This branch is currently under development and is provided as is
- Convert All Code to TypeScript (done)
- Use FFmpeg Layers for Lambda instead (done)
- Make the thumbnailer more moduler (in-progress)
- Implement Tests (to be done)
This package was created to automatically generate thumbnails from s3 video uploads.
Install this package using the following the command then require the package in your code.
npm install lambda-video-thumbnailerYou can use the provided s3example.js to get you started. Or you can use the following code snippet for a lambda function.
// Require the library
const lvt = require("lambda-video-thumbnailer");
exports.handler = function (event, context) {
// Returns a S3Thumbnailer Class
const t = new lvt.S3(event.Records[0].s3);
// Create a thumbnail will return promise<string>
return t.createThumbnail({
width: 100,
height: 100,
time: "00:00:05",
type: "webp",
});
};createThumbnail() takes a thumbnailConfig object with that has the following default options
{
"prefix": "thumbnails/",
"width": 180,
"height": 180,
"time": "00:00:05",
"type": "jpg",
"quality": 2,
"thumbnailOption": "default"
}To configure the output size of thumbnails simply supply a width/height number. If you want to automatically calculate width or height based on the aspect ratio of the input, just use -1. For example:
{
"width": 700,
"height": -1
}This will produce a 700x300 image when the input aspect ratio is 21:9
If your output needs to be a multiple of n then supply -n and it will output an output that is scaled by the ratio and a multiple of -n. For example
{
"width": -3,
"height": 250
}This will produce a 441x250 image when the aspect raito is 21:9
To configure the output type of thumbnails just set the type to "jpg", "png" or "webp"
{
"type": "webp"
}NOTE: ffmpeg with png tends to output very large filesizes even with -pred mixed
The output quality of the thumbnail is controlled by the quality property. It is controlled on a scale of 1-10 with 1 being the best quality and 10 being the worst quality.
{
"quality": 1
}This option determines which filter to use for capturing thumbnails via ffmpeg. The options currently are default,none. The default option will set the filter to thumbnail this will look for the best frame within the select timeslot, this can be use quite a bit of memory on large files but will generally give the best screenshots. none will not use any filter and pick the first frame at the given timeslot.
{
"thumbnailOption": "none"
}The output bucket is defined by the outputBucket thumbnail config option, by default this will output the files to the source bucket if left blank.
{
"outputBucket": "mybucket"
}By default the output file name will use the source file with the new type suffix. For example movie_trailer.mp4 becomes move_trailer.jpg when used with the jpg type output.
However you can customize it with the outputKey option in your thumbnail config. It will still attach the type or replace it my_new_filename.hello becomes my_new_filename.jpg
{
"outputKey": "my_new_filename"
}This package requires a ffmpeg layer to function. You can use the following arns:
Or alternatively use the serverlesspub build (only available in ap-us-east-1) arn:aws:lambda:us-east-1:145266761615:layer:ffmpeg:4 or build your own from their github repo serverlesspub/ffmpeg-aws-lambda-layer
Then just go onto AWS Lambda create a new function and give it a trigger of put objects with a suffix filter of _.avi, _.mov etc..
Although you don't need to do this it just reduces the number of invocations. Apply a policy of LambdaExecute to the IAM role (this gives read and write access to s3) then set the memory limit to the maximum* and set timeout for 30seconds (again it normally takes around 6 seconds but its good to be safe).
*NOTE: I have never reached the maximum with this function, its normally around 700mb even for 4gb mov files
arn:aws:lambda:ap-east-1:260572601982:layer:ffmpeg:1- Hong Kong (AP East 1)arn:aws:lambda:ap-south-1:260572601982:layer:ffmpeg:1- Mumbai (AP South 1)arn:aws:lambda:ap-northeast-2:260572601982:layer:ffmpeg:1- Seoul (AP Northeast 2)arn:aws:lambda:ap-southeast-1:260572601982:layer:ffmpeg:1- Singapore (AP Southeast 1)arn:aws:lambda:ap-southeast-2:260572601982:layer:ffmpeg:1- Sydney (AP Southeast 2)arn:aws:lambda:ap-northeast-1:260572601982:layer:ffmpeg:1- Tokyo (AP Northeast 1)
arn:aws:lambda:eu-central-1:260572601982:layer:ffmpeg:1- Frankfurt (EU Central 1)arn:aws:lambda:eu-west-1:260572601982:layer:ffmpeg:1- Ireland (EU West 1)arn:aws:lambda:eu-west-2:260572601982:layer:ffmpeg:1- London (EU West 2)arn:aws:lambda:eu-west-3:260572601982:layer:ffmpeg:1- Paris (EU West 3)arn:aws:lambda:eu-north-1:260572601982:layer:ffmpeg:1- Stockholm (EU North 1)
arn:aws:lambda:ca-central-1:260572601982:layer:ffmpeg:1- Central (CA Central 1)
arn:aws:lambda:sa-east-1:260572601982:layer:ffmpeg:1- São Paulo (SA East 1)
arn:aws:lambda:us-east-1:260572601982:layer:ffmpeg:1- North Viginia (US East 1)arn:aws:lambda:us-east-2:260572601982:layer:ffmpeg:1- Ohio (US East 2)arn:aws:lambda:us-west-1:260572601982:layer:ffmpeg:1- N. California (US West 1)arn:aws:lambda:us-west-2:260572601982:layer:ffmpeg:1- Oregon (US West 2)
Originally this was a tutorial I wrote back in July 2017, I updated it a bit since then. The original also contained mediainfo. Available Here