images, but using keyframes is super fast compared to the traditional way to create them.
So tl;dr is to enable the keyframe option if you want this to go by fast.
