Kai Vehmanen is a CS student in Turku, Finland as well
as being a musician and audio programmer. He has
developed ecasound and ecawave to record and modify
sound files in the Linux environment and has used them
to produce his own music.
We talked to him by email recently about music and making
music with ecasound.
> Just to set the scene: I think you're in Turku in Finland which is not far
> along the coast from the capital Helsinki and across the water
> from Stockholm in Sweden. Is Turku a university town? Should tourists
> go there? (heh heh)
> You've got some nicely disparate musical influences including Jimi
> Hendrix (what do you think of Vernon Reid?) and the Orb.
Well, yes, Turku is an old university town. I'm still a university
student myself, majoring in computer science. Turku, in fact, used to
be the capital of Finland, but that was a long time ago (-1800). I'd say
it's a good place to visit. Not too big, but not too small either. :)
> When you're setting down tracks do you find you're in a Jimi mood
> (whoa, feedback!) or Orb, or do they coalesce?
I must admit that I haven't listened to V.Reid that much. After
Hendrix I soon drifted into more heavy (Black Sabbath and its
followers) and progressive (King Crimson) stuff. But for many years,
guitars were always there, whether it was rock or ambient.
> ... Electric Ladyland was headed that way too I think ... and by
> "ambient" I guess you mean all the textural stuff that goes on which
> doesn't overtly "Rock".
In fact, Jimi and Orb aren't that far away from each other. In my
twisted mind, many Hendrix classics are very ambient. ;) Voodoo
Chile(sr), All along the watchtower, 1983, etc. are good examples
of this. But back to your question. Nowadays I'm most interested
in ambient and idm (armchain techno a'la Aphex Twin and Autechre)
style material. But I still like to have as much live material as
possible in my music, and this of course makes a huge difference.
I'm also playing guitar (after a long break) in a more traditional
band, and this has unavoidably influenced my solo stuff.
> heh heh, Brian Eno sometimes enjoys counter-intuitive explanations
> and there's something zen about music-less music (the way most
> people would define "music").
> When I was playing guitar a lot I used to jam with anything that moved
> ... including CDs! Mr. Hendrix was very hard to jam with - great leaps
> with those huge hands of his and inspired "out of the box" stuff. Did
> you ever try and jam with him?
For me, ambient is something that creates a strong atmosphere.
You know, you put something playing and it fills the whole room
even if you aren't listening. Although I'm a huge Brian Eno fan,
I don't totally agree with his definition of ambient. In my
opinion, a good ambient piece works in the background, but is
interesting enough for active listening.
> Do you do your ambient stuff live? It's nice to do but in quite a few
> parts of the world it's a bit difficult to put together an audience...
Same here, I'm a huge fan of jamming. CDs, TV, radio, mp3s, anything
goes (even other musicians :)). I've noticed that listening and making
ambient makes it easier to jam and improvise, no matter what the
style is. When you just listen to rock/jazz/whatever, you easily
start repeating yourself, and it's hard to compete with the "masters".
At least now I feel comfortable jamming along with Mr. Hendrix himself. :)
One amazing thing about Hendrix is his live performance. We've
tried to play some of Jimi's songs with a trio band, and sometimes it
feels like we'd need three guitarists to play what Jimi played alone! :)
> ... speaking of which - what are the most popular music genres in
So far no. I'd need to get more gear (ADAT, minidisc or something
similar) to do it, and it still would be quite a challenge.
> Is there anywhere Turku readers can go in the next month or so to
> check you out?
Ah, mainstream audience listens to mild dance and techno at clubs
and bars, while rock/pop prevails on the album charts. So pretty
close to the world-wide average. :) We do have some oddities. For
instance, we still have an active heavy/death metal scene.
> One of the appealing things about electronic music is you can make
> your own timbres so you're never necessarily tied to existing voices
> (once you get the gear together).
> Was this something that attracted you?
I can't promise anything. Due to constant lack of time, this band
project is still in early stages.
> On the downside there are so
> many variables (even putting a system together can be a challenge)
> that many people get bogged down and go back to voice and guitar
> or whatever - a simpler world! Do you feel that at all?
And even more important, you have total control of everything.
