Hi!

For my own, I’m a user of Magix Video Deluxe 2006 Plus.

I use Magix for video editing almost since it exists.
It is a very friendly software with professionnal features at a very affordable cost: about 50 Euros in the standard version with 16 tracks supporting all kind of medias (image and sound) or effects (SFX, Transitions, Titles), and about 90 Euros for the Plus version (32 tracks and more features).

This said, I use Blender video editor for some effects that are not available in Magix, like the Glow effect. I also use it to mix scenes and video, which is only possible inside Blender.

As I like a lot the Blender Glow effect, I’d want to have more amplitude available in the setup of this effect, to get stronger halos from low lights or small sources.

Having a video editor inside Blender is a good thing, but as there are efficient video editors at low cost, I wonder if spending too much time on the NLE part of Blender isn’t a waste of time.

It is only my humble opinion, but I think that the energy of the developpers would be better employed if concentrated on the
main purpose of Blender : Modelling and animation.

Users of other 3d softwares will probably not switch to blender because of it’s NLE, but because it is a very handy, efficient and complete 3D package.

Philippe.

@@roubal&dave62:
“…there are efficient video editors at low cost, I wonder if spending too much time on the NLE part of Blender isn’t a waste of time.”
“There are a lot cheap (or free vdub) video-editor solutions. Why should i use the kind of unhandy nle editor for compositing.”
This may be true if you are on mac or windows, but on the linux side there are very few stable NLE’s. Blender has become the only usable video editor on “odd” hardware platforms such as PPC.

I really like vegas too! I don’t I have time to explain every details right now but let’s say that the main advantages are:
- Can do lots of stuff in just one click (really fast workflow)
- It’s a very intuitive program. You just need 30min to learn how to use it!

One thing that is really hard to use in blender sequencer right now is alpha between clips. You wasting lot of time to select clips in the right order, select the good alpha effect and redo the whole thing if you didn’t choose the right alpha FX.
A alpha propretie in the Srip Propreties can improve the workflow a lot!

a small overview about the whole film would be great. i mean, like the navigator in adobe premiere. so you shouldn’t move all the time from beginning to the end, instead, you had an small overview over the whole sequences.
it would be fin, if that navigator could be in such a small floating window, like the curve-widget or the transform-properties (shortcut N in the 3d-view)

Yes,the easy way to import .tga or .png files and they can see lower layer not to add alpha effect.

Easy to cut /move frames ,option of export file.

And How can I import .tga file to edit in Gimp. Gimp change my image. Alpha part change to black and color part change to alpha.

I think we really need for it to be able to open audio files besides .wav (e.g. Mp3, ogg)?

One thing and one thing only.

Sound. Allowing to control fade in and fade out of sound. Right now all you can do is cut down the track and reduce total sound level. And fix the mentioned things in wiki page. (chosing one sound system and perfecting it ;))

I believe that integrate Blender using Verse Protocol with http://www.virtualdub.org or http://www.jahshaka.org or any another free tool is better than create a complete editor sequence inside of the Blender.

Use the sequencer as an interface to create nodes in the compositor. When you add a movie to a strip, what is accually happening is you are adding an image node and setting it to movie then setting its frames to the length of the movie and finally setting its output to a mix node to allow you to add it to the current composition. If you want to add an effect, you can add the specific effect node between the image node and the mix node. You get the idea.

i do some free lance video editing and would be realy interested in a stable linux video editor so i think it is great that blender offers a nle. i also think that when there is a good nle blender can be a good allround motion graphics production solution when there is a good plugin structure for the nodes (to add effects).

features i would like.

*preview of first frame (the firs frame shown in the clip bar on the timeline)
*audio wave forms in timeline (use them a lot when editing conversations an surounding sounds)
*easy linking and unlinking of sound to a clip on the timeline.
*crosfading (when you have two clips do fast scalable crosfade between them ==clip1====clip2==) like in final cut pro
*linking clips to effects for the overview (the effects are a part of the clip bar on the timeline)
*effects keyframing
*scale rotation location effects (keyframable)
*doing a render to disk of imported blender scenes in the timeline.

