Alan, your idea is very good, but for me it's look like a dream ...

Look at the time needed just for have kdenlive .. so try to imagine the time for a complete integration suite ..

so you can begin your help in debugging .. you will see the time needed for a good soft ...

nyme

Nyme you are completely correct that this is a dream. You are also correct that it would take an incalculable amount of man hours to complete, but let me make a counter argument.

In 1963 M.L.K also had a dream about racial equality. Last year America saw the election of the first black President.

On September 12 1962 John F. Kennedy made a speech about sending a man to the moon. He said “we choose to go to the moon and do the other things, not because they are easy but because they are hard”. We made it to the moon a mere seven years later.

And more recently in the 1982 when Richard Stallman discussed his dream of an open computer environment I'm sure people told him to think of the time it would take to develop. To think of the work and the man hours and to know that in the end it would only equal what was already available. So why try? But the GNU project was created and is nowadays a vital operating system.

Yes this project would be hard. Yes it would take an unimaginable amount of time, but it needs to start somewhere at sometime. If we are to just say “it will take too long” we would be shooting ourselves in the foot and not living up to the true focus of the GNU project.

The formation of a FOSS video production suite is something that will happen. The community is to motivated and intelligent to not let it. Whether it starts now or ten years from now I can't say, but it will happen one day. I think we should get a jump start on it now.

So even if it is a large dream I want to know the first steps. How do other projects get formed?

From a programing aspect color is currently integrated through the FreiOr, there site is down right now so i can confirm color depth, but i think it is above 8bits per color channel (RGBA). so that should make it possible to have smooth color controls. as long as they user interfaced correctly.

AlanJames1987 - its not a dream, its merely the next step. yes kdenlive needs work, and so do many of the other packages, but integration need to be thought of. I'm no great programmer but user interface is important and often lacks in Linux, but more important is integration on a deeper level. we need to be able to take video/audio/frames into any other package easily. This needs to be defined and clearly. there a number of projects that do this in different ways, for instance http://en.wikipedia.org/wiki/Verse_protocol. the up side of a project like verse is that its network based meaning that the colorist and sound engineer could be working on it at the same time and if the director changed a cut the would have instant access to the change. someone working on the DVD authoring would not have to wait till every one else is finished to start as he would have active access as well. especially if footage could be marked in a unselected/selected/stared/incut manor. If anything i would like to see more incorporation of this type of technology.

uni-verse (http://www.uni-verse.org/) is unfortunately restricted to 3d at present but if that could Incorporated a fuller work flow i think it would be perfect, especially if versions of renders could be maintained.

MillionMonkeys - EDL looks quit possible, form the structure of Kdenlive. though I'm looking at the structure of an edl at the moment cus i have no idea from a technical point of view.

ultimately, if we being serious we need to look at 16bit per channel, 10-bits not bad, but there are those who will need 16bit as professionals. 12bit is great on a 2k image (http://www.max-t.com/press/cmn/2009/14apr2009.html) and the machines do that already but if you goes as high as 4k, you going to need a minimum of 16bit per channel. that comes down to a desition if they seriously want to include pros or not.

4K is essential, red cameras are using it to shoot at even now. and if you need to composite, rotoscoping or other special effects, it nice to have those kind of options.

I would like to see a deeper level of integration than simply comandline between programs. this means a separate tool that talk not just on one machine but across a network, so colorist, editors, rotoscopers, animators, audio are all in touch real time with the same project. this would be bigger than just Kdenlive....

I think we need to completely conceptually design the perfect system for all of us. We then need to create a (PDF) document outlining what we want. We should include workflow pictures and perhaps a rough design of each programs GUI. Included in this design would also be what currently active projects we feel would be best to use and what projects need to be created.

After that is done we should present it to each active project's head developer. We should explain that we are small group of professionals working in the video and film industry and further explain that this plan is what we would want from an open source workflow. We should also make it clear that we don't want to run their projects or even tell them what to do. The plan presented should be considered an outline and if they feel it is a good idea they should integrate it into their own develop goals.

