datavudeja

Stream.sh

stream.sh is a Bash script designed to stream multiple video sources to multiple screens using ffplay. This script ensures that each video stream is displayed on a specified monitor and restarts the stream if it fails. The script is optimized for real-time playback by using various ffplay options to reduce latency and handle frame drops effectively.

Features

Streams multiple video sources to multiple screens.
Ensures each stream is displayed on the specified monitor.
Restarts streams automatically if they fail.
Optimized for real-time playback with minimal latency.

mostly from the url here:

https://stackoverflow.com/questions/24961127/how-to-create-a-video-from-images-with-ffmpeg

-pattern_type glob

This great option makes it easier to select the images in many cases.

Normal speed video with one image per frame at 30 FPS

Send video file as clip

How to use this:

Prepare your video file by using the clip metadata snippet in the Requirement section below.
Press Ctrl+Shift+I to open DevTools and copy everything in Main Function section.

Note

If there is nothing poping up, you can:

Download the ptb client, or
Use this to enable DevTools on stable client.

	#!/usr/bin/env python

	import subprocess
	import tempfile
	import os
	import signal


	class ShellCmd:
	"""Helpful class to spawn commands and keep track of them"""

	import json
	import os
	import sys
	import tempfile
	from typing import cast

	import requests
	from requests.models import Response

	CLOUDFLARE_API_TOKEN = "YOUR_API_TOKEN"

	import os
	import json
	import base64
	import sqlite3
	import win32crypt
	from Crypto.Cipher import AES
	import shutil
	import click

	#from gptcache import cache
	#from transformers import BertModel, BertTokenizer
	from datasets import load_dataset, concatenate_datasets
	from sentence_transformers import SentenceTransformer
	from sklearn.cluster import HDBSCAN, DBSCAN
	from sklearn.metrics import pairwise_distances_argmin_min, silhouette_score, pairwise_distances
	from sklearn.model_selection import ParameterGrid
	from sklearn.neighbors import NearestNeighbors
	from tqdm import tqdm
	from transformers import AutoTokenizer#, BertModel, BertTokenizer, BertConfig, DistilBertModel, DistilBertTokenizer

	# Add a subtitle to a video, set the language=eng, make it the default subtitle and store it as mkv
	# $1 = video file, $2 = subtitle file
	ffmpeg -i "$1" -i "$2" -map 0 -map 1 -metadata:s:0 language=eng -c copy -disposition:s:0 default "$1.mkv"`

	# Add a subtitle to a video, copy just one audio stream, set metadata language=eng for the subtitle, make it the default one and store it as mkv
	ffmpeg -i "$1.mp4" -i "$1.srt" -map 0 -map 1 -map -0:a:0 -c copy -metadata:s:s:0 language=eng -disposition:s:0 default "$1.mkv"

	# Scale down every file to H265 720p using Intel QSV hardware encoding
	find . -type f -exec ffmpeg -i {} -c:v hevc_qsv -s 1280x720 -global_quality 20 -c:a copy -c:s copy {}.mkv \; && find . ! -name '.mkv.mkv' -exec rm {} \; && rename 's/.mkv.mkv/.mkv/'

	# Downloads the video and coverts to audio, then clips it from $start to $stop time and saves as $file
	# The time formats must be "hh:mm:ss" and are relative to the beginning of the source video.
	$youtubeUrl = "https://www.youtube.com/watch?v=T6SyM0Kw6GA"
	$start = "00:00:36"
	$stop = "00:04:55"
	$file = "clip"

	$youtube_id = [regex]::new("[^/=]+$").Match($youtubeUrl).Value
	md $youtube_id -ErrorAction SilentlyContinue
	cd $youtube_id