*assign [spacebar] to play [c] to cut [f] to crosfade [e] to effect

*make a video editing panel so you dont have to use the same tools you use in the 3d space for nle stuf (change render to export etc. to make it more accesable for users)

for those who want to disable previewing- I don’t know if this is what you want, but you can just turn of the sequencer image preview in visible windows. In the default blender sequencer layout, that window has no header, but if you add one to it, you’ll be able to change from image preview to sequencer strip display, or change it from a sequencer to something else.

* the distinction between HD and RAM for adding strips in the editor isn’t very clear, nor is t intuitive that you can load diff. formats with it (linux)- RAM seems to mean “the old way” while HD means “the ffmpeg way”
* audio ram clips preview the waveform , but audio HD clips do not
* relative paths works for HD soundclips but not for RAM; RAM soundclips have absolute path soundblocks, even though the strip itself uses a relative path- worse, it fails silently, you don’t realize untill you move your files.
* very easy to inadvertantly create one frame gaps between abutted clips- several remedies possible here-( a change in background color of a frame if it is “empty” vs “has content”- gaps would show up as darker lines, or , suggested by matt, when you push one strip into another and it displays red, it snaps next to it with no gap instead of jumping up to another track)
* inserting clips is in the middle of a full edit is cumbersome- there needs to be a way to “push” the strips after backwards, without zooming out, selecting, moving, zooming in.. etc.
* when zoomed out on a small enough clip it becomes very hard to select the whole clip for moving, instead of the end markers.
* personally I don’t like the visual noise from frame previewing in strips; if this is planned, please make it optional.

these are ideas/suggestions, not demands :) I’m quite happy with the recent developments in the sequence editor, and look forward to whatever the developers have up their sleeves.

Being a Linux user with a limited tool set available I see great use and potential for the sequencer. I currently use cinelerra for my video editing but the blender sequencer is light weight with an opengl interface which makes it usable on lower end machines. Cinelerra recommends ( or at least they did before they went opengl effects ) a dual processor AMD server with 4 GB of ram to run. That said blender sequencer would really change the way I did things.

I recently was working on a project where I had a video captured at 10 fps that I wanted to sync with an audio clip. The sequencer would not let me sync my clip to my audio. Cinelerra detected the length of my clip based on actual time of the clip where as the sequencer pulled it in based on the frame count. It would be nice to be able to adjust frame rates for individual clips. I hope that makes sense.

I, personally, really like the sequencer. It surprises me how much I use it. There are a few things that would make it significantly faster for me:

1) Automatic Alpha Over. You import a strip of background images and place them on a layer, then import a strip of foreground images with alpha transparencies and place them on a layer above the background images. Done. No more Add images, add images, convert to premult, right-click, shift-right-click, space, Alpha Over, (c, switch ab if neccessary)… I go through that a lot!

2) MP3 import. Yeah, I know it’s not as high quality as WAVs, but It’s what most of my stuff is saved as. This would save me the step of converting my MP3s to WAVs

3) This one is a bug. I don’t know if it belongs here, but I thought I would bring up because it costs me a lot of time. I’ll take a meta-strip (with two image strips and a filter) and use it with an image strip and another filter. When I preview or render, the meta-strip will occassionally strobe (random frames get dropped). If I go into the meta-strip and preview it, then go back out and preview or render, the strobe is gone. It would be nice to not do that anymore.

–> does *anyone* use the “reverse frames” feature?

Anyway, I think that the sequencer has come a long way since I first fiddled with it. look & feel & probably code are better now. The demo at siggraph blew me away, it’s really that good.

I think we should look at it as a good bases for something bigger, and I agree with everything said about programmers doing what they want to in the open source community. It’s good peter took this over.

As for me, I think more control will be good, more help from the tool (when I first tried using it I actually opened the code to understand what is up), many other plug ins , and a better SDK for them – last time I checked it was per-image plugins only.

