alright see so
hello everybody
thanks for being here can everybody hear me directly in the back and in that
side yes
but
so if you hear in this room you probably already know what it is it's
a video editor resonant typhus and made of also
it's cheese weaponry includes such elements as a huge file formats but it see you
can pretty much throw anything in to the T V like
you could we makes windows media video files with your cameras quicktime files and you're
used videos
it has the best of breed you are design it's very user centric it values
efficiency and
it's designed in collaboration with filmmakers and if this experts
in a rather well so and
it's also an avid it's or that aims to not have
this notion of market segmentation that you see in the proprietary well where in the
propriety world you would see well we're making this low and videos it is or
less meat super simple with limited features and the super high in super complicated thing
with five hundred how options in the preference dialogue just so that it looks professional
and we're gonna price differently and this features more interest act and we don't have
that here
so
i want to spend a few minutes to actually talk about the mission the vision
we see why we're working on it why it matters
because i haven't done that very often
i was based i sat down and i started thinking about this after all these
years of being involved in one way or another
with this that project there ought to be a powerful reason
why and other contributors are stuck around and continue pushing things forward
of course everybody starts by scratching an image so i and i remember how in
two thousand five
i want just wanted a replacement for big S
and i thought well this projects really cool it's open source so we don't need
six months right
so that's the reason you get involved in the beginning
but then as time passes
as you make friends to the project
as you see people being enthusiastic about what you do and people putting trusting to
you even this takes a review notes i mean you're fighting against impossible odds
you need to find something more compelling to keep you motivated
and here's what came to my mind
the T V in pretty we believe in allowing everybody on the planet
to express themselves through filmmaking
we choose that they can own and improve
unsurprising yet but that what segment on the website in a surprisingly
recently it's you can to be assisting to means that wow that really resonates with
me
that's exactly my thoughts
so okay at first and this might seem like a very
draw insignificant state
but now it's very powerful actually
it gives you a goal in life to run these people around it
and it's a very important facets of our goals
like allowing everyone that touches upon them know market segmentation thing i mentioned earlier that
means ground much as well as the kids in school doing short films or an
activist recording public protests or and event independent professional filmmaker wants to be a master
of for tools
we want to capture to all those people
and the notion of mastery brings a second element in that statement
which is owning and improving the software it has to be open source
it has to be easily tractable
so it's disgusting partition
of course made of also
you have to clean code base
and has a clear documentation around how to get involved in the project
so how do we do all these things
well that this means a couple of things
this means the at the forefront of technologies the latest educated latest use your
we have this upstream first approach which
people in the just in the community and of to the right i mean see
will probably be familiar with already it basically means no axe
you discussed we actually such as opportunity streamer three P K and everything
and you work out solutions in corporation with them and you don't put some stupid
have in your application downstream
in that a fixing the problem for everybody
and you each your own that food of course so you and you talk with
filmmakers i have to make a right to consider myself a filmmaker and go on
with are rather we have
and to T V also acts as a catalyst
it's a torture test for G streamers so these guys we kind of like you're
scared of me way ready and running a
at chasing them down conferences
so
completing this mission is the property it's rather difficult
technically speaking it's a very complex piece of software
and to add some more difficulty well your users are typically not developers otherwise they're
of every we are a there are your users are typically not developers and your
developers are often not users
or the ones that where users and became developers are not uses anymore sometimes a
very interesting station
and there is no sustainable business model around an open source we do it it's
or because it's just too much of an each
and you have too much competition from the provided video editors which are actually pretty
good
and that's just the surface of this thing if you're right interested in details of
the economics of open source video editing software and everything i you can check out
the presentations i gave that did like graphics meeting two thousand ninety two thousand eleven
or maybe we can discuss those things and now something but that those presentations are
actually on the P G website so they are easy to find
and for this for the next the or for this a talk or the next
lose you slightly i'll be focusing more on the technical side and a little a
little bit of gonna mix
i'll be mentioning what we've been doing for about one or two years you're doing
now what will be doing next
so in two thousand and ten
G S came out
it's a nice you guys very there is part of this tumour and allows you
to create
audio and video editing applications without requiring at the H T in astrophysics
so it's meant to be a complete highly and high level solution and not just
for everything but also playback and coding media discovery et cetera
and it's basically the result of the years of experience in developing it's easy another
applications in the despair world "'cause" you have seen on and but you know i
mean it's nineties timing and dynamic pipelines and all that that's nearly not enough
for you need you need far more than that to be able to make a
real we do it at the