In a band, you can tweak your marshall and pedals, but many times
you'd like to go beyond that. When you do everything yourself,
there's nothing stopping and slowing you down. But this does have
its downsides. You have to have a clear vision and know what you
want to do, or else you end up trying various alternatives until
until you get frustrated.
> Oh yeah, sequencing drums can be a bit tedious. The only "easy"
> input way I can think of is to play pads and record it with a
> sequencer, or set up groove samples to be triggered by different keys
> on a keyboard (so that you have a bar or two of drums per key)
> - but still you have to set the thing up and that sort of keyboard
> isn't all that cheap (or the pads either).
Well, almost constantly. I spend quite a lot of time
playing my acoustic guitar. If needed, I can play most of my songs
with an acoustic - even the ambient ones. ;) I think it's a good
way to refine your ideas. If it works with a bare guitar, you
have a winner.
One thing I've really learned to appreciate is live drums. After
spending hours and hours sequencing drum tracks, a good drummer
is almost like magic! With a band, you can just ask the drummer
to play something groovy and that's it. Compared to sequencing,
this feels like cheating! :)
> You wrote ecasound and ecawave to help your own music
> production process. What brought you to do these on Linux instead of
> some other operating system? (heh!) ... aside of the 'finix' factor :)
I tend to use all these techniques. Sometimes I play a whole
track with drum machine's pads, while sometimes I just record short
takes and use them as loops. When I need to sequence stuff, I usually
use Michael Krause's SoundTracker.
> Is that an implied criticism of the ALSA API or are we just talking
> about what existed in 1998?
This is a long story. It all started when I bought my Tascam 4-track.
I wanted to record my tracks to computer, so I could put them
on my web site. I was using Os/2 back then (1997, if I remember right)
and there weren't any programs that could handle large cd-quality
audio files. So I decided to write one myself (wavstat was the name
back then). Ecasound project started in 1998 when I got frustrated
with Os/2's multimedia API and decided to move to Linux. At the
same time I rewrote the whole thing in C++. I tried Linux for the
first time already in 1994, so it was the natural choice for me.
I especially liked the OSS API. It was simple, well documented and
> Looking at ecasound, it seemed to me that one's way of working might
> be a little different for those who've worked with other systems - in
> particular the mix-down process.
No, no. Although I've been quite happy with OSS/commercial drivers,
I nowadays use ALSA. ALSA adds a new level of abstraction, which is
absolutely necessary for supporting more advanced audio hardware.
In addition to ecasound and ecawave, I'm also maintaining native
ALSA support for SoundTracker.
>. Could you take us through a track -
> say, "traaginen", and outline the steps you took?
It's important to note that the whole ecasound project is based on
the libecasound library. I've tried to keep user-interface separate
from the actual processing side. Ecasound and qtecasound are very
generic interfaces. They aren't meant for any specific tasks. This
makes them powerful, but it also makes them somewhat difficult to use.
This suits my purposes, but I know there are plenty of people who
like more traditional and easier to use interfaces.
Because of this, my aim is to make writing a new ecasound
user-interface as easy as possible. There already exists a few
projects, which use ecasound directly. For instance Janne Halttunen is
working on gtecasound, a GTK-based interface for ecasound that is
meant specifically for multitrack recording. And of course
now there's ecawave. It's also based on libecasound.
I don't need these traditional user-interfaces myself, so I don't
think I'm the right person to implement them. Motivation is the key
here. If you try to write free software like commercial software,
you'll soon lose interest and the project never gets finished. I
intend to continue developing ecasound in such a way, that it keeps
my motivation up.
> So at this stage you set the relative levels of the tracks by
> trial and error? Is there a way of dynamically altering levels
> as you go (eg. track 1 starts soft and goes to max while track 2
> is fading from the start to the end) or do you have to set that up
> when you first record the track, or in an intermediate stage after
> the original record?
Ok, let's see if I remember what I did. :) Ok, this track
is based on a simple, repeating 909-drum beat. I created
it with my drum machine (boss dr660) and then recorded it
At this point I write a few ecasound chainsetup files - one for
recording, one 'mixdown-to-wav' and one 'mix-to-soundcard'.
Then begins the creative phase. I try out various ideas and
record loads of material (using the 'recording' chainsetup).
