Cache and multiple file resolutions

[Thread] It was being discussed on github the adoption of microsservices/lambdas for storing multiple informations regarding different media res on minting time.

Could the formats field of the TZIP-21 spec be used to track the various image resolutions:

{
    "name": "Test",
    "description": "Just a test",
    "tags": [
        "art",
        "3d"
    ],
    "symbol": "OBJKT",
    "artifactUri": "ipfs://QmXu6QUvZGUhzPEABCqscyvaycekCQV9KMNRg3NLi6J45p",
    "displayUri": "ipfs://QmbV5e6f9o8hjniXqRMQzoWxippLyTWZxr2mY6HzZkT5xK",
    "thumbnailUri": "ipfs://QmTcFRqJk8zyR1jE8TWFLwYNbDSBAFn2Tr5j3J1rLAJsVq",
    "creators": [
        "tz1XtjZTzEM6EQ3TnUPUQviCD6WfcsZRHXbj"
    ],
    "formats": [{
            "uri": "ipfs://QmXu6QUvZGUhzPEABCqscyvaycekCQV9KMNRg3NLi6J45p",
            "mimeType": "image/jpeg",
            "dimensions": {
                "value": "6144x6144",
                "unit": "px"
            }
        },
        {
            "uri": "ipfs://QmbV5e6f9o8hjniXqRMQzoWxippLyTWZxr2mY6HzZkT5xK",
            "mimeType": "image/jpeg",
            "dimensions": {
                "value": "1024x1024",
                "unit": "px"
            }
        },
        {
            "uri": "ipfs://QmTcFRqJk8zyR1jE8TWFLwYNbDSBAFn2Tr5j3J1rLAJsVq",
            "mimeType": "image/jpeg",
            "dimensions": {
                "value": "350x350",
                "unit": "px"
            }
        }
    ],
    "decimals": 0,
    "isBooleanAmount": false,
    "shouldPreferSymbol": false
}

Also, couldn’t the various image formats such as cover images and thumbnails be generated in the browser during minting without needing any back image processing?

processing in browser is best if it’s possible, but a lot of devices may not be able to handle video processing via browser. It also sounds a lot more error prone compared to a service. I think it’s worth exploring more though.

Another consideration is we can’t update our needs later if it’s only done client side at mint. We already have hundreds of thousands of objkts that will require some kind of service or batch processing. I’ve been trying to explore the batch processing side but it takes a while just to get all of the metadata and ipfs files and then the actual processing isn’t simple. Tarwin did a bunch on that front a while back GitHub - tarwin/hicetnunc-collection-generator

Just curious what kind of processing would be needed for videos?