so G S is nice a uses also uses very subgoal centric technologies like G
to G object so it uses G T K dot still i think
and it's fully introspective so you can use it in parts and you can use
it in that in the other languages
but fight and it's best of course
and then in late two thousand ten so one year and something later and early
two thousand twelve we decided in the P C project well
it's time to switch to gas it would be a good thing to do because
it doesn't make any sense to maintain your own if you think back and anymore
and it would be stupid not to make a reference usable for other projects
so at one last year
this was a status report that we have on this was are tremendous that's is
around E S
you had the core features in the first group
that were implemented
and some features the second group that we're being worked on as power to the
summer of code
and the remaining features on the road map or left to the future
and this does not include the features in an improvement simply T V itself that's
pretty much just centric around the port to gas so i don't really have time
to go over this in this presentation but if you're curious you can go to
the T website
there is a on the showcases a recording of my talk and what it two
thousand twelve which turns out to be the only surviving record of what the two
thousand twelve and it was it would be T V coincidence i think not
and then in early two thousand well without okay the plan is we stay will
actually a gas we try to get a release out of the door a sad
before looking at working to just one or one point zero because i just run
point there is not yet ready you know we weren't sure
but then assess you once we were in a situation where
well nobody was looking at system is over that's anymore
and were quite behind already the are initial schedule so at this point in time
to make incense
to keep going with zero that ten
and to get any amount of support we needed to use just or one point
zero
well other than that
we didn't foresee a huge amount of work left to be
able to make a newbie to release
the biggest obstacles we predicted
where that we would have to switch to discover one point zero and you take
a three and you to construct spectrum all at once
because of the complex tonight's topic sixteen this picture
and
so
not such a big deal
yes
and then
okay
and system of one point zero and one of so will be here
and i'll see all those little a you know like with the
you know if we just take care of those where we're done right but that's
really vote
is a hard thing you
i'm sorry
a violent
this however
this one point zero
wait what going off like
i think that's
but exact holes
over the
really that is gonna take care of this time
my i think she it is
all this work
there is a stuff can in the just replace plug-ins
it's not a summary of what has happened the last few months
and that's here we are when you're to what it two thousand well with that
bunch of new regressions that have been fixed and well and you features implemented and
at time of worked on two dates use your
and contrary to the last year this light only talks about the president's
the first list of items are what's don and either all features that we're reimplemented
usually much better than before or new features
and the lower the last is the buttons that we items or stuff that we're
still working on this summer
so let's look at the future what are we doing right now what's
in the compare
well i have always have a bunch of plants based on my own intuition and
my experience as a filmmaker once in a while is good to take a good
hard look at yourself and we evaluate the priorities and ask your user days what's
the most urgent and important then
so we recently made this survey asking well what the users want because it's you
know all about users
so we have a sample of that thousand three hundred and eighteen people when certain
these survey
the source the internet's
so i don't know trying to be T V social networks
likeable close all my god we want to thanks mate i have to say thanks
to joyce a then there is always a nice a nice to
help
because the allowed us to get a lot more coverage from different from a different
audience than
if it was just from planet the home or
did you fans
so we have a sample that is a lot more representative of the population at
large sets is to be statistically speaking
some pretty confident in these results
for once
you may think that a sample size of a on a little bit over a
thousand users is not that much
actually i that's something you can find out from my talk that L G M
it's pretty close to the size of the population of people who do video editing
on the next
because you have like you could to estimate to the amount of people are using
and actually this are one percent of people in general do you anything and then
one percent of four percent of one percent or something you end up with not
many people
so it's pretty close to the population size
and so before we and the lies to results let's take a closer look at
the sample
i don't know why i said that the is just
we have a
i a good portion of people who actually strike a T V or arctic to
be users
but we also have a bunch of new people that's a never heard of it
and some people who
just like other software and that's fine it's interesting
because this the surveys not just use for us it's also useful for whoever is
interested in that we do it it's or the point so as to do it
it's a nice
we that could very sample which mostly follows and all normal colour sure
so i have a question in the survey that says how do you rate from
the scale from one to ten how familiar you are would video editing are your
professional or you be so we have a bunch of newbies on the lead well
it's not just like a bunch of newbies and a bunch of professionals that we
have a nice
normal curve
this can work things up a little bit in the next chart because you like
mixing professionals with normal people
but that's okay because we need to we want to capture to everybody and remember
market segmentation is a construct of the mind
rather unsurprising result here
mostly nobody cares about able to platform
or these nobody in our sample
and among those who do only a small fraction would be willing to pay for
an open source we do with the church reported created platforms