I don't want to fiddle with computer while I'm recording,
and ecasound makes this easy. I edit the file names to be
recorded, start processing and that's it. Sometimes I do
temporary mixdowns, which I then use as background tracks
This time I ended up using the original 909-drums, four
different guitar parts, live bass guitar part and
a live drum track (yes, live drum machine playing!).
You can hear the live drumming during the first minute.
Filter effects were done using ecasound. I wrote a separate
chainsetup for the filtering effects. During processing,
I controlled the resonating lowpass filter with my synth's
Then I start ecasound and load the 'mix-to-soundcard'
chainsetup. Using amplify and pan effects I try to
simulate a tradional mixing environment. Once I find
a good balance, I write the whole thing into a 2-track
That was it. Sometimes I spend more time playing
around with ecasound, while sometimes I do almost everything
with outboard gear and use ecasound just for recording.
> Have you looked at the Linux soundforge clone yet? I think
> it's called waveforge - I haven't tried it myself.
> Is ecawave close to being able to normalise tracks yet?
Actually even the console mode user-interface allows you to change
effect parameter values during processing. If you run out of cpu
power, you can always mix non-realtime. Ecasound's controller system
is also quite flexible. You can assign a controller (oscillator,
linear envelope, etc) to any effect parameter. For instance, you can
use MIDI-CCs as controller sources. This way you can control the
mix session with various MIDI-gadgets.
Now I usually have to boot to Windows and use SoundForge for
finalization (volume normalization and possibly eq fixes).
Hopefully I can soon avoid this phase altogether and use
Linux for everything (this is why I started to work on ecawave).
> What sort of minimum machine requirments are we talking about
> for ecasound to do say, 8 tracks with heavy "chaining"
> of effects on 4 tracks?
Well, I've tried practically all linux sound editors out there.
There are many promising projects, but snd seems to be the only
one that is ready for serious use. I guess these new projects aim
a bit too high. There're SoundForge clones and GIMPs of audio.
To be honest, they're competing with some awesome pieces of
software. SoundForge is the best GUI application I've ever used
(whatever the platform), and GIMP is another amazing project. There
still aren't many succesful Linux projects that aren't network
related. Anyway, ecawave doesn't do much yet, but it works. My
aim is to maximize code reuse between ecasound interfaces. For
instance, ecawave itself doesn't do any audio processing. But
although this makes development easier, it'll still take a lot
of work, before I can start comparing ecawave to editors like
SoundForge. I try to be realistic.
> You have a link to the free music site at your music page.
> How well do you think this idea sits with musicians who hope to
> make a living by what they do or do you think there might be a
> future where musicians get paid for live gigs but not recordings?
> (once Big Music is broke and can't afford lawyers :)
This is really hard to say. First I must admit that ecasound
performance is not optimal. I'm interested in high-level programming
and design methods, and ecasound has always been my test bench
for new ideas. I prefer simple solutions and try to avoid unnecessary
hacks. The resulting code is easy to maintain and more stable.
But back to your question, with my P166 box I can mix about 14-16
mono tracks in realtime. But I'd say that no machine is too slow or
too fast for ecasound. If you don't have enough hardware resources,
you can do separate submixes (like bouncing with analog 4-tracks),
apply effects separately (to free cpu-power), or if nothing else
helps, you can do everything non-realtime.
> Bulky Material ... nice title :)
> Thanks Kai.
Yep, this a hot topic right now. Like with software, I think
it's important to make a distinction between 'free' and 'gratis'.
Free music is more about freedom to create, copy and distribute.
For instance, I'm about to release my 2nd demo-cd, Bulky Material.
Just like with my first demo-cd, all the tracks are freely
available on my web site. But this doesn't stop me from selling
the cd version. Similarly, distribution of mp3s is not free. Someone
has to pay for the web space and the net bandwidth. But this is
not necessarily a problem, because when more people start downloading
music from the net, mp3 distribution sites will attract more
advertisers. And of course, mp3-sites will then be interested in
talented artists and will even pay money for them. In fact, mp3.com
is already doing this! So whatever happens, money will always be
involved. And if you are good, you'll get your money, one way
Finland Resource Map