Mixed editing of 8bit and 10bit files

Report · Dec 17, 2022

Do I need to do any processing or setting in advance to edit 8bit and 10bit files together? Will the system reduce the precision of 10bit files for editing? When outputting, the maximum color depth is selected, which means outputting a 32bit file? thanks

Report · Dec 17, 2022

Premiere internally processes in 32-bit float. So ... an 8-bit clip is converted mathematically into 32-bit float, just as is a 10 or 12 bit file. The program doesn't care about the differences. But more bit depth in the clip means more data for the program to work with.

That's the working pipeline. (A colorist term, but useful.)

The image sent to the monitor depends on the bit space of the file being viewed at this moment. Sort of. (It's a complicated discussion to completely cover ... and yes, I have dealt with that in published works.) This is another 'pipeline'.

The export depth of the file will be what you choose to do. There are presets built that are in format/codec likely to be 8-bit, some that will more likely be 10 bit.

You can set the export to be 10 bit via the options in the Export process.

However ... exporting an 8 bit file in a 10 bit export does not magically get more data into the file. It simply 'spreads' the data over those 10 bits.

So back to your question for a timeline with mixed 8 and 10 bit files, a very common occurence.

If you export with 10 bit settings, the exported file, the file itself, will be fully 10 bit. But ... there will be more image data for those images that are 10 bit than for the sections that were 8 bit. More likely to get say banding in the sections of the file from an 8 bit original clip.

Neil

Report · Dec 18, 2022

thank you for your reply. I chose “render at maximum depth” here, doesn't that mean the output file is a 32bit video?

Report · Dec 18, 2022

No. And first, there aren't any 32 bit video formats period.

Second, that setting means that Premiere will process the output to the depth of the preset/settings. Most ProRes for example is 10 bit. Some RGB formats, mostly DPX like image-sequence formats, are 12 bit.

So at most, you will get a 10 bit file, as that is the bit-depth of most relatively 'heavy' video.

But again, choosing a say 10 bit preset when you only have 8 bits of data is just making a bigger file. There's no more data there. Just spread out with gaps in-between.

Neil