so i'm not gonna care about that much
and because frankly we've got enough on our plate already
but if someone offers
so you help or think to be to other platforms are even write a new
video editor based on to yes
you're welcome
do experienced users have a different requirements require regarding the multi platform thing i just
talked about well there's a slightly bigger proportion of then
that care about other platforms but still the vast majority does not care about other
platforms and the next
interesting
and now what do people in general want the most out of an open sort
we do it or
surprise people want basic shape to work first and foremost
and
no one cares about collaborative editing actually
well at least until you've got everything else nailed down
and people didn't believe me when i said that two years ago
well let me say this now i told you still
but i can this is for the entire sample
what if we limit ourselves to a subset that sample let's check what happens if
you take only those we evaluate their skill level to seven or more on a
scale of one to ten
interesting enough
the number one feature is still simplicity and stability
and then comes proxy editing anti free mobile effects properties
suddenly to becomes pretty much just as important as stability
and then really simply to number one feature that instead of stability
all the other features suddenly become a little bit more important
so the take away is that pros want everything and we want all the rock
solid no big surprise
but it does confirm the order of priorities that i didn't suspecting for years
and it validates the features that i'd be in prioritising in the upcoming to be
released
the good news is that much of it also has implemented the key frame able
effect properties recently and that and some of that is a thing is laid the
groundwork for proxy that thing and then so is that it is actually working towards
these slow and fast motion feature
so
we're on our way
be people one quality and stability
and in one of the top users and wanted for free and they want to
know well we knew that
it confirms what we suspected features are nice when stuff works
historically we have a quality and reliability problem including being sure
see those that nice reading we got among which is users
and there's so much that can go wrong in a nonlinear but you know to
meet a pipeline that
below the surface of the you why
we have to address that some
stability is far
stability with a mind boggling complex underlying a non linear fitting application is even harder
it's time for quality so let's try so that's that
they with been thinking about various things
that we might need to do this might be one of the
actually we probably won't field in on in the traditional sense
we're hoping to do some phrases have you respect me around
and we're entertaining the idea of merging it with gas or merging it into that
system or basic workings
and the argument for remote merging it into this same because it's re as yes
is that would make it easier to harmonise releases
the argument for merging to non in into the G C plug-ins space is that
welding on is an elements not library like D S and S
well
dft plug-ins basis meant to have elements of all types including to do with the
use cases
so you are the big advantage of putting on this debate is a would be
under the responsibility of those guys the commentators and that would be better tested i
hope which is certainly an interesting prospect
so if you're curious about things we brainstorm the little bit you can see about
number know about seven hundred one two the two eight seven for details
but this thing
test all the things
we recently started experimenting with and automating
automated test server for everything for P T V N G S N G stream
originally and everything so everything it runs pretty much on every time someone come it's
something in G stream or
which is quite often
and it's allows us to spot regressions and know where they come from instead of
just spending two days trying to track down wanting is in everything broken why is
the my quicktime file not seeking anymore or something like that
it's much easier with that at least i've been told
the interface is and french sorry about that but that it's a difference conspiracy
and the other things you can do
well we also have these don't tell tests to just you want
automatically so right now we have the most common to be like scenarios that as
well as some corner cases known to figure bug like try to split a hundred
times it and also makes it very easy to write you access you just tell
a sequence of things to do you like click this button and then search for
the existence of this you why elements with this label or something
and it's actually really good way if you're looking to get involved in project in
the quality
i if you know some of the corner cases or
some of the things that are annoying to test and to reproduce bugs
come right to test for first test for your issues
systematically so that we keep working
well
so
stability is paramount
but people including me expect to have you know all some exciting stuff to make
complex sitting possible
so let's take a look at some of this exciting stuff from the last few
weeks
we killed good and that's
this is thanks to matt suitable sell you be factored and for to our time
like to clutter
in two weeks
see so if you don't that just before this summer of codes project started this
year
just to show how much of about that is
and
and the all this is another awesome
you can finally animated properties of your effects
so let's take a moment to marvel at how also places
here on my old let since there are key frame curve for this clip and
i can click on the key printer and miss the target but i can click
again and moving some T from upwards to nine controlling the saturation
and
this will change the amount of like how popping the colours are and it fits
so well way too much and
you'll be able to see the institute and result soon
so it goes from black and white to full colour
and black to back to black and white so you can animate here stuff now
which is super importance for crazies like me
pretty cool
speaking about that see
matters some real good project can be summarised in two words
fix
everything
