This article began with a question from David:
“When exporting media, would ProRes 422 have a speed advantage over ProRes 422 LT?”
I was fascinated by this question, because I haven’t look at the speed differences between exporting different versions of ProRes before.
David was asking in the context of Apple Final Cut Pro X, but I thought it would be cool to compare the speeds of Premiere with Final Cut. Except…! As I got into my tests, I realized that comparing Premiere to FCP would be like comparing oranges to ducks. They handle rendering and export very differently.
There’s no “right” approach – but it is important to know the differences.
For these tests, I’m using a 2017 27″ iMac. Different computers will export at different speeds, so, in these tests, look at the relationships between speeds, not the actual speed itself.
All files were stored to the internal SSD; here’s the speed of my storage. (As you’ll discover, exporting does not max out the bandwidth available from the SSD, which means that the storage bandwidth of this system is more than sufficient.)
For a source file, I took one of my recent webinar masters:
To this I applied a single effect: scaling the entire image 50%. All tests were run with the same file and the same effect, on the same computer.
NOTE: Applying more, or different, effects will create different export times. Scaling renders reasonably quickly, which is why I chose it.
Then, in Premiere, I exported a media file and, at the top of Export Settings panel, selected each of the relevant ProRes codecs in turn to measure speed.
NOTE: I didn’t export ProRes 4444 with alpha channel, as that is an interim format used to create media for other projects. And, while I can’t envision a reason to export an entire project using ProRes 422 Proxy files, I figured someone would ask, so I timed this, too.
One last comment. Export speeds may slow when Premiere is not the foreground application or when the computer is doing other work. To avoid this potential issue, I left Premiere as the foreground application and avoided doing other work with the computer. (I, ah, read a book…)
Here are the results:
THOUGHTS ON OUTPUT
Where possible, I always recommend exporting a master file at the highest quality possible. It is easy to convert a high-quality file into something smaller. It is impossible to convert a small file into high-quality.
The only exception to this rule is when getting a clip posted quickly is more important than quality. In which case, export a compressed file – H.264 or HEVC – and post that. However, as soon as possible after you get your story posted, go back and export a high-quality master file and save that for the future. Just in case…
To be most effective, Premiere requires a fairly beefy system. Even though I was exporting a 720p project, the 8 GB of RAM on this system was insufficient. During export for ALL versions of ProRes the computer needed more, as evidenced by Swap Used being greater than zero.
Based on this, my recommendation is that any system that primarily uses Premiere should contain at least 16 GB of RAM. (The trade-off is potentially significantly slower exports.)
Premiere accessed storage at wildly different rates but, on average, around 50 MB/second. This is pretty slow. As you saw above, the storage bandwidth was much faster than this.
While I don’t have screen shots for all versions of ProRes, I found storage access to be fairly consistent across all the codecs.
As with storage, CPU usage varied. However, in these tests, it ranged around 75% of maximum capacity. Again, Premiere is not maxing out the system.
The good news is that Premiere is now accessing the GPU. (This is new with the 2019 version of Premiere.) The bad news is that it is not using all of it, yet.
While I’m sure that Adobe will improve this in future versions, the current version – and definitely all older versions – are not taking full advantage of whatever GPU is installed on your system.
WHAT DOES THIS MEAN?
What struck me in these results was how consistent these times were across all the ProRes 422 codecs. Three of them were separated by only 17 seconds, even though file sizes doubled!
NOTE: ProRes 422 Proxy has only 1/4 the resolution of any of the other ProRes codecs. It is designed as an aid to editing, not for creating a master file.
I’m not sure why ProRes 4444 was twice as fast as the ProRes 422 codecs. Whether that was due to the source media or a general efficiency in the 4444 codec, I don’t know.
NOTE: FCP X renders ProRes 4444 more slowly than ProRes 422, which is the opposite behavior to Premiere.
My takeaway is that there is no time-savings to exporting smaller versions of ProRes, unless you are really tight for storage space. Plus, the benefits of outputting a higher-quality file in the same amount of time will pay off every time you need to create another version from that master file.
My recommendations from the Thoughts on Output, above, still apply.
NEW & Updated!
Edit smarter with Larry’s latest training, all available in our store.