The creation of this project or the formation of a video project alliance is not something that will be made over night. I think for now we need to plant the seed and let it grow.

MillionMonkey (or anyone) if you need to contact me so we can coordinate my contact information follows:
Phone: 813-310-4868
E-Mail: Alan.James@9IronProductions.com
AIM: AlanJames9Iron

@AlanJames1987: Now that Kdenlive incorporates a complete set of features, it is very likely that the development team will follow some directions from one release to another. But not all directions at once.

For the next releases, we plan GPU decoding, a better H264 support (timestamps). Then, there are other directions like a better color management. Maybe a better keyframe and effect GUI.

All this demands efforts and a lot of development time.

And of course stability is a major issue.

Also, don't forget that Kdenlive is MLT user interface. At each step, you need to ask yourself "How can I interface this feature with MLT and which library do I need". This is mostly a free-software hunt for other free software libraries.

Last of all and pardon me if I am clear as possible:

We NEVER ever look at commercial products. There is plenty of free software to look at. For example showfoto for effects and color management.

If you keep on taking about this and this software which sometimes release patents, we may become very upset. This kind of information "Kdenlive should look like product X" is of no help. We do not copy closed software and do not intend to do so.

Understand we don't even run Windows and I bought a Mac OS X computer for compilation only. But only recently. So a better focussed description of needed features is better. Describe what you need, not what Product X provides.

Don't even make sceenshots of commercial products. We will not read them.

Is that clear?

I understand your reasons for not looking at closed source software for inspiration however I feel that is a very closed minded way of developing. I will not presume to tell you how to run your project though.

I believe that if open source developers will not be open minded enough to give their users what they want or to examine the competition they should expect to not have any users. Ignoring either of those is ignorant and bad business practices.

As much as I would like to use Linux as a production platform I can not currently for lack of software. I don't know how to write code or how to bug test and thus am not able to contribute directly. I can only provide you with what I as a professional video editor would like to see from KDENLIVE and projects like it. If this is not of value I understand.

This has no bearing on this discussion but I feel the desire to inform you that I am currently using Windows 7 and plan on using it into the future. I will also continue to use the Adobe CS4 suite. Until a usable suite is developed for Linux it is of no use to me.

I believe you are making good head way on your project and I wish you good fortune into the future, but for now you have lost me as a user because of your lack of features and integration.

I hope this is not the end of the Video Project Alliance... This thread has been great sofar and a wishlist of professional video editors needs and some research into how to implement them will surely benefit everyone. An open and sometimes little emotional discussion is just the open source way to do it.
(I can remember some GNOME, GPL3 and Mono heat in the past :D )

Let the seed grow!

Dyngo and Weevil I would like to see my and other people dream of a united video production suite fulfilled but because I am not a developer I can't do anything but suggest what changes should be made.

I think projects need to follow what they set out to do at the creation of the project. I also think that this is the developers project, but they will not be the people using the software every day. I further think developers should listen to the people that do use this software every day but I will by no means say it is their responsibility to listen to us. I believe in their right to progress their project in any way they choose.

The capitalist system is a rough system and if you can't measure up in the eyes of a user (of any product) the user has the right to leave you of their own free will.

I feel that KDENLIVE is going to be a very good product one day but only by listening to and integrating the needs of a user. If the developers decide it is to hard to create a needed feature necessary to a user then they should expect that users will leave them and use a product that does have that feature.

I will continue to monitor this thread but I leave the ball is in your court developers.

"I feel that KDENLIVE is going to be a very good product one day but only by listening to and integrating the needs of a user. If the developers decide it is to hard to create a needed feature necessary to a user then they should expect that users will leave them and use a product that does have that feature."

Sure kdenlive is not a jail .... the user coud also learn how to program and help the other developpers ..i know it's not easy but we are talking about dreams ... and mans need a lot of work to walk on the moon ...

