Recherche avancée

Sur d’autres sites (6036)

• Inserting clips into mp4 via ffmpeg

  15 septembre 2024, par lfgan

  I am looking to create mp4 files by combining multiple static mp4s and mp3s.
I have a script already to convert mp3 to mp4 and using the album art as a static background image, and inserting the intro.
What I am having problems with is that it cannot concatenate the videos reliably, or at all for that matter.

  


  This is the code that I currently have, it is ChatGPT, however I gave up as it seemed like it was just running in circles.

  


  setlocal

:: Set variables
set "mp3file=myfile.mp3"
set "coverart=extracted_cover.jpg"
set "outputfile=music.mp4"
set "introfile=intro.mp4"
set "outputendfile=Outro.mp4"
set "wowfile=Wow!.mp4"

:: Step 1: Extract cover art from MP3
ffmpeg -i "%mp3file%" -an -vcodec copy "%coverart%"
if %errorlevel% neq 0 (
    echo Error: Failed to extract cover art from MP3 file.
    exit /b 1
)

:: Step 2: Convert MP3 to MP4 using the extracted cover art
ffmpeg -loop 1 -i "%coverart%" -i "%mp3file%" -c:v libx264 -tune stillimage -c:a aac -b:a 192k -shortest "%outputfile%"
if %errorlevel% neq 0 (
    echo Error: Conversion from MP3 to MP4 failed.
    exit /b 1
)

:: Step 3: Randomly generate two positions for Wow!.mp4 insertions
:: Random positions will be between 10 and 50 seconds
set /a "pos1=%random% %% 40 + 10"
set /a "pos2=%random% %% 40 + 10"

:: Ensure pos1 and pos2 are different
if %pos1% EQU %pos2% (
    set /a "pos2=%pos1%+10"
)

:: Print positions for debugging
echo Random insertion positions:
echo Position 1: %pos1%
echo Position 2: %pos2%

:: Step 4: Split the music mp4 into parts at the random positions
:: Part 1: From the beginning to pos1
ffmpeg -i "%outputfile%" -c copy -ss 0 -to %pos1% -y "part1.mp4"
if not exist "part1.mp4" (
    echo Error: Failed to create part1.mp4
    exit /b 1
)

:: Part 2: From pos1 to pos2
ffmpeg -i "%outputfile%" -c copy -ss %pos1% -to %pos2% -y "part2.mp4"
if not exist "part2.mp4" (
    echo Error: Failed to create part2.mp4
    exit /b 1
)

:: Part 3: From pos2 to the end
ffmpeg -i "%outputfile%" -c copy -ss %pos2% -y "part3.mp4"
if not exist "part3.mp4" (
    echo Error: Failed to create part3.mp4
    exit /b 1
)

:: Step 5: Concatenate the files in the required order
(
    echo file 'intro.mp4'
    echo file 'part1.mp4'
    echo file 'Wow!.mp4'
    echo file 'part2.mp4'
    echo file 'Wow!.mp4'
    echo file 'part3.mp4'
    echo file 'Outro.mp4'
) > concat_list.txt

ffmpeg -f concat -safe 0 -i concat_list.txt -c copy final_output.mp4
if %errorlevel% neq 0 (
    echo Error: Failed to concatenate the video files.
    exit /b 1
)

:: Cleanup
del part1.mp4 part2.mp4 part3.mp4 concat_list.txt "%coverart%"

echo Done! The final video is saved as final_output.mp4


  


  exactly what it is supposed to do :

  


  

  1. Convert "myfile.mp3" to "music.mp4" with image (works)

  


  2. Add "intro.mp4" to the start of "music.mp4" and "Output.mp4" to the end (works)

  


  3. Split "music.mp4" into 3 parts, randomly, at a minimum 10s after intro.mp4 and 10s before Outro.mp4 (can for some reason only create part 1)

  


  4. Put all clips in order (unknown, probably works but part 2 and 3 wont generate so it cannot get to that step.)

  


  


  Edit :
I forgot to add the error message that I am getting. Here it is :

  


  [out#0 @ 000002e024a6ed00] -to value smaller than -ss; aborting.
Error opening output file part2.mp4.
Error opening output files: Invalid argument
Error: Failed to create part2.mp4


  


• Pyinstaller Hidden import 'ffmpeg-python' not found

  13 août 2024, par petunia rose

  Trying to convert Python scripts to exe with PyInstaller.

