soundchaser128 / clip-mash Goto Github PK

If a user decides to stop generating clips, goes back, changes video settings and continues generating clips, this will probably lead to a corrupt video

Idea: Don't make the user enter a folder name before being able to see videos, show all videos that exist in the database (mark files that no longer exist at the original location) and add a button to let users add new folders to the library.

Hello

My network has a proxy, and no direct internet access to port 80.

In this case clip-mash dies/locks after just looking for updates

ERROR clip_mash::server::error: request failed: Report(
   0: error sending request for url (https://api.github.com/repos/soundchaser128/clip-mash/releases/latest): error trying to connect: tcp connect error: An attempt was made to access a socket in a way forbidden by its access permissions. (os error 10013)
   1: error trying to connect: tcp connect error: An attempt was made to access a socket in a way forbidden by its access permissions. (os error 10013)
   2: tcp connect error: An attempt was made to access a socket in a way forbidden by its access permissions. (os error 10013)
   3: An attempt was made to access a socket in a way forbidden by its access permissions. (os error 10013)

Location:
   src\service\new_version_checker.rs:39

Hi,
after starting the container on docker, I got the error from the title on the starting page.
I run it on iPad.

• Allow setting a project name before selecting the video source
• Allow resuming existing projects (give a prompt on startup of the UI)
• Store JSON blob for each project in the database

Open questions:

• Version management

Use cargo chef?

I'm running the latest version in docker, some videos (eg one with: 854x480 resolution) gives me an error when the clips encode.
Running the docker with RUST_LOG=debug, this is the ffmpeg command causing the error.

ffmpeg -hide_banner -loglevel warning -ss 7.509 -i "input.mp4" -t 1.877333641052246 -c:v libx264 -preset medium -crf 24 -filter_complex "split [original][copy];[copy] scale=ih16/9:-1, crop=h=iw9/16, gblur=sigma=120, eq=brightness=-0.125[blurred];[blurred][original] overlay=(main_w-overlay_w)/2:(main_h-overlay_h)/2, fps=30, scale=1280:720" -acodec aac -ar 48000 output.mp4

which gives error:
[Parsed_crop_2 @ 000001c442d2d700] Invalid too big or non positive size for width '852' or height '480'
[Parsed_crop_2 @ 000001c442d2d700] Failed to configure input pad on Parsed_crop_2

I commented here:
515cfd3#r137951506
with a possible fix, not sure if it works in all situations. It may have some unwanted side-effects when te target video is 'portrait mode' aspect ratio.

For example, when creating a video is finished, the notification shows up 20+ times.

Notify users when:

• Video generation finished
• Video download finished
• Music download finished

On the clip preview page, users should be allowed to change the order of clips

The API key is actually not required if credentials are disabled in Stash.

Make the video library show all scenes with markers in Stash in addition to the local videos. Move the Stash configuration dialog to the video library as well.

Could expand this to a “PMV mode”, where the length of the video is dictated by the music used.

Allow users to specify the order in which markers (by their title) should occur in the video.

See https://discuss.eroscripts.com/t/clipmash-generate-video-compilations-automatically-v0-15-1/102278/96?u=soundchaser128

Currently, if I create a Tag compilation with two Tags selected, all Markers with either Tag are chosen.

I would like to have a toggle which would allow me to select only Markers with both Tags, instead of markers with either tag.

I'm attempting to use Clip-mash.exe (V0.20.1) on Win 11 for the first time.

What I've done:

1. added videos to my library
2. chose a handful of these videos
3. added markers to the chosen videos
4. Uploaded an mp3 music clip
5. Set the video options to their defaults

When I try the next step (preview clips, I think), I am returned with a Status Code 422 error.

I'm at a loss here, what am I doing incorrectly?

Thank you!