I know this would require a lot of work, and I’ve mentioned it berfore… BUT, I still think blender would be better off if it ditched FFMPEG and migrated all video and audio IO to gstreamer. Here’s why:

-It’s multiplatform and acts as a nice abstraction layer to system specific APIs
-It’s fast, low latency
-It takes care of sync and frame accurate seeking issues
-It still offers access to FFMPEG codecs, as well as others
-It’s less code, smaller footprint and easier to manage
-It would unify the sequencer and game audio io
-It’s isn’t stuck in CVS limbo (it has release versions)

Second, I like that you can have a http based frame server. This is nice, especially if you need access to it from a remote machine. But not all programs can use it. Could you look into implementing a file node based setup like AviSynth3? (it’s still experimental, but is to be multiplatform)

@peter:

to prevent to be misunderstood: i meant something like this:

http://www.tutorialguide.net/images/adobe_premiere/0001/008.jpg

it’s a small map of the whole editing, with all the videos, transitions and sounds visible

i’m still kind of new to blender/3d, but started a 3d vlog a couple of months ago. (part 2 is finished, i’m just waiting for the music now.) i wanted to use the blender nle for finishing my animation, but just could not do it because:

the plug-ins/filters i needed either did not work or did not exist (those i found only gave some error message, after looking in forums i tried to use an older blender version that was supposed to work with those filters, but i could not get it to work) – so what would be needed is:

• more composite modes/basic filters (many are missing, i needed “darken” combined with gaussian blur)

• it would be great if the nle editor was more intuitive, simpler. i personally like final cut pro, but most nles today work similar: when you add a filter you just drop it on a particular shot/timeline instead of having an extra timeline for the filter/effect/title. it’s nicer to work with, easier to learn and just looks cleaner if you have lots of effects.

i am very impressed with blender and as a film maker i now plan to move more towards 3d/combine more traditional methods with 3d. while i found blender rather hard to learn at first, i soon realised i could trust the app (if something does not work it probably is not a bug but some detail you’ve missed) and because everything is very well documented.

still it took me quite a while to understand and get used to the blender key frame editor – this was the greatest source of frustration for me so far. and i did have some experience with keyframe animation in final cut pro. i mention this because:

• it would be great if the keyframe editor could be more intuitive so that you can easily use filters in the nle editor via keyframes (i’m not sure how/if this can be done at the moment)

another thing i needed for my last project was:

• slow-motion/speed manipulation for clips (again i don’t know if this can be done, i just did not find a way to do it)

so all in all maybe:

• a simpler, modern ui à la final cut pro (or avid etc.), all standard composite modes, standard set of filters (like gaussian blur, sharpen etc.) that work with the latest blender release and a simple, easy to use keyframe editor. and of course a good documentation.

to the above discussion about the usefulness of a nle in blender:

for mac os x there is no open-source nle editor that i could find that works on my system (i tried jahshaka, but i looks like it needs os x 10.4 and i only have 10.3.9 at the moment). i’d love to be able to produce/finish a complete animation just with blender – if this can be done without risking to make the app buggy…

@Reuben:

Switching from ffmpeg to gstreamer: don’t like the idea, see why:

“-It takes care of sync and frame accurate seeking issues”
doubt that, looking at the code. I had a lot of twiddling with ffmpeg and I’m still not finished doing timecode based seeks.
They simply seem to assume, that ffmpeg can seek by itself, which is not true on mpeg files for example. But maybe some gstreamer expert can enlight me.

“-It still offers access to FFMPEG codecs, as well as others”
some old version, not updated, when I checked last time.

Additional problems: maybe I misunderstood the documentation, but I was under the impression, that you have to hand over the event loop to gstreamer to make it work. I doubt, that you will convince Ton (or even me, and I’m a little bit crazy ;-) on this.

———-
To make a long story short: I will add a plugin interface for audio / video to blender. If you want to write an input / output plugin using this interface, that uses gstreamer, I will not step into your way. But I don’t think, that gstreamer is the final answer to everything…

“Second, I like that you can have a http based frame server. This is nice, especially if you need access to it from a remote machine. But not all programs can use it.”
Hmm. That is interesting. Just write a plugin for the other program?

“Could you look into implementing a file node based setup like AviSynth3? (it’s still experimental, but is to be multiplatform) ”
Don’t know, what the sense is here. Let’s take a look: within windows, you can overload the DLL-calls to AVI-opening. That is the way, virtualdub does the trick. Within Linux you can… wait there is no AVI-library to overload. So the best thing here is again to simply write a small “glue” daemon, that translates the calls.