I am the lead maintainer of MLT and dvgrab as well as casual contributor to Kdenlive. In the past, on this forum, I have reacted negatively to threads like this, but I am trying to give this one a fresh approach; I do not want to be discouraging. In that light, I suggest that you approach this process carefully with the developers. Many open source developers are sensitive to taking direction and hearing opinion about priorities. Also, take into consideration that some are only hobbyists, some are part-time professional open source devs, and some full-time. Consequently, it might not be good to present them a grand master plan that represents the professional market. No one wants to feel like they are being asked to work for someone for free. It is fine to make a plan and to publicly collaborate on its development. However, please take into consideration the current state and the small steps needed to get to what you want. Then, lightly approach the developers in incremental steps in the form of bug reports, suggestions on how to rework something specific into something better, and sensible, next-step style feature requests. In other words, behave like a contributing power user.

For example, if you write to me now and ask for 10-bit or higher processing in MLT, I would respond by saying that I am willing to work on this under (paying) contract. Currently, I have more basic things to work on in addition to the occassional MLT-related contract side job. Simply put, a contract affects priority. Outside that, I take into account personal interest and things I think would benefit my fellow casual home user. On the other hand, I can forsee that higher depth processing will become more commonly beneficial and that I will develop a personal interest in it. However, it will just have to wait until I feel like more important things have been addressed. But also, this is an open process, and if someone else wants to do it, great! Assuming no one steps up to that challenge, and outside of a contract offer, what do you think you can do to help me get to that point?

+1 Dan.

Dan you are always very polite, to the point and willing to explain the thought behind an act rather then just acting on the thought. Thank you for that.

I will try to rework my schedule to do bug reports once a week for KDENLIVE. As long as I know there are developers interested in developing the features I want I am (or course) willing to help in the way they need me to help. I don't expect someone to just give me something with nothing given in return, so if its bug reports you want, bug reports you shall receive.

Dan if it is okay with you, MillionMonkeys and I will be in contact with each other to develop a long term plan. After it is developed we will send it to you to revise. You have a better knowledge of the inner workings of how open source software is developed and thus can inform us of how to “pitch” the idea most effectively. I am used to dealing with paying clients and I bet that approach will not work well here. As I said we will only send it to you if you are willing to help us revise it.

Again Dan thank you for being polite and informative.

Hello,

Please excuse my terrible English, sometimes in this post I will be rude but i's because I don't how to say things in the good way. (Jean-Michel, if you have time maybe we could talk in French at phone or via skype. If you want we can also met)

I'm a French movie director and I know a lot the industry. I worked as editor and know all the existing application. I worked too as a cinema teacher with kids and teenagers.

The only reason why I still have Windows in my computer is my need of a good NLE. For the moment, I can't switch to Kdenlive because of a lake of basic features. This week, for an educative project, I propose to the people I work with to use Kdenlive. Because of those missing features we couldn't and they decide to buy 5 FCP licences. That was very disappointing for me because I want to teach kids how to make movie with no money. I don't want to teach them how to work with expensive not free applications.

I really think that Kdenlive could be a killer apps : the engine is efficient, the design is clear, etc. but it can't be use by professionals right know.

As a professional, I'm not (really) interested by color management, compositng and effects. I just want to be productive and accurate in editing and to be able to communicate with the other actors of the post-production. I have to warn you that I will compare Kdenlive with Avid Media Composer, FCP and Premiere. I don't propose you to copy those, maybe you can imagine how to be as productive than those 3 applications. But I have to explain you what they have and you haven't.

So, after those long preliminaries (a French specialty) here are my suggestions.

I) Productivity in editing.

1) Separation between video and audio.
For a pro, sound and images are always different. We want to see immediately where is the video and where is the audio. But we want also see i we loose the syncro or not. In FCP, you cant easily link and unlike audio. In avid, when you loose the syncro, you cant read, in the track, for how many image. Actually, in kdenlive, the split audio function is nice, but i would like the possibility to choose "alway split" in the settings.

2) Nice layer logic.

As said in this post http://kdenlive.org/forum/change-timeline-logic, the layer logic should be this one :

Vn..
V4
V3
V2
V1
-
A1
A2
A3
A4
An..