• Create marker for video
• Edit the marker
• Change the range
• Refresh the marker page
• There's two markers when there should only be one

See https://trac.ffmpeg.org/wiki/AudioVolume

Parameters:

• Output codec (h264, h265, VP9,, AV1 etc.)
• Quality (needs to be translated in to the codec-specific options)

• Would have to let the user specify markers in the app
• Would allow to operate entirely without Stash

Should also come up with a different name for the app if this is implemented, since it would no longer rely on Stash

Additional thoughts:

• Either let the user "upload" videos (not sure how much sense that makes, since everything is on the local drive anyways, but it's probably better from a UX standpoint) or just instruct the user to put them into a specific folder (easier implementation). Could also look into local filesystem access like FapInstructor does it.
• For each video, let the user specify markers, i.e. start- and end-points with a specific action. Ideally re-use the existing marker data structure.
• Should markers be persisted somewhere? (if yes, probably in a sqlite database)
• UI-wise, would have to build a page before the current mode selection to let the user choose between accessing videos from Stash or from the local drive.

New name idea: clip-mash

• Start time must be before end time
• Duration must not be zero
• Show validation errors in UI

Hi,
Thank you for sharing the great tool, it gets me nice compilations with ease. but i encountered a problem with dispaying thumbnails in 'select marker/scene' pages.

stash-compilation-maker doesn't give any error messages and I'm wondering how to solve this issue. I appreciate your help.

The current video library might be a little overloaded with functionality, it might make things easier to have one page to manage videos (add, remove, add markers) and another one to just pick the videos to include in the compilation

Currently, the marker names are meaningless. Users should be able to see the list of marker names they specified (show distinct titles from the selected markers) and optionally be able to set how much of each type of marker to include (not sure how to do this yet)

Need to pin down this issue.

Thoughts:

• Need to be able to opt-in during folder scanning (default should be off, I think, at least for now)
• Allow adding images from Stash as well
• Need filters for the grid views.
• How to represent images in the marker view? (possibly just have them as different-looking cards that let you specify the length)

For ffmpeg: https://www.bannerbear.com/blog/how-to-do-a-ken-burns-style-effect-with-ffmpeg/

• Included videos (including their source, so either link to Stash or local path)
• Included music
• Encoding parameters

See for example https://github.com/chrisguttandin/web-audio-beat-detector

The new ideas probably warrant some changes, something more akin to the strategy pattern. Idea:

pub trait ClipCreator {
    type Options;

    fn create_clips(&self, options: Self::Options) -> Vec<Clip>;
}

The user should be able to see if there's a new release available to download.

Report via https://discuss.eroscripts.com/t/clipmash-generate-video-compilations-automatically/

When I hit the reset button at the top after generating a video, the app doesn’t seem to actually reset. It looks like it has, but when I go to generate a new video, it doesn’t process the clips and then it tries to give me a JSON file if I click download?:

Marking scenes is pretty easy, BUT it would be nice to be able to manually enter timecode rather than having to scrub to in/out points

Check if all the source files for the clips have the same encoding parameters, i.e.

• Encoding (h264, h265 etc.)
• Resolution
• FPS
• Not sure if bitrate has to align as well, but probably not?

If yes, don't re-encode every clip, just cut out the videos and concatenate them.

https://discuss.eroscripts.com/t/clipmash-generate-video-compilations-automatically-v0-15-1/102278/100?u=soundchaser128

When a source video itself has cuts, the user should be able to mark them in the UI to guide the clip generation to not include that cut in a clip

Would probably require using channels for communication instead of a shared mutex for the progress, as is now

Probably makes sense to wait for further dependencies that have transitive dependencies to hyper 0.14 to go to 1.0

• reqwest
• sentry
• youtube_dl

I recently upgraded from clipmash 0.16 -> 0.17. It seems harder now to create a compilation of all Markers with a specific tag from my Stash in 0.17.

For instance, if I want to create a compilation of all markers from stash with the "Riding" tag is there a way to do that without needing to add every stash video to my clipmash library

Acceptance Criteria

• Allow users to select all markers with A specific tag from entire stash library
• Allow users to select all markers with ALL selected tags from entire stash library
  • eg) "Riding" AND "Red head"