—
Don’t want to bash here, but I don’t think, that it is that easy. If I simply have misunterstood your ideas, please feel free to correct me.

- crossfading. Since I do not own Final Cut Pro, please explain a little bit more.

for example:

==clip1=========| |=========clip2=====
end^ ^start

now when you rightclick on the point where the enpoint of clip on touche the begin point of clip2 you ge a menu with options and one of them is ad crosfade (or however they call it) when you select that option the picture looks like this:

==clip1========={> | effect

this is a clip with 2 effects:

___________
| clip
|___________
=========== >effect
=========== >effect2

this is a clip with a effect and a transition effect:

___________
| clip
|___________
=========== >effect

Hi im also some of those that like the sequence editor quite a lot, what i miss is more integration between nodes/sequencer, i mean if we have nodes why not use it with the sequencer and vice/versa. I dont know exactly the amount of trouble would be to mantain such code, but i think that it is a resource not being well used.

hy sorry for this unorthodox way of posting but when i just posted in reaction to Peter Schlaile i made some ascii explenation of features and blendernation now onely show’s half of my mesage so i copied my post it in a textfile and uploadet it to a fileshare lot off banners thing you can find here:
http://w10.easy-share.com/695529.html

sorry for this i think my ascii graphics are filtered out by blendernation becouse of the sql insert securety or something like that.

@gman:

if I understood you correctly this time: crossfading will somehow close the gap between to clips. May I ask, what video information should be used in between? Just the still begin or end frame?
My solution to this would be more like: just let people drag the two clips into each other. That will simply cross fade. If you want to do advanced fading, you will have to add a seperate effect track as before. Does this also makes you happy? (Is even easier then the final cut solution…)

Close gap doesn’t really work right now. Has to be fixed.

“linking clips to effects for the overview”: I’m not very convinced after the explaination. I add effects to more then one input track all the time. And I don’t want to get confused. You seem to only stack effects on one track all the time. Correct me if I’m wrong or show me, what you want to do any better in the case of 2 input tracks to an effect.

@simhar: When I said to focus in the future UI redesing I rather meant to have it in mind. Maybe this is a good opportunity to do it and focus in real editing functionality instead of things that could change in the next months and freeze aspects that could be work when the interface is reworked.

——

@Roubal:
I understand your point, but:
-That division is the standard in almost every single NLE in the market
-That structure has a purpose: Audio and video tracks are linked and have common track locations: When you drag an AV clip into track V1, its audio goes to A1, and so.
-I’m not sure about the synchro problems with that structure: if you have a muzzle flash in V1, you move the play marker to that point and insert the audio clip of the “bang!” there in A1. If you can see the waveform and the playback bar, you can synchronize everything.
Anyway, I do not hate the current layout (I think it’s more flexible), but I think the other structure is more effective in production work because you can have better track of the assets and their locations.
-The division between audio and video allows to have specific switches in a track header (for example: a gain slider in audio, an opacity slider in video)

——

@Peter Schlaile:
The crossfade system gman described works together with trimming in the sense that it uses the preroll and postroll material previous to the trimming of the clip.
If the clip has preroll or postroll information, then it’s used to the mix, if there is no extra roll, the first frame freezed is used (generally with a warning of insufficient media).
The mixing takes place in the same track/layer. You put both clips in contact (outpoint of the first by the inpoint of the second) and you drag the efect over the contact point.
The complete duration of the transition is over that part, in halves by default (but you can change the both sides duration independently).
Cinelerra uses almost the same method, you can check it there.

Another idea for a feature:
-timeline markers (allow to put flags to identify easily and faster important points in the timeline).

@simhar: When I said to focus in the future UI redesing I rather meant to have it in mind. Maybe this is a good opportunity to do it and focus in real editing functionality instead of things that could change in the next months and freeze aspects that could be work when the interface is reworked.

@Peter Schlaile:
The crossfade system gman described works together with trimming in the sense that it uses the preroll and postroll material previous to the trimming of the clip.
If the clip has preroll or postroll information, then it’s used to the mix, if there is no extra roll, the first frame freezed is used (generally with a warning of insufficient media).
The mixing takes place in the same track/layer. You put both clips in contact (outpoint of the first by the inpoint of the second) and you drag the efect over the contact point.
The complete duration of the transition is over that part, in halves by default (but you can change the both sides duration independently).
Cinelerra uses almost the same method, you can check it there.