That's the way the editors use to think, think today and will alway think because V1 and A1-2 are the main tracks. Other are complement and need to be far of the center.

For now I will use this for all my examples.

3) Always know who goes where.

We want to chose witch track of the player goes in witch track of the recorder. For example :

---------
- | V2 |
V1 | V1 |
---------
- | A1 |
- | A2 |
A1 | A3 |
A2 | A4 |
---------

Source V1 goes in V1 recorder.
Source A1 goes in A3 recorder.
Source A2 goes in A4 recorder.

Another example :

---------
- | /V2 | Inactive V2
V1 | /V1 | Source V1 try to go in inactive V1
---------
- | /A1 | Inactive A1
- | /A2 | Inactive A2
A1 | A3 | Source A1 goes in A3 recorder.
A2 | A4 | Source A2 goes in A4 recorder.
---------

In this case, you just put the sound in 2 track and don't touch a the other tracks.

4) 3 point editing. VERY IMPORTANT.
edit: See also 3 point editing —Granjow

3 point editing is the key. A NLE where you can't do 3 point editing can't pretend to be use by professionals. You can imagine other logic but : 1) why reinvent the wheel 2) the editors love 3 point editing 3) (for the moment) Other logic are slower and less permisive.

So, what is 3 point editing ? I'll try to explain but it's not easy for me in English.

This is the content of my source and my timeline (for the examples, I will use only one layer), each letter in a frame :

SOURCE | V1 | 1234567890
TIMELINE | V1 | ABCDEFGHIJKLMNOPQRSTUVWXYZ

Case 1 :
Now, i put a mark IN and a mark OUT in my source (to say, i want this part of the rush) an Mark IN in the timeline (3 marks -> 3 point)

SOURCE | V1 | 123 i 456 o 7890

TIMELINE | V1 | ABCDEF i GHIJKLMNOPQRSTUVWXYZ

Now I have 2 solution (corresponding in 2 buttons in the interface or 2 shortcuts in my keyboard).

a) I press REPLACE : my selection of source go after mark IN and replace the same length.

TIMELINE | V1 | ABCDEF-456-JKLMNOPQRSTUVWXYZ

a) I press INSERT : my selection of source go after mark IN and everything after it move.

TIMELINE | V1 | ABCDEF-456-GHIJKLMNOPQRSTUVWXYZ


Case 2 :

SOURCE | V1 | 123 i 456 o 7890
TIMELINE | V1 | ABCDEF o GHIJKLMNOPQRSTUVWXYZ

REPLACE :
TIMELINE | V1 | ABC-456-GHIJKLMNOPQRSTUVWXYZ

INSERT :
TIMELINE | V1 | ABCDEF-456-GHIJKLMNOPQRSTUVWXYZ


Case 3 :

SOURCE | V1 | 123 i 4567890
TIMELINE | V1 | ABCDEF i GHIJK o LMNOPQRSTUVWXYZ

REPLACE :
TIMELINE | V1 | ABCDEF-45678-LMNOPQRSTUVWXYZ

INSERT : NO


Case 4 :
SOURCE | V1 | 123456789 o 0
TIMELINE | V1 | ABCDEF i GHIJK o LMNOPQRSTUVWXYZ

REPLACE :
TIMELINE | V1 | ABCDEF-56789-LMNOPQRSTUVWXYZ

INSERT : NO

I hope you will see all the advantages of this logic: rapidity (just one click or key is quicker than drag'n drop), precision (I know exactly what's happens and why) liberty for the editor (for the moment, in kdenlive I can't put a clip between two).

5) Cut and lift in the timeline. VERY IMPORTANT.

No need for words to explain this but a letter drawing (# is a hole in the editing).

TIMELINE | V1 | ABCDEF i GHIJK o LMNOPQRSTUVWXYZ

CUT : TIMELINE | V1 | ABCDEFLMNOPQRSTUVWXYZ
LIFT : TIMELINE | V1 | ABCDEF#####LMNOPQRSTUVWXYZ

6) Trim mode.

