How to use FFmpeg within lambda function via layers

[@nanascho3]

I've been using a pre-existing lambda layer that was available on SAM but it was based on Node 10.x. Can't wait to set up my own with the help of this video!

[@ShubhDevotional]

Thanks buddy, I'm a beginner in AWS and your video helped. You're great

[@SrceCde]

Glad I could help! Please like, share & subscribe. :)

[@hikerbikeromar]

Excellent tutorial, thank you!

[@SrceCde]

You are welcome! Please like, share & subscribe :)

[@kedarundale6850]

Awesome quick tutorial.

[@SrceCde]

Glad it was helpful, Kedar! Please like, share & subscribe :)

[@NonnoSgrenf]

Amazing tutarial! I just have a question: is it strictly necessary to create the zip spinning up an EC2 machine? Can I do it directly in my laptop?

[@SrceCde]

Yes, you can do it on your laptop. Please check out below resources. Video: kzbin.info/www/bejne/e5K1fphneJqDd9k GitHub: github.com/srcecde/aws-lambda-layer-creator I hope this helps. Please like, share & subscribe :)

[@VinodV-uz9wt]

Thanks so much brother!!

[@SrceCde]

You are welcome! Please like, share & subscribe :)

[@mohitkale6879]

Good tutorial for a short time running transcoding job. Can you please create a tutorial for real use-cases where we have large transcoding files like 1080p, 2k, 4k, and 8k videos of size 1-3GB. Lambda function won't be the ideal choice due to longer transcoding time and intensive GPU/CPU usage. Is there a way we can mimic a Transcoding workflow like KZbin does when a video is uploaded?

[@SrceCde]

Thanks for stopping by Mohit Kale. I am glad that you found the tutorial helpful. Yes, you are right that this will work great for a short transcoding job or small task. I will definietly work on the use-case that you suggested soon. Please stay tuned. Please like, share & subscribe. :)

[@mohitkale6879]

Thanks, will be waiting for your tutorial

[@nanascho3]

I'd be interested in this as well -- I've got a hunch that either the transcoding done for larger files for such services is done on the fly via streaming protocols. If not, dedicated instances/virtual machines are running the jobs. An idea that comes to mind but that I haven't tried is to split the video into "chunks" and then run a lambda to process each chunk and then to concatenate them back together. This does introduce a LOT of variables and potential failure points so it's probably not the best way to do it.

[@chinmaykunkikar]

@@SrceCde Hey! Any update on this?

[@aabya134]

really helpful, Thanks for sharing.

[@SrceCde]

Thanks for stopping by Tanmay Dixit. I am glad that you found the tutorial helpful. Please like, share & subscribe. :)

[@vishapatel3434]

Thanks for sharing another excellent video!!

[@valentinoforever]

good tutorial! Why not just using your mac instead of running an ec2 instance? Thanks

[@maghani.official]

Yes, you can compile FFmpeg on macOS, but it's important to remember that the resulting binary might not be compatible with the Amazon Linux environment used by AWS Lambda. If you're compiling FFmpeg for local use or development on your Mac, this approach works fine. However, for AWS Lambda, it's recommended to compile FFmpeg on an Amazon Linux environment to ensure compatibility.

[@ankursingh8377]

When we download file from s3 in aws lambda and again save it to s3(as it was in the video to download audio for transcoding) does that incur data out bandwidth cost or it is just a transfer inside same machine infrastructure .

[@tanmayapatra3001]

"cannot execute binary file" error while executing ffmpeg inside /opt/. Please help.

[@romanempire5333]

Bro can you please tell me how you added cv2 layer in lambda please guide me am trying that for past 1 week 😢

[@saikishorebeeram1131]

amazing tutorial :)

[@SrceCde]

Thank you! Cheers! Please like, share & subscribe. :)

[@lucahighton7648]

This is cool but I don't get it. Why spin up an EC2 instance for that? Can't you just do that same process locally? it seems expensive and more complicated for no real reason. Not trying to offend anyone, this is a great video. I'm just curious.

[@McQuinTrix]

I think this could be useful to do video/audio processing thru API when your files are stored on ECS and dont want to download to your machine & transcode it on AWS itself

[@lucahighton7648]

@@McQuinTrix yeah yeah, the FFMPEG lambda is useful for that. The part I didn't understand was at the beginning when he spun up an EC2 instance and created the FFMPEG zip from inside it. instead of just downloading it from his browser and loading it into s3 storage and then into a lambda layer, or directly into the lambda layer from the browser console or from the CLI SDK

[@saloninagaria1317]

Hi I have uploaded the zip file directly to s3 and then when I am trying to run ffmpeg command then I m getting permission denied error. Also lambda function has the s3 bucket access.

[@SrceCde]

Can you please paste the full traceback?

[@saloninagaria1317]

@@SrceCde sh: /opt/SampleUploader/bin/ffmpeg: Permission denied this is the error which i m getting while running the below line command='/opt/SampleUploader/bin/ffmpeg -i "concat:/tmp/audio1.mp3|/tmp/audio2.mp3" -c copy /tmp/audio3.mp3' os.system(command) help me on this :)

[@SrceCde]

I did test this command and implementation on my end and it's working fine for me. In terms of exception, it's hard to guess the exact issue without looking at your implementation. I can certainly help you with a quick remote session. Please DM me on Instagram/Twitter (The links are in the video description). I hope this will help. Please like, share & subscribe. :)