  


  In my code, I use ffmpeg-python :

  


  import ffmpeg
....
def ffmpeg_save_clip(self,output_video: str, clip_start: str, clip_end: str): 
  (ffmpeg 
   .input(self.file.get_videopath(), ) 
   .output(output_video, vcodec='copy', ss=clip_start, to=clip_end, acodec='copy') 
   .global_args('-y') 
   .run())


  


  So Ii call PyInstaller from terminal with related flag :

  


  pyinstaller --windowed --hidden-import "ffmpeg-python" --noconsole --icon=app.ico --name "app" main.py 


  


  I checked also :

  


  pip install ffmpeg-python 
Requirement already satisfied: ffmpeg-python in c:\python38\lib\site-packages (0.2.0) 
Requirement already satisfied: future in c:\python38\lib\site-packages (from ffmpeg-python) (0.18.3)


  


  But I get from PyInstaller :

  


  Hidden import 'ffmpeg-python' not found


  


  App works in visual-studio, but when I run pyinstaller exe and try to save clip, it freeze without response.

  


  Note : I also :

  


  


  1. try to add ffmpeg.exe into root folder of pyinstaller app.

     


  


  2. try to use .spec file with :

     


     binaries=[('C:\\ffmpeg\\bin\\ffmpeg.exe', 'ffmpeg/),('C:\\ffmpeg\\bin\\ffplay.exe','ffplay/'), ('C:\\ffmpeg\\bin\\ffprobe.exe', 'ffprobe/')],

     


  


  


  Nothing changes

  


  UPDATE
Tank you @happy_code_egg, you explained me such a things.
I tried what you said and I don't have error any more.

  


  But I can't make work ffmpeg anywhere on Pyinstaller exe.
I added a try/except block to isolate problem
(when I launch app on Visualstudio it works, but not when use exe).

  


  block is (very symilar to past def) :

  


  def ffmpeg_save_clip(self,output_video: str, clip_start: str, clip_end: str):
    input = 'C:\Users\x\Desktop\input.mp4'
    output = 'C:\Users\x\Desktop\output.mp4'
    try:
        (ffmpeg
        .input(input)
        .output(output, vcodec='copy', ss=clip_start, to=clip_end, acodec='copy')
        .global_args('-y')
        .run(capture_stdout=True))
    except Exception as e:
        self.logger.error('Video window -> ffmpeg_save_clip -> error ' + str(e))
        self.logger.error('Video window -> ffmpeg_save_clip -> input file: ' + input)
        self.logger.error('Video window -> ffmpeg_save_clip -> output file ' + output)
        raise ValueError(str(e))


  


  Note : (input and output fixed are only as simple path examples)

  


  When I open log I have :

  


  2024-08-09 11:28:50,293 - ERROR - Video window -> ffmpeg_save_clip -> error [WinError 2] File not found error
2024-08-09 11:28:50,293 - ERROR - Video window -> ffmpeg_save_clip -> input file: C:\Users\x\Desktop\input.mp4
2024-08-09 11:28:50,293 - ERROR - Video window -> ffmpeg_save_clip -> output file C:\Users\x\Desktop\ouput.mp4


  


  (I tried with '/' and also '\' to create path)
File paths are correct and, on Visualstudio, they can be processed by ffmpeg (no error), but the exe seems to not find (input or output)

  


  I tried solution described here :
Python ffmpeg won't accept path
(without effects)

  


  I found also these discussions :

  


  github ffmpeg-python issue1

  


  github ffmpeg-python issue2

  


  in both discussions seems to be a problem with ffmpeg + ffmpeg-python libs cohexistence.