@Roubal:
I understand your point, but:
-That division is the standard in almost every single NLE in the market
-That structure has a purpose: Audio and video tracks are linked and have common track locations: When you drag an AV clip into track V1, its audio goes to A1, and so.
-I’m not sure about the synchro problems with that structure: if you have a muzzle flash in V1, you move the play marker to that point and insert the audio clip of the “bang!” there in A1. If you can see the waveform and the playback bar, you can synchronize everything.
Anyway, I do not hate the current layout (I think it’s more flexible), but I think the other structure is more effective in production work because you can have better track of the assets and their locations.
-The division between audio and video allows to have specific switches in a track header (for example: a gain slider in audio, an opacity slider in video)

Gez is right about what i meant with the crosfades. this works realy fast especialy when you combine it with the close gap tool. it allows you to edit a un edited camera shot and just cut out all the crap close the gaps and tweak the cuts and add some crossfades where needed.

and i agrea with Gez, timeline markers would be great to! if you could snap clips to the markers. and when you assign a key to seting markers, you can “tap” the key to the beat of the music you’r editing on and then just snap the clips to the markers.

the crosfading can also be done in a other way, i have used it on a hardware based editing suit last year but i’me not shure what it was caled.
in that system the firs 2 tracks would always crosfade all the overlaping material, so you could easely drop all your clips on the first track and when you neadit a crosfade instead of a hardcut you would just drop that clip to the second track and overlap it and the overlap would crosfade.

this workes realy fast if you had to edit some material down from like 1h to 20 min. And you could do all the more complicated editing on the other track without having to fear tha something was overlaping..

it would also be good to have a inpoint outpoint meganism on the timeline. where you can watch a whole track and define inpoints and outpionts on the track while watching, and when you’r done you will only have the clips you wanted on the timeline and not the ones you dont want whitch would be still on the timeline if you where just cutting the clips.

@Peter Schlaile
“linking clips to effects for the overview”
i understand that it is a good feature to be ableto add a effect to more clips
but what i meant is that it is not always easy to see whitch effects are linked to whitch clips.

but maybe this would be no issue in the future when you can assign the effects with the nodes sytem :) if that is possible…

@ shawn fumo: thanks for the link, i’ll have a look at it. (but it looks like there is one important limitation: it is only dv. this is of course not really an option for 3d animation…)

@ gman: dissolves/fades, effects etc. in final cut pro are little icons you drag and drop either on a clip (for an effect) or over a cut (for a dissolve). when you doubleclick a clip or a dissolves you can access what would be a properties window in blender (in fcp the player/viewer window changes to this when you doubleclick a clip/dissolve in the timeline). here a general screenshot: http://en.wikipedia.org/wiki/Image:Fcp5_screen.jpg (http://en.wikipedia.org/wiki/Final_Cut_Pro)

@ peter schlaile: thanks for the detailed tips, i’ll try and see if i can get what i need this way. about the nle (and blender’s) interface in general: from what i have read (and again i am still kind of new to blender) i think there are discussions about how-to make blender more accessible for more people. final cut pro (that “borrowed” a lot of its interface from avid) and most other nles i’ve seen (exceptions might be very expensive high-end ones) all tend to have this drag and drop effects onto the timeline/clip solution. the way blender does it (extra strip/timeline for an effect) is more the way this was done 10 years ago. maybe i have not used it long enough, but to me this seems not very intuitive and tends to clutter the timeline (or might require an extra step to bundle it all in a new strip). i’m sure i can learn how-to work with it, but if blender wants to gain a larger userbase than simplifying the interface seems to me very important. i agree that the keyframe editor does work well once you figured it out, but not every creative person is also a technical one (many are just the opposite and get pretty scared when i show them blender). of course in the end ui discussions are also about taste and this is of course something very personal…

I also don’t like the way that the seq editor automatically rescales an input strip to the render size. It would be nice if there was an option under strip properties to import a strip at it’s origional size – it would certanly make the tranform effect much more effective

