Is Premiere Pro due for a full code rewrite?

[ah_photo]

I know this is primarily a support forum, but here’s a bit of musing for fun. Hopefully this type of discussion on these forums is allowed, if not I apologize.

"Is Premiere Pro due for a full code rewrite?”

By no means am I saying this is an easy task (on the contrary it’s probably the hardest thing for any software company to do), and by no means am I stomping my foot and demanding it. That said, I do think it’s a fair question to ask.

Unless I’m mistaken, Premiere Pro’s last full code rewrite was 2003. A lot has happened in 17 years, and while Adobe has done a good job of expanding upon their foundation, there are some artifacts that I feel are the result of hanging onto the dated foundation.

1. Rec 709 and the changing video landscape
   In 2003 Premiere Pro (and all NLEs I guess) were primarily used for broadcast and cinema. YouTube wasn’t even created until 2005. The landscape of digital video production has changed drastically since. Not everyone using Premiere Pro is using a Broadcast Rec 709 workflow. The fact that people frequently cite issues with “desaturated exports” is noteworthy. While Premiere is doing nothing wrong and is fully working as intended, this identifies a clear issue that editors may not be given the opportunity to define the intent of their workflow, or be able to convert between color spaces for footage and/or export.
   
   
2. Presets, configurations, and the editor's perogative to personalize.
   The way presets and configurations are managed seem very much like a relic of 2003. There is no clear manager for presets, and there are inconsistencies between how things are modified. Effects Presets, Workspace Layouts, Track Height Presets, Sequence Settings, are all managed differently and in different locations. The local storage of these settings is baffling as well. The directory differs depending on if you have synced to CC or if you haven’t synced to CC, and sometimes the wrong directory is referenced when upgrading (My 2019 -> 2020 update referenced my outdated non-CC directory as the foundation for my 2020 settings). A unified in-app manager tied to a single directory on your machine would solve a lot of this. This is critical because I'm sure many editors, like myself, like to define our work environment. We setup our panels, our keyboard shortcuts, and our presets so we can focus on the task at hand of creating stories rather than spend time fussing with the software.
   
   
3. Workspaces to accommodate newer Premiere features
   The customizable workspaces in Premiere have been a highlight feature for me, but seem to have suffered some drawbacks over time. For example, project panels consistently do not load in the correct location when working across projects, assumedly a result of Adobe adding the ability to open more than one project at once. There was also the dreaded “workspace reset” issue that was ultimately addressed by preventing default workspaces from being deleted or modified at all, which was a bit of a band-aid than a fix. A new system that helps workspaces work consistently with newer features in mind would be top notch.
   
   
4. Performance
   This is a bit more vague, but one of the more common complaints about Premiere is performance and stability. Adobe has taken a tougher stance on this and personally I have noticed gains in this space in 2019 and 2020. Still, would a code rewrite allow for this to be even better when compared to building off of a foundation from 2003? I genuinely don't know, but again: food for thought.

I have been using Premiere Pro since 2015 and to date I still think it is one of the best NLE's around as far as overall editing functionality and interfacing goes. I enjoy using it to this day, especially alongside other Adobe apps that help with workflow. None of the above ultimately ruin the experience for me, but in an era where competing NLEs like Resolve and FCPX are vying for editor interest by leveraging foundations that are nearly a decade newer than Premiere, the question still remains: are we due for a recode? I would imagine Adobe has learned a lot in the past 17 years and could incorporate a lot of that learning into a new NLE framework built for the modern era.

Curious to hear others' thoughts, especially considering I am severely underqualified when it comes to software development. Cheers!

[Ann Bens]

No problem this type of disscussion but moved to video lounge.

[ah_photo]

Great thanks! Didn't know this was even a thing haha.

[R Neil Haugen]

Patrick Palmer was brought in as head of all video apps around two years back. Prior to his term, Premiere seems to have been headed for simply "Cool New Features!" all the time. He stopped that, as notably, the number of actual new major features plummeted over the last year and a half. And he brought in Ivo Manalov (hope I spelled it right) to oversee a complete rebuild of the development structure.

Patrick, Ivo, and several engineers were at Adobe MAX in LA last fall, and spent a couple hours talking with the ACPs from the Premiere forums that were there. Was quite blunt and fascinating. We ACPs tend to have our own bones to pick with the program from many different directions, and yes ... we routinely see the troubles other users are having. And we're not a shy lot by any means. Ahem.