  


  I tried (as said) to uninstal both and reinstall ffmpeg-python,
but nothing change.

  


  Note :
I've also three other warning during Pyinstaller :
(I write them only to complete context, but i think they have no repercussions on this problem.)

  


  57283 ERROR: Hidden import 'fiona._shim' not found
57298 WARNING: Hidden import "fiona._shim" not found!
57392 WARNING: Hidden import "importlib_resources.trees" not found!


  


  tried with :

  


  --hidden-import fiona._shim --hidden-import fiona.schema


  


  tried also with :

  


  --add-data="fiona/*;fiona”


  


  but warnings still remains.

  


• FFMPEG Concatenating videos with same 25fps results in output file with 3.554fps

  5 juin 2024, par Kendra Broom

  I created an AWS Lambda function in node.js 18 that is using a static, ver 7 build of FFmpeg located in a lambda layer. Unfortunately it's just the ffmpeg build and doesn't include ffprobe.

  


  I have an mp4 audio file in one S3 bucket and a wav audio file in a second S3 bucket. I'm uploading the output file to a third S3 bucket.

  


  Specs on the files (please let me know if any more info is needed)

  


  Audio :
wav, 13kbps, aac (LC), 6:28 duration

  


  Video :
mp4, 1280x720 resolution, 25 frame rate, h264 codec, 3:27 duration

  


  Goal :
Create blank video to fill in the duration gaps so the full audio is covered before and after the mp4 video (using timestamps and duration). Strip the mp4 audio and use the wav audio only. Output should be an mp4 video with the wav audio playing over it and blank video for 27 seconds (based on timestamp) until mp4 video plays for 3:27, and then blank video to cover the rest of the audio until 6:28.

  


  Actual Result :
An mp4 file with 3.554 frame rate and 10:06 duration.

  


  import { S3Client, GetObjectCommand, PutObjectCommand } from "@aws-sdk/client-s3";
import { createWriteStream, createReadStream, promises as fsPromises } from 'fs';
import { exec } from 'child_process';
import { promisify } from 'util';
import { basename } from 'path';

const execAsync = promisify(exec);

const s3 = new S3Client({ region: 'us-east-1' });

async function downloadFileFromS3(bucket, key, downloadPath) {
    const getObjectParams = { Bucket: bucket, Key: key };
    const command = new GetObjectCommand(getObjectParams);
    const { Body } = await s3.send(command);
    return new Promise((resolve, reject) => {
        const fileStream = createWriteStream(downloadPath);
        Body.pipe(fileStream);
        Body.on('error', reject);
        fileStream.on('finish', resolve);
    });
}

async function uploadFileToS3(bucket, key, filePath) {
    const fileStream = createReadStream(filePath);
    const uploadParams = { Bucket: bucket, Key: key, Body: fileStream };
    try {
        await s3.send(new PutObjectCommand(uploadParams));
        console.log(`File uploaded successfully to ${bucket}/${key}`);
    } catch (err) {
        console.error("Error uploading file: ", err);
        throw new Error('Failed to upload file to S3');
    }
}

function parseDuration(durationStr) {
    const parts = durationStr.split(':');
    return parseInt(parts[0]) * 3600 + parseInt(parts[1]) * 60 + parseFloat(parts[2]);
}

export async function handler(event) {
    const videoBucket = "video-interaction-content";
    const videoKey = event.videoKey;
    const audioBucket = "audio-call-recordings";
    const audioKey = event.audioKey;
    const outputBucket = "synched-audio-video";
    const outputKey = `combined_${basename(videoKey, '.mp4')}.mp4`;

    const audioStartSeconds = new Date(event.audioStart).getTime() / 1000;
    const videoStartSeconds = new Date(event.videoStart).getTime() / 1000;
    const audioDurationSeconds = event.audioDuration / 1000;
    const timeDifference = audioStartSeconds - videoStartSeconds;

    try {
        const videoPath = `/tmp/${basename(videoKey)}`;
        const audioPath = `/tmp/${basename(audioKey)}`;
        await downloadFileFromS3(videoBucket, videoKey, videoPath);
        await downloadFileFromS3(audioBucket, audioKey, audioPath);