The ability to use realtime opengl effects! Then we could all make our own effects in the 3d view window and these effects would be realtime.
Some examples that wouldn’t need rendering …
Bluescreen like in the game engine video plugin (you could even use the code from it!)
alpha fade ( use multiple materials on a plane and change alpha value)
video on spinning 3d cube while another video playing in background
shatter effect
shrinking circle
blinds
rotations
other effects like in the free windows video editor “Wax” – which you can even import 3d objects into effects.

well if you’re adding realtime effects bases on elements from within the 3d view i would start with a good “stroke path with brush” function. this is realy usable in many motion graphics projects (think about apha masking combined with strokes resulting in “growing” elements etc.)

I use xjadeo + ardour for make 5.1 sound for video. It will be nice to connect ardour directly to blender. To do this you need to add jack support with slave mode.

screenshot of ardour+xjadeo: http://sourceforge.net/dbimage.php?id=46744
ardour’s home page: http://ardour.org/
jack’s home page: http://jackit.sf.net/

High end Avid systems have a feature that allow them to store video in a database like structure. I have done several 15/20 minute videos in the past. the most recent one had over 10 hours of footage and ~400 clips. Searching for those files is a pain. It would be great if it could be setup to be able to search for metadata in those clips. Sqlite+Verse?

Verse support would be nice. Multiple people working on one project.

@Peter Schlaile
”
“well if you’re adding realtime effects bases on elements from within the 3d view i would start with a good “stroke path with brush” function. this is realy usable in many motion graphics projects (think about apha masking combined with strokes resulting in “growing” elements etc.)”
Since this is rather obscure: Why not just use the 3D-editor for creating the animation? This is BLENDER!
”

these “Motion graphics” functions i desctibe are verry hard to make in the 3d window. you can do some stroking with the particles but this is limited and not realy intuitive. I dont think the nle is the right place to add sutch functionality tough.

Maybe a motion graphics screen layout with a bunch of speciffic motion graphics functions like stroking vector paths. Easily transforming vector paths with keyframes without having to use the shape keys (slows me down a lot!) and importing bitmap graphics (or even movie clips) with an alpha chanel right on the canvas without having to map pictures to a plain. ofcourse with a good previeuwing in the 3d space (also some previeuwing of the aplied effects to the ellemnts).

that would be fun :)

I know my previous posting doesn’t seem to have anything to do with the current purpose of this page, but on another website people mentioned that when you hit Alt+A nothing happens in the compositor and there was no IPO access, this would be one way to fixing this problem.

@Peter:

maybe the “preview pictures in the timeline” shouldn’t be for every frame, but one (maybe the first) picture on a movie-strip, to identify it easy…

I want, I mean, I need chroma key !!!

I would love to see some of Blenders weaker areas strengthened. Thats what this is about.
I have Premiere Pro with a 1000 real time editing card which I hope not to use… :) I want to use Blender its all there. I hate pulling a program from one area to another it gets so messy.
I think that Ton is working on the the image editor or is it the browser?

Are these compositing tasks or video???
I would love motion tracking for video help to intelligently help with all the things we want to with blender.
So I am asking for is a path in for motion tracking and the sequencer handles the video so that is where I think it should be. Then it could be exported for autoroto scoping in the image editor. Or stablization.

Also a way for things that you could do in the image editor(Painting (frame by frame)) to be put into a sequence automagicly. Is this a composting task? Its a little unclear. The image editor is lacking in the video dept.

@Peter Schlaile:

“Don’t want to bash here…”
You’re not bashing. You know a lot more of the complexities involved here, so it helps when you explain why some ideas won’t work.

“They simply seem to assume, that ffmpeg can seek by itself, which is not true on mpeg files for example. But maybe some gstreamer expert can enlight me.”

I don’t think they use ffmpeg for seeking or sync. I believe the Gnonlin plugin handles that type of thing. (Gnonlin is a gstreamer interface to provided functionality needed by video/audio editors)

“maybe I misunderstood the documentation, but I was under the impression, that you have to hand over the event loop to gstreamer to make it work.”