so this goal is to fix bugs all atrocities from is that so that we
can actually really something without causing protests and riots in the streets
and so since the beginning of this do you some project from what i can
see from long but see that he's been involved in something over forty bugs in
G stream or
and
well we have a lot of the
nice stuff coming up we services
project for the some of code is quite different than matthew
i
which is basically you can be we are the clip
so that's example of one but roads are is will be ready finishes
no pressure
so regarding motion renting i have some thoughts on how to make an easy to
use you why for it and would like to discuss that with the advanced filmmakers
in the room us with your we do anything and this these can talk to
me interest about this
okay
other stuff when rendering we actually cast use humour pipeline errors now and we show
them to you instead of just randomly freezing up without saying anything
i think a
is a feature that about
and we have a nice with high resolution logo i can to which is always
nice and we also have new icons for the buttons on the in the you
like the timeline which is our pixel precise which is nice until you're running how
you guys we know
i'm also working on a more than fully responsive website design for P T V
which i will publish continue please comes out any day now
and i should hold of other stuff the stuff i mention here is
just what has happened in the last few weeks
the list of improvements is why that last year is really don't i'm not sure
i would have time to cover here
actually been working on it for the lot
well
i've been working on my talk for the last few weeks but
did still didn't have enough time to prepare for it today
and there's too many interesting people in glottic so this is annoying
so anyway you're curious you can take a look at or blog posts
and you have a nice very nice summary of the improvements
since last year
so for now simply invite you to join us and share the software we've got
a great technology got nice contributors and mentors very find friendly i see channel
call these will architect the that you want is separate from the core we have
a long project is three and we have a lot of experience
don't make the mistake of reinventing the wheel like countless others well unless you reinvent
the wheel using G S that's okay
we have a three weeks merge policy which we try to follow as much as
possible so if you try to contribute something to be a pasture artwork or something
where at least gonna try to we're gonna review which and if possible marriage at
the main that one branch and there's plenty of other reasons why it's cool to
and it to be to be on the website
and we want to make this stuff using so we have an automated build report
so this stuff will check if you're disrespect is recent enough and will allow you
to build everything in one go so it's a pretty easy
to get started it to be hiking and it's also useful a starting point if
you're trying out development versions of thirty streamer without touching rest of system is actually
useful to also test some simple bugs like those quicktime files actually play or see
so as you can see there's many ways you can help us
even you can writing new specialised it using of vacations
that's helpful for example
do you guess could be used to create a give you answers based realised for
recording and processing talks at conferences such as glottic
i'd be happy to help with guidance a not by support that
so i wouldn't code because well there's just so many hours in a day but
to come and see us and will help you get started if you're interested in
subdirectories project
and if you want to do you want work what we have some stuff but
this is just an example please finish my title it for brunch i'm sick of
from i think this feature and there isn't much work left to do and it's
a fun project that so come see us if you're interested
is save lot see in the room
i'm looking at you
so we will be having a fact is the this week so not sure where
we will be located
maybe we'll just be in the auditorium or that's a lot of tables
or something will see we can check our google plus event based or might sort
of what i would there be for details
and that's pretty much it thanks for your attention
any questions
my killed them all
you use P T V on any of the was you were showing us
want to right
i use that you were showing inside
your presentation i cheated and then use at id marks or something else
just to cut the middle and the end because i didn't want to re encode
the videos or maybe i did you speak to be for one or two of
them i don't remember
but
i had to fight but id marks box
that's that
very frustrating when it crashes and you have to start over again and again
every we do it in supper socks in a different way
other questions
insults
rotten fruit
there's rotten fruit over there
so i'm sorry i'm is the beginning readable
down but i know that about the T what's your plan in
because there's gonna be a lot of issues of the me and it's really in
the end
as you already said
what plans you have to have as many tests as possible
in P T V would you yes
so that in the end us G stream of can
and something break speaking
i guess needing to find additional time to write those
did you see is that i was i missed the beginning so did you missed
the automated testing thing
the want was the dock no before that actually me remind for five no point
ultra melissa like
in this i think that so long and that has always that you have seen
this i think the menorah it was in my point i don't mean you know
to make you know driving spend well point of test
and how many are gonna do right saying and so forth there's another member of
the for instance there is over there wants to answer
i should check out the give actually we have
lot of do yes
of course with a plane pipelines stuff like that
basic pipeline with the you do mixing origin mixing or transition stuff like that we
just wondering which it but the
they wanted five is the is actually correct
and then the idea is to have a like verification with the human eye
we just check
usually it's right
points
we just take this easy to magically
like to be sure it's right just
my
no other questions
wow that's like process questions how will i think or is very well