        //Initialize file list with video
        let filelist = [`file '${videoPath}'`];
        let totalVideoDuration = 0; // Initialize total video duration

        // Create first blank video if needed
        if (timeDifference < 0) {
            const blankVideoDuration = Math.abs(timeDifference);
            const blankVideoPath = `/tmp/blank_video.mp4`;
            await execAsync(`/opt/bin/ffmpeg -f lavfi -i color=c=black:s=1280x720:r=25 -c:v libx264 -t ${blankVideoDuration} ${blankVideoPath}`);
            //Add first blank video first in file list
            filelist.unshift(`file '${blankVideoPath}'`);
            totalVideoDuration += blankVideoDuration;
            console.log(`First blank video created with duration: ${blankVideoDuration} seconds`);
        }
        
        const videoInfo = await execAsync(`/opt/bin/ffmpeg -i ${videoPath} -f null -`);
        const videoDurationMatch = videoInfo.stderr.match(/Duration: ([\d:.]+)/);
        const videoDuration = videoDurationMatch ? parseDuration(videoDurationMatch[1]) : 0;
        totalVideoDuration += videoDuration;

        // Calculate additional blank video duration
        const additionalBlankVideoDuration = audioDurationSeconds - totalVideoDuration;
        if (additionalBlankVideoDuration > 0) {
            const additionalBlankVideoPath = `/tmp/additional_blank_video.mp4`;
            await execAsync(`/opt/bin/ffmpeg -f lavfi -i color=c=black:s=1280x720:r=25 -c:v libx264 -t ${additionalBlankVideoDuration} ${additionalBlankVideoPath}`);
            //Add to the end of the file list
            filelist.push(`file '${additionalBlankVideoPath}'`);
            console.log(`Additional blank video created with duration: ${additionalBlankVideoDuration} seconds`);
        }

        // Create and write the file list to disk
        const concatFilePath = '/tmp/filelist.txt';
        await fsPromises.writeFile('/tmp/filelist.txt', filelist.join('\n'));

        const extendedVideoPath = `/tmp/extended_${basename(videoKey)}`;
        //await execAsync(`/opt/bin/ffmpeg -f concat -safe 0 -i /tmp/filelist.txt -c copy ${extendedVideoPath}`);
        
        // Use -vsync vfr to adjust frame timing without full re-encoding
        await execAsync(`/opt/bin/ffmpeg -f concat -safe 0 -i ${concatFilePath} -c copy -vsync vfr ${extendedVideoPath}`);

        const outputPath = `/tmp/output_${basename(videoKey, '.mp4')}.mp4`;
        //await execAsync(`/opt/bin/ffmpeg -i ${extendedVideoPath} -i ${audioPath} -map 0:v:0 -map 1:a:0 -c:v copy -c:a aac -b:a 192k -shortest ${outputPath}`);

        await execAsync(`/opt/bin/ffmpeg -i ${extendedVideoPath} -i ${audioPath} -map 0:v:0 -map 1:a:0 -c:v copy -c:a aac -b:a 192k -shortest -r 25 ${outputPath}`);
        console.log('Video and audio have been merged successfully');

        await uploadFileToS3(outputBucket, outputKey, outputPath);
        console.log('File upload complete.');

        return { statusCode: 200, body: JSON.stringify('Video and audio have been merged successfully.') };
    } catch (error) {
        console.error('Error in Lambda function:', error);
        return { statusCode: 500, body: JSON.stringify('Failed to process video and audio.') };
    }
}


  


  Attempts :
I've tried re-encoding the concatenated file but the lambda function times out. I hoped that by creating blank video with a 25fps and all the other specs from the original mp4, I wouldn't have to re-encode the concatenated file. Obviously something is wrong, though. In the commented out code you can see I tried specifying 25 or not, and also tried -vsync and no -vsync. I'm new to FFmpeg so all tips are appreciated !