Re: [linux-audio-dev] Poll about linux music audio app usability

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [linux-audio-dev] Poll about linux music audio app usability
From: David Olofson (david_AT_gardena.net)
Date: Tue Jun 11 2002 - 04:02:21 EEST


On Monday 10 June 2002 07.10, Juan Linietsky wrote:
> I thought this may be of interest to the list.
> In a k5 poll about usability of linux audio apps,
> ( http://www.kuro5hin.org/poll/1023512126_OSelOkZS )
> So far, out of 38 answers the results are:
>
> -How do you like music software for Linux?
>
> 2 % - Great! It has everything I need.
>
> 13 % - Good, but i wish apps were more userfriendly (Like Reaktor
> or SoundForge)
>
> 31 % - Could be better, I think the apps are not yet mature enough
> for my needs.
>
> 15 % - It's unusable, the apps plain suck.
>
> 10 % - Dont care about composing on computers
>
> 26 % - Dont care about composing.
>
> ----
>
> I think this raises some questions.. My feeling is that most people
> aiming to write music on this OS is expecting to have apps with
> super easy and intuitive interfaces, where you only go trough
> displays, knobs, sliders and paintabe areas.

Personally, I'd rather have something that sort of feels like the
applications you mention below, but with a unified backend that
enables the whole set of tools to be thruly *integrated*. I'm not
talking about host applications displaying plugins in little pop-up
windows, and some next to useless keyboard shortcut system based on
drop-down menues, but rather a visual DIY "Super Tracker" toolkit,
where the user can *really* build his own, highly streamlined
studio-in-a-box.

Oh, and users shouldn't have to learn a scripting language and hack
pages of code. That will *never* work, regardless of what kind of
users we're targetting.

I know I wouldn't like to mess with that, even if I could do it in
*asm* if I had to - I'd much rather just create music.

> Why we dont have apps such as Reason, Reaktor, Sonar, Sound Forge,
> etc? I dont mean full apps, but at least projects aiming for that
> kind of thing.

Well, building good GUI intensive applications is *lots* of work,
even with strong visual tools (they don't build much of the logic for
you...), and it seems like most audio hackers prefer hacking DSP
code. Meanwhile, few GUI hackers and designers are into audio and
music.

So, the part of the audio community that would care to hack that kind
of stuff is too small. There will either be one collaborative project
that will eventually break down because not even two developers can
agree on anything - or there will be multiple projects with one or
two active developers. (The sequencer projects currently in existence
all seem to be of the latter sort...)

Finally, I might have touched on another major issue above; "I'd much
rather just create music."

Right; so if I'd start hacking an audio sequencer, there's a big risk
that, as soon as it's mature enough for my own needs, I'd stop adding
features and just use it - happy and feeling generally good about
having solved The Number One Itch of mine; needing Windoze for music.
I might not even function as a maintainer of the project any more, as
I've effectively turned into a happy user with no more itches strong
enough to have me dig out the development tools again. (All right,
once a hacker, always a hacker - but you get the idea... :-)

The only hope for such a project to *really* mature before it's
obsolete would be that when the original author is happy, the project
is passed on to a new maintainer who's just dying to add his/her
favourite features. Continuity. Allowing the project to go where the
visions are.

> We do have very powerful tools, but i have to admit that for most
> of them we have to learn some script programming.

Yeah... (Feeling guilty for having invented Yet Another Synth
Scripting Language.)

Anyway, it might be easier to wrap a script base system in a glossy
GUI if one keeps the related issues in mind when designing the
system. It isn't exactly trivial to "generate" code for pure
scripting languages, but as soon as they approach markup languages,
everything gets *so* much easier - the language essentially becomes a
very flexible, human editable data file format.

> Do we lack good APIs? Alsa MIDI api is the best I have seen for
> it's kind.

APIs are great, although we do have an issue with GUI toolkits - or
rather, getting different toolkits to live together in peace. (That
would be even more of an issue with something like the DIY Super
Tracker Toolkit outlined above, as you can't get away with pop-up
windows there...)

> Also, sould linux apps really take this windows approach
> of making huge bloated interfaces with lots of eye candy,

No, but that sort of design does have it's advantages from the GUI
POV. Problem is that it takes a single team of quite potent hackers
to create something like that - unless we can figure out a way to
integrate multiple projects well enough to form something that looks
and feels pretty much the same.

> or should
> we try to improve on making our apps intercommunicate between
> eachother, while still giving some importance to ease of use?

Yeah. As to "ease of use", though, I don't really think there's room
for compromize in that area. IMHO, MIDI, soft synths, external
mixers, FX etc already make it a major PITA working with music, so
the only *really* important goal is integration - and that is
*complete* integration, throughout the system.

For example, a MIDI sequencer should know how to control every aspect
of every machine connected, or it will be forcing the user to keep
track of trivial "numbers" all the time. The way most sequencers
handle SysEx data, and even plain MIDI CCs, actually makes it look
more complicated and painful than it is. (It's not a good sign when
you think "Damn! It would be easier to hack a C program to do
this...!")

When I'm working, I want to be able to switch between projects
whenever I want to (I often work on several songs at once), and it's
an absolute requirement that *all* information is automatically
stored in the project files. Selecting another song and pressing
"Play" should Just Work(TM).

> What do you think about this issue?

Well, I think it has to be solved if Free/Open Source is ever going
to be a major force in audio and music. (Which IMHO, it has to be, if
applications are to become mature enough for serious use. "Becoming a
major force" is not a goal; just a way of attracting more users and
hackers.)

//David

.- M A I A -------------------------------------------------.
| Multimedia Application Integration Architecture |
| A Free/Open Source Plugin API for Professional Multimedia |
`---------------------------> http://www.linuxdj.com/maia/ -'
.- David Olofson -------------------------------------------.
| Audio Hacker - Open Source Advocate - Singer - Songwriter |
`-------------------------------------> http://olofson.net -'


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Tue Jun 11 2002 - 04:03:10 EEST