You can’t believe everything you read on the internet 😉

 

I have made over 1,000,000 million 16-bit and 32-bit TIFFs compressed LZW over the past 14 years on this film and not aware of single one that was larger than uncompressed. Compression is not super-efficient but it works. That’s a user error or software bug – there have been bugs in 16-bit LZW compression but not aware of anything current and certainly never see it.

Here’s hard data for you from current tasks. As a control, I made a 8192 x 8192 file in Photoshop, 16-bit, black layer and applied 100% gaussian chromatic noise.

 

1. Uncompressed: 393 MB
2. TIFF LZW: 106 MB
3. I currently have 19,661 16-bit TIFF files from AE 2019 renders, all 8192 x 8192. Largest is 347 MB – very high noise and detail, smallest is 1 MB (all black frame).

 

So unless you can post results that show a problem with LZW and 16-bit, I call BS on this.

 

Stephen

 

From: After Effects Mail List <AE-List@media-motion.tv>
Sent: Tuesday, February 26, 2019 6:13 PM
To: After Effects Mail List <AE-List@media-motion.tv>
Subject: Re: [AE] Preferred intermediate image sequence for Ae Pipeline

 

 

On 27 Feb 2019, at 10:00 am, Stephen van Vuuren <AE-List@media-motion.tv> wrote:

 

And I’ve never seen a problem with compressed files larger than originals – you need to back that one up with motion footage sequences. Uncompressed anything 6K and 8K frames are massive and would have broken workflows if LZW  compression does not work.

 

 

I can see how LZW might work OK for mograph elements on a back background, but otherwise it’s just not a good algorithm for real-world footage.  I just googled it, here’s the first link that came up.  Note the file size comparison at the bottom as well as the cautionary note:

 

https://havecamerawilltravel.com/photographer/tiff-image-compression/

 

LZW works fine for 8 bit, just not 16bit images.

 

-Chris