I’ve been told that’s not the case: “It runs it’s own threads. The only mainloop integration is via the GstBus, to receive messages from the pipeline, but that doesn’t have to use a mainloop either”

@Gez
I understand that you don’t want blender to be a “everything in one package” that seems logical to me. You dont want to be botherd with functionality that is not interesting to you.

I (and i gues many others to) use 3d is very intergrated with motion graphics and video editing.
I am not saying that these are the things every 3d program should be able to do, but in my oppinion the best 3d sofware would be a package whitch allows it users to do everything they want to create the best endresults. And blender is in manny way’s the best solution and one of the main reasons is it has a lot of the things i need in my motion graphics pipeline.
So why keep these parts of the program basic? they are very valluable for a lot of users and if there are people interested in developing these functions why not?

i get your poin that is has no use to include spreadsheets, but i just dont get what is wrong with pro like video editing functionality?

(i hope i am not boring you :) but i realy think video editing is a essential part of many 3d pipelines and good dv video editing (on material similar to wedding video material) is a great feature in a 3d program so you can easaly do special effects and motiongraphics in the same environment where you do the editing.)

@Gez: Hmm, to be honest, I haven’t used the Sequencer ever for serious 3D-work. In fact, I needed a stable NLE to edit our companies celebration video, which was 3 camera live footage 2 hours long… ;-)

Since I haven’t managed to get anything usefull out of the other OSS solutions (Cinelerra, Kino, etc.), I tried the Blender Sequencer and was surprised, that it reacted the most like I expected it.

It is rather my relatively conservative view on what a NLE should do and what it shouldn’t do, that made my code usefull for others.

The nice thing about Blender is, that it “thinks” in an orthogonal way. Other programs try to build single purpose plugins (most probably to make advertisements, that they included another zillion effects in the next version, that I have to buy), instead of making a basic tool set, that can be used in a general way.

Why should I try to include _any_ 3D-transition effect into the sequencer? The single purpose effects are basically useless, and if I need something special I’m doomed anyways with a commercial NLE. In Blender I simply add a scene, create my effect and throw it onto the timeline.

Why should I try to add _any_ compositing functionality except the most basic one? There is a compositor, only one screen away included, ready to help.

In that way, I think by simply focusing on making a very straight lean and mean NLE, that integrates well with the rest of Blender. we could satisfy both worlds, the 3D-artists _and_ the wedding video guy. (Assumed, that the later had time to learn the interface…)

In fact, I wanted to add something soon that especially should make our live video freaks very happy: The (optional) ability to seek based on timecode instead of frame number. If you have several cams and want to synchronize later it is a real pain if someone decided to stop his camcorder in the middle of the show several times. The funny thing about that is the fact, that it makes also mpeg-seeking in certain cases more stable (there are mpeg files out there, that have variable frame rates… yiieks!).

I still think, that we can make both worlds happy, if we don’t try to add features “just for the sake of it” and try to stick to a minimal feature set, that works very robust and gracefully.

To make it short: we both have different reasons for our opinions but most probably will arrive at a very similar destination… ;-)

P.S.: I simply can’t believe, that you would try to edit a wedding video with Cinelerra… ;-) I will go nuts, if I have to use it again…

@mppic: sorry, the easiest way currently is indeed using the IPO in frame locked mode, using two control points (one at start, one at end) and adjusting the second one to your strip length. Reason for this: the speed controller doesn’t really know, in which way it’s input changes length (it can’t know the original length of the input strip!). If this task is very common, one could add a hack to accomplish it though, but it would be a hack, really!

What do others think? Is this very common?

@mppic: thought about it again. You are right. Is in CVS now. (Works only for enlarging right now and only for the right handle, not the left one. Will add that later.)

@Reuben: “I don’t think they use ffmpeg for seeking or sync. I believe the Gnonlin plugin handles that type of thing. (Gnonlin is a gstreamer interface to provided functionality needed by video/audio editors)”
I read the Gnonlin sources. The seeking is handed over to the input plugin by passing up a seek message through the tree. The input plugin then performs the seek operation. This is rather naivly implemented and will certainly not work with mpeg files. Looking at my latest version of the seek code maybe it is possible to work around this. Have to think about it…

““maybe I misunderstood the documentation, but I was under the impression, that you have to hand over the event loop to gstreamer to make it work.”
I’ve been told that’s not the case: “It runs it’s own threads. The only mainloop integration is via the GstBus, to receive messages from the pipeline, but that doesn’t have to use a mainloop either” ”

That sounds interesting. Is there example code out there, that shows that?

