The Future of Editing...

dantewaters

dantewaters

Well-known member
Hey guys I was mulling over a few things and what do you think of this idea for not just FCP but non-linear editing.

Since we are constantly advances as far as cores, and gigs and gigs of RAM, and since
the system (windows or OSX) has it's native codecs (WMV, MOV)

Why not have these output files be updated thru the non linear system vs. rendering out a new file each time. For example:

If we have a shot of a guy kicking a soccer ball, and we want to change the music
all we should have to do is update the music within the non-linear editor and save the project file,
and this adjustment will only be made on the music end leaving the video information the same (hence saving rendering time, and processes), but beyond
this you will not have to re-render the entire movie over.

This would be similar to using a motion file in final cut where all you do is updated the motion file and the items will update in FCP without have to render out a whole new movie.

This is just my 2 cents and I know it will take a massive amount of CPU power, but it can be done I just know it.

How would you guys feel about this?
 
Or you just media manage your project and keep it on a hard drive you can get to whenever you need it. Thing is- keeping a file as a container for a bunch of different media- i.e. a reference movie can cause major trouble when you move that file say to another system or online. I prefer to have everything completely mixed, finaled and in one file with no dependencies. Otherwise it's not really a 'final cut' if someone else can come along and mess with it.

Noah
 
No you are missing the point...

If you do a 45 second clip for a client in 4k, and it took your macbook pro / Sony viao 4hrs to render out.
So you show him he likes the flow of it, and then tells you that you didn't follow the color scheme so you
have to make adjusts to the file and send it back to him (lets say he's watching this online).

That process requires you to re-render the timeline (another 4hrs) Lets just say there is
no re-rendering to the footage, but it just takes the laptop 4hrs or just a long while to export to the format you require.

Rather than re-render you change the colors and choose save...
the entire process happens (updated colors, and whatever you want changed directly to the movie file WMV / MOV).
You send the file in about 5secs and he approves, sends you the other 1/2 of your payment.
 
I don't think online viewing is the best example for this. 4k preview online seems a little much. Plus if you change the colors how would it not have to re-render the video from top to bottom. The audio I could see but video not so much.

With that said I do get what your trying to say (from the original post) it would be nice to not have to re-render. It seems like one of those questions that sounds super easy and the technology is there but in the end not so possible. Maybe in the future.
 
dantewaters said:
Rather than re-render you change the colors and choose save...
the entire process happens (updated colors, and whatever you want changed directly to the movie file WMV / MOV).
You send the file in about 5secs and he approves, sends you the other 1/2 of your payment.
Click to expand...

Uh?... when the video stream is 'rendered' it is done with specific intensity and 'color', and a lot of processing done to encode in the exported form... so when you 'just change' the color of a clip, or series of clips, or the whole video stream, the 'render' re encodes those changes. In a word, it is not like changing an initial 'look up table' to move all 'blues' to 'greens'...

Many editors do incremental 'renders', either automatically, or per command. However, if the material is maintained by the editor in one format, say DV, but you have requested a H.264 based output... a 're-render' and compression operation must be done if you 'just change the color'...

If it really is 'just change the music', the use QT Pro or something like that to eliminate the previous 'audio' track of file, and replace it with the new track. But anytime the 'video' appearence changes, one must 'render' the video again at some point...
 
I do understand what you're saying clark
However With the right algorithms say a specific system that handle color, one that handles sound I feel this is achievable...

Now Noah makes a great point... Having one file instead of many...
But my thing is having the ability to update your project (the 4k was just an example) on the fly right down to the final export whether i be:

mpeg 2, H.264, 8Bit , 10bit etc...
Save updated (done).

I know it's far fetch but I'm sure either the computers will render at light speeds
or someone will figure out a way to update just parts of an entire exported file, rather than having to re-export.

The other option is for flash or a media player to be able to break up an entire moving clip into parts so if you do an update you can just render out that section, and the media player will just read the updated file along with the rest of the queue.
 
In some NLEs this is already the case. For example in FCP I can put color correction on as a real-time effect. The actual processing is done by the gpu in real time. So I think we're kind of there already in terms of functionality. In terms of sending a file in some sort of "unrendered" state, that's sort of interesting but I still don't see how it would work without sending all of the raw media to the client first so their system could recreate the realtime effects.

Noah
 
You must log in or register to reply here.
Back
Top