I don't know the English word for "raccord" witch is the cut between 2 clips en the timeline.

Well, with a trim mode you cat adjust with a lot of precision the "raccord". For example you can say : I want 2 more images more in the left clip and 3 less in the right one.

Do you need a drawing ?

7) Many editing in one project.

We need many different sequences in the project for different reasons :

- We need to keep older version of the same movie and, the possibility to get back just one part who was better finaly.
- For a long movie I can edit the different scenes in a separate sequence and, after, edit the sequence together.

When I put a sequence in the player I don't expect to get it like a complete movie (like in FCP) but like an editing with many clip (like in avid).

8) Shortcuts for everything.

In Avid, every key of my keyboard is a shortcut and I can change them easily. A good editor barely touch is mouse.
The more important are INSERT, REPLACE, CUT, LIFT and J-K-L that should work like in this post : http://kdenlive.org/forum/change-way-j-k-l-buttons-work


II) Communicate with the other actors.

1) With the color-guy (how do you call him ?) and the laboratory.

I won't do the color stuff in my editing computer : It's not my stuff ("Etalonneur" is a difficult job) and I won't paid for RAID hard drive and DLP screen necessary for a good job.

So I need the possibility of "offline" work and to provide to the lab a list of my editing decision. That's the definition of an EDL. So I need EDL, OMF, AAF... something that allow me to have a conformation of my work.

2) Communicate with the sound editor.

Between Avid or FCP and Pro Tools or Nuedo we have OMF.
We need something between Kdenlive and Ardour. That would be a first step.

But I scared to need also to export to Pro Tools for a long time : every sound editor I know paid a lot for Pro Tools and they won't switch to Ardour for me.

Maybe tomorrow if Ardour become the reference it should be.

So, one more time, OMF or AAF.


III) Small little stuffs.

1) Fullscreen in my second monitor.

2) Categorization for effects.

3) Easy access of fade.

Fade should be possible between 2 clips in the same track. Doing a fade is not compositing (for most editors).

3) GUI profiles.

Screen organization for editing, for recording, for titling, etc.

By the way, all the philosophy of editing is to compare 2 images : the one I have and the one I what to add. So the default set should be with 2 monitors (clip an project).


OK, I think that's it for the moment. Sorry for this very long (and hard to read) post. I repeat, I really like kdenlive and hope I can help you to understand professionals needs to make it even better.

Romuald

Romuald, thank you. Your English may not be perfect, but you communicate these ideas quite clearly. And it's actually quite good to see stated clearly the needs of a professional editor, and to have some exact suggestions for how this app might be molded to become a better tool for this purpose.

I think it's very important that these ideas be considered. They are clear and logical and they are necessitated by the way movies need to be put together, and this goes for both pros and amateurs.

Right now I think Kdenlive is developing into a first rate amateur/prosumer video editor. And by this, I don't mean second rate quality, I mean that typically one person on one computer is going to edit and completely finish and render the final product, with no or very few other outside apps. This is a very different kind of environment than the film industry with it's divisions of labor into editing EDL, colorist, sound, titles, effects etc. as you indicate. So the prosumer NLE must do it all to some degree, and walk a difficult line between power and ease of use. In fact, it must have some color correction and sound features etc. that are maybe beyond what is necessary by someone who is purely editing. In other words, most Kdenlive users will not be correcting color in a special outside app, nor will we be massaging 10 tracks of audio in Ardour, for the most part.

However, Kdenlive seems like a very flexible video tool. I think that it may well be possible for it to be a simple yet powerful 'prosumer' tool, while giving the pro editor some options to work in a more 'industry standard' environment with things like 3-point editing.

In fact I think the 3-point editing and trim editing procedures you talk about are the things I miss the most in Kdenlive as it stands right now. Being an efficient 'cutting' tool is job #1 for an NLE. These cutting modes go to the heart of the editing process and I would very much like to see Kdenlive get these things really right.