What WOULD make a "jaw-dropping" NLE?

David Jimerson

David Jimerson

Administrator
Staff member
Let's put it on the table and design one. Let's talk about what needs to be done to get something from raw footage to finished product and come up with the best way we can to get it done.

Ground rules: this is not about any existing brand. This is not about any existing NLE, or what could be done to improve it. This is ABSOLUTELY not about X NLE vs. Y NLE. This is about designing something FROM SCRATCH to get the job done efficiently and effectively, whatever that might be.

So, let go of all preconceived notions. Let go of how things have been done. Think about the task and the best way to do it. Start with raw footage and build from there. Heck, start with the SCRIPT and integrate it if you can. Blur the lines between preproduction, production, and post if you want. Let all the ideas fly.
 
first, is there a reason for that? you think something is missing in the current crop of nle's

problem definition,

what's in the scope of NLE? what functions should it do?

data management?yes, more management controls :no, NLEs are not supposed to be database like
video editing
audio editing
export
after effects plugins
color correction?yes:why not use something like resolve

if there is a reason or need for it, maybe you can incorporate the use of current tablet touch screens or a stylus, or include a hardware similar to a jog/shuttle wheel in the price of the NLE or something.

usually specialized application do the job better than the collection all in one apps.
 
didn't notice the last paragraph, you want pre production, poduction and post produciton. the all in one corporate mindset is just a series of restrictions on the applications.
 
The reason is curiosity and a mental exercise, a brainstorming session. This is NOT about current NLEs. I thought I made that clear.

I didn't say I "wanted" anything. I simply suggested that no one limit their thought processes unnecessarily.

Look, if you think there's no purpose for this, then don't participate. I'm going to moderate this discussion pretty aggressively to keep it on the course I defined. Negativity is unwelcome.
 
No sure if this is what you're looking for, but:

How 101 do we have to get? I can ASSUME we can go ahead and state that cutting, sliding, slipping, rolling, penning, and the like are already incorporated. If not, I want those.

A viewer and a canvas.
Nice organization of files.
Audio syncing that works.
Compositing in the same program (no extra program for that). You could just change your view to "compositing" for this to get to one window to work in with all of your tools on the side. And this compositor needs to be powerful. I want it to be able to playback at real time as well.
Basic 3D modeling. I say basic because I don't need something so complex as to make a human head with individual hairs, but I would like to be able to work in 3D with basic shapes that can be manipulated.
Strong audio editing. This could be done the same way the compositor window is done as well. (View, Arrange, Audio Edits)
No codec issues.
Multicam editing that works by selecting the files you want used, and the program configures how to line it all up by aligning the audio.
No mixed framerate issues.
Compatibility to send my project to another editor working on the same program 500+ miles away without any loss or degradation.
Motion tracking in 3D.
 
Hawk Teflon said:
How 101 do we have to get? I can ASSUME we can go ahead and state that cutting, sliding, slipping, rolling, penning, and the like are already incorporated. If not, I want those.
Click to expand...

I don't want to make any assumptions, or automatically include or exclude anything. Those are all things which make perfect sense if you go a particular direction with it (e.g., track-based), but perhaps that direction isn't necessarily the best one. Or perhaps it is. I really do want to start from scratch.

In other words, I'm not assuming that this hypothetical NLE would resemble ANY NLE which currently exists. I don't want to start there, because it may limit the flow of ideas.
 
No rendering ever.
Real time playback of hundreds of layers with a ton of effects (guasian blurs, particle effects, etc.).
Real time playback of 8K video.
Use any codec without transcoding.
Instant (100x faster than real time) export regardless of the export settings or length of the timeline.
Iris Tracking; move a clip by simply looking at the timeline.
 
A few of those things would be greatly dependent on hardware; was thinking more along the lines of functionality and workflow. There are, of course, physical things which just can't be gotten around. But, surely, "as close to as possible" should be in the mix.

Don't know if you were serious about "iris tracking," but surely the control interface is part of the equation. Right now, it's all pretty much keyboard, mouse, shuttle controls, etc. Does it have to be?
 
Anyway, thoughts on total integration -- being able to keep footage linked to the script from the beginning would be pretty cool; perhaps even some integration with the continuity script would be possible. Imagine being able to feed in a script, which then has reference links to specific footage, even footage marked as circle takes, and the basic edit is done for you. Something like that.

Then, of course, there's the interface -- how you work with the footage, the sound, etc.
 
There's a plug-in for Avid that recognizes individual lines in the footage, based on the script, as well as variants of the line, and categorizes them accordingly. Almost enough to make me want to switch to Avid.
 
Other directions -- as referenced above -- is the entire "track" model necessary? Is there a way to work with clips without using tracks, or at least not rigidly adhering to them? Is there a way to do it where things like rolling edits, slip edits, etc., wouldn't even be necessary?

Think about tools, like color-correction -- how can that be done differently than the drop/window/slider/color wheel, etc., way it's done now? Audio? Same question. Is there a way to do it better? Directly manipulating the clips themselves in some way? Like, for example, being able to "touch" someone's face with your fingers and manipulate the skin tones as if you're actually painting? Just an example.
 
Also, following up on milksac's post (or my response to it) -- what's the balance between performance and accessibility? Certainly you could have a semi-trailer full of hardware and achieve a lot of those performance goals, but is that really something which is desirable? Or is it? Does portability or the ability to run it on the hardware we already have take precedence? Again, no preconceptions. I'd personally think it ought to be something accessible to us all.
 
Hawk Teflon said:
My mind can't wrap around a non track method. I'd be eager to see something else, but I can't imagine it.
Click to expand...

That's because the track method is the most visually representative way of how we process the idea of time.

As for my idea, I want an NLE that has an assembly language graphics engine that allows cutting H.264 and its variants (AVCHD, HDV, XDCam EX, ect.) like DV footage. I've been working with DV again for a project and, man, if all the footage was this fast, that'd really be something.
 
The track method could be the best way -- I just don't want to assume it is. Let's try to come up with something different before concluding anything.
 
David Jimerson said:
being able to keep footage linked to the script from the beginning would be pretty cool; perhaps even some integration with the continuity script would be possible. Imagine being able to feed in a script, which then has reference links to specific footage, even footage marked as circle takes, and the basic edit is done for you. Something like that.
Click to expand...

That would be cool.
 
This is an interesting exercise. I cannot help but believe that the conversation is akin to the ones they had as they developed their NLE at Sonic Foundry.
 
There's always something I've wanted coming from an audio background, and that's to be able to enter a tempo (bpm) and snap to a grid of 1/8, 1/16th notes etc. Just like in a daw, but you have the option to either lock to smpte timecode, or keep events positioned at their tempo markers moving with any change in tempo. It could be cool. For me it would make a world of a difference... dunno if that's jumping too far ahead in this thread, we're still at ingesting footage? I like this thread. It's nice to constructively dream, it helps you think about how much of 'your way' is actually your way of doing things. I had a similar session with thinking of a camera to no material end, but the stuff we came up with is nuts.... nice thread anyhoo. My kingdom for a tempo based option to edit with.
 
You must log in or register to reply here.
Back
Top