@tom: The binary size of gstreamer is essentially the interface library which is very small. (around 30 kb or so). The problem I had with it was more that I was not sure, if it could be used as a serious ffmpeg replacement because of seeking issues and not up to date ffmpeg version.
And it should be multiplatform by advertisement. ;-) One has to test, that’s sure…

Wow, that’s amazing! Now the speed control behaves exactly the way I imagined – thanks a bunch, you just made my day!
Another tiny idea just came to my mind: Would it be hard to integrate frame blending in the transform effect? This would help me to keep a consistent look while I’m zooming/panning/rotating in storyboard stills inbetween speed controlled image sequences. It’s not a problem now and certainly not a request, would just be nice and consistent ;-)

I just want to mention another reason why I think asset management is important even for purely 3d work. Creating just what you need and assembling it in the timeline may be ok for smaller projects and meticulously planned bigger ones, but there can still be problems. For instance, it seems like a lot of time was lost in Elephant’s Dream because things were planned a certain way in the storyboard but had to be changed at the end of the process. I think the individual scenes were roughed out and the completed, but when putting the scenes together, pacing issues came out, etc.

So, perhaps a workflow could be done more similar to filming a live-action movie. When you’re still in the animatic stage with very basic models and animation, you could create longer takes and even multiple takes of the same action exploring different ideas. At this point, you can depend on the asset management and the strength of the editing tools to create a relatively solid edit of the whole movie. Then the pro voice acting can be done. Now you can go back and flesh the portions that were actually used.

So, in this case the power of the editing suite would be more important because it would be tightly integrated into the process of creation instead of just assembling finished pieces of animation at the very end. In this kind of situation, interesting integration options could possibly be done. Like if you select a clip in the assets or timeline and do a special option, it could open up the scene in blender, marking the ranges of frames that are actually being used in the final edit.

So, I think a powerful editing section (including assets) is actually more important for a project like Elephant’s Dream than something like a live-action film that happens to have one 3d scenen in it. In the latter case, it may very well be better to use a dedicated video app which is more likely to have video-related features like de-interlacing, capturing directly from a DV source, etc. (though the lack of a good open source editing program on windows is still pretty glaring). But for the former case, it could potentially save a lot of time by having a version of the final edit throughout the whole project to make sure that a final render only happens for what is absolutely needed.

@Peter
you are right, of course, there really isn’t anything to blend as the transform tool already calculates all needed inbetween frames. Which is my ‘problem’ exactly.

Maybe I wasn’t too clear about what I’d like to achieve. I have two strips that I want to have a similar look: First I stretch an image sequence via speed control, frame blending turned on – everything fine so far. The second strip is a single image which I zoom into via transform, which (correctly) gives a smooth animation. To make this strip look like the previous (with blending), I’d have to ‘drop’ some of the frames transform calculated and replace them with blends, which isn’t possible right now (correct me if I’m wrong).

I realize now that integrating frame blending in the transorm tool is more complicated than I originally thought because you’d also need some sort of ‘strobing’ or ‘step size’ that tells transform *not* to create all inbetween frames.

A more general approach would be to turn strobe and frame blending into independent effect strips. You could then take transform output, drop frames as needed with strobe and after that blend the remaining frames.

Quite flexible, but even more complicated to implement – and maybe there isn’t any other ‘need’ or use for this except my case, so I guess we might as well forget about my ideas ;-)

It will be nice to have background rendering linke in cinelerra. Just store rendered frames in jpegs and play it in real time without caching.

ive been using blender for a year now i am still not that good at it but i know how to use it i go to school for multimedia and i have found that some of blendr features in the video i can compare to final cut pro so why not make them alike. what i am saying is use final cut pro as a template to fix blender if you use it you will see what i am talking about. final cut is onely on mac we have lots of video programs for pc, we should make the video editor like final cut.