First they noted Ivo was tasked with rebuilding their internal structure and every flipping process their developers go through so that stability, performance, and bug considerations are top of the list at the begginning and all the way through every step of the process. From discussions, I would guess that there were a few separate people handling those issues, that most developers weren't tasked to do so. That's gone. And charmingly, the engineers present clearly liked their new structure and processes a lot better than the old.

Second, your comments about old code are spot on, but ... it's worse. Yea, some of the code is downright ancient, but it's been updated in bits & pieces so it was a complete mess. Workable, surprisingly, but ... far less than optimal.

So ... in order to get stability/performance constantly improving while killing bugs, they went through and simply junked all old devices they connected to and some old formats/codecs and tasks. Not worth the time to update. Next, they started rebuilding what was left especially certain heavily used things like the "pro" media internal relations with RED, Arri and other such media to get it working in at least stable if not "performant" operation.

Oh ... and changed from only a private beta program to adding a public beta option for users also. One of the biggest problems they had with performance/stability/bugs was simply as this app runs on a wider array of hardware and media than any other NLE had been, even with I don't know, probably several hundred Macs and PCs internally and whatever they had in their private beta system, they couldn't replicate most of the surviving bugs and performance issues.

What the engineers can't replicate, they can't fix. So the "public beta" process was added to get a wider array of tested input.

They've started to roll out some major changes from rebuilding the app recently. Improvements in things like using the GPU for hardware decoding/encoding of H.264/HEVC, continuing RED issues, Canon file support, quite a few other things that are getting tested across a wider user-set.

Understand, I've still got the list of My 20 Biggest Gripes with Premiere, followed by the other list of My Next 40 Issues with Premiere. But me being me, I have that with AfterEffects, Resolve, Windows ... yea, everything I use.

Every editor I've ever watched does things differently than every other editor I've ever watched. One of the fascinating thing about video post. And of course, a bit of a pain if you're a developer for that divergent and always opinionated crowd ... ha!

Neil

[ah_photo]

R_Neil, it's a privilege to hear your insight, as always.

I was not aware of the new head of video apps 2 years ago, maybe that explains in part why the last 2 years Premiere seems to have had positive momentium. I do think the open beta approach has allowed me to upgrade to newer builds with more confidence than before, and 2019 and 2020 have been quite stable for me compared to previous years and I attribute that in part to having that feedback loop.

I suppose I should be grateful for that at the very least. If I'm understading you correctly though, it seems like while there hasn't been a full, ground-up, re-write, it seems like Adobe's approach to updating has changed in the past 2 years? Perhaps one that isn't afraid to break a few traditions and legacy functionality in the name of improved performance and stability?

"Every editor I've ever watched does things differently than every other editor I've ever watched. One of the fascinating thing about video post."

Agreed. Hence why I wanted to bring up the #2 point in my post. I think that editor perogative should extend to the feeling of being able to export a full config package, import it on a different machine, and essentially take your Premiere setup in the same way you do a briefcase to a new office desk.

[R Neil Haugen]

A good share of the complaints here on the forums for the last year+ have been from users with older gear and/or older little-used video formats. Why? Because writing code for every freaking out-of-date card/chip/format out there is a massive job and one that requires creating "translation" code to get from newer code to ancient unused code.

There of course are the comments "but they don't have to rebuild that code, why is it a problem?"

Well ... the problem is in the first paragraph. They've been systematically prepping for then starting to go through the app for code rebuild, and yes, they would have to spend time and testing to create new instructions ... and put them in every function of the app ... to translate new code into old to run some bit of gear or format if X hardware/format is involved.

So, yea, they've been dumping older stuff all over the app.

The Titler is an example ... it's one of the oldest bits of code in Premiere. By the time they rolled out the EGP, tThere was no way to 'update' the old Titler code other than completely rebuilding it.

So they built a new graphics/titles "machine". The Olde Tyme Titler is still available, but I sure wouldn't use it for anything. I'd guess the metrics are showing very low usage over the user-base, and when it gets to some X figure, I expect it will simply disappear. Nearly everything you could do in the old thing you can do in the new, and you can do things in the new you can't in the old.

Yea, I want more shapes to simply select from, who doesn't? But in general, the EGP/mogrts works.

I'm rather ... charmed ... by how fast my H.264 encodes are going now with the new use of the GPU for that. Even my "old" GTX1060/6-Gb card does a pretty good job of speeding up exports to H.264. That's been a great help. It's the second item off my "top 20", the first they adopted being changing ALL the keyboard shortcuts that used to work only with the olde Titler to just being a "graphics" key-short, so they work with either the Titler or the EGP/mogrts work as well.

Now ... if they'd start with the other 18 things still on my Top Twenty list ... 😉

Neil