Tag Archives: digital audio

Ardour 4.0 Released

Great news, ardour 4.0 is here:

Read more:


some more information in the great libregraphicsworld site:


Tagged , , ,

Ardour 3.0 released !!!

Finally it arrived! The long awaited, award winning DAW for GNU/linux reaches a new historical milestone. The major new feature of this version is the introduction of MIDI track recording and editing, but there are LOTS of new features in other departments too.

Also the website has changed a lot, and for the better:

ardour new website

ardour new website











Check out the announcement and the “what’s new” page:



Big kudos to Paul Davis and all the ardour crew for this wonderful product.


Tagged , , , , , ,

Linux digital audio editors roundup [part one]



Today I came across this list of “25 free digital audio editors” (and that definition has been stretched a lot here) where I’ve been pleased to find many of our beloved linux applications.

Some don’t even fall in the category, in my opinion: for example I would define hydrogen as a drum machine, and rosegarden a notation editor or a MIDI sequencer, like qtractor.

I take the chance to resume the current situation in GNU/Linux about digital audio editors, which is not a bed of roses compared to other systems/proprietary applications, there are some corners that can be smoothed. Some editors however can be used in real life with good results.

First of all, I leave ardour out because it is much more than a simple audio editor, it’s a full fledged DAW. It can for sure be used as a simple editor but for the task is overkill. For one, you have to create a new project before you can even think of opening an audio file.


One of the most known and succesful editors in the scene is audacity, an almost complete solution that is cross platform, easy to use, actively mantained and not too buggy.  There are lots of documentation and tutorials around for using it, coming from the huge user base that audacity has, thanks in part to the cross-platform nature of the program.

The problem with audacity is that it plays not so well with JACK, the low latency sound server that is almost a must if you want to make music professionally on Linux. JACK is supported, but through portaudio, a middleware library that makes things a little confusing on JACK side: for example it creates a JACK client on the fly when the user presses “play” on the interface of audacity, and then destroy it when the user presses “stop”. This almost always lead to problems and unwanted clicks (xruns, in JACK lingo) that make audacity not the best choice in a setup that relies heavily on JACK.

Instead, audacity plays well in a desktop system (without low latency audio) and it’s quite useful indeed. It plays nicely with pulseaudio and would be the perfect choice for the user willing to make podcasts, simple audio cuts, simple voice-overs (audacity supports multitrack, to some extent), audio cleaning, and so on.


One of the audio editors that performs better in a JACK environment  is mhwaveedit (not listed among the “25”),  a one person effort that is much less advanced than audacity and has a far smaller user base but is stable, useable, actively mantained and performs the basic editing operations quite well.

In mhwaveedit many functions of the program are bound to shortcuts, to speed up common operations such as navigation into the wave file. Horizontal zoom in/out is conveniently bound to the mouse wheel.

Integrated effects are not many: only basic facilities like fade in/fade outs, normalization, and sample rate/format conversion. A lot of other effects are available through the “effects window” where you can find a host of LADSPA plugins and sox algorithms in a long list (not very user friendly for a newbie, I admit) that you can organize to your liking, e.g. put your favourites on top.

Audacity and mhwaveedit are so far (in my opinion) the most reasonable choices for an audio editor on Linux, next time I will explore other (more exotic or unmantained) options for destructive, simple audio editing.

Tagged , , ,

Cavaco vou levar – another track made with GNU/Linux

The new album cover

The new album cover

Here you can listen to a sneak preview of th 6th track of the upcoming album “Il capitan Pepito Amarilla e altre storie samba” by Orchestra Rosichino , an italian band where I play guitar that does a strange genre that we call “spaghetti samba”, because we would like it to sound like samba… but with a touch of spaghetti western.

The album backing tracks have been recorded onto analog tape with a tascam TSR-8 recorder. It is one of the latest analog recorders produced; it records eight tracks into half inch tape.

With it I recorded all the band toghether (drums, bass, fender rhodes piano, electric guitar and cavaquinho) into a small room with some acoustic treatment. I used three microphones on the drums (kick and two “overheads”) and one mic for every other instrument.

Playing toghether greatly improves the groove. In the previous album we recorded every instrument alone and while this approach gives the best separation, the interplay between musicians is negatively affected. This time we choose the opposite and played almost everything (except voices and trumpets) toghether. I think the result is far superior in terms of groove.

Then I trasferred all the analog tracks directy from the recorder to the ardour DAW using the excellent RME hammerfall/multiface soundcard. The tracks sounded already very good with no effects or processing, but I added some compression and eq (invada mono compressor and the LADSPA triple band parametric eq).

At this point we recorded the trumpets (two trumpets with a ribbon EHX  microphone, for the most part) and then the main vocal part, with a sennheiser MD441 through a TC helicon Voiceworks processor. Finally we added some backing vocals (again, two voices into a single microphone).

The mixing has been done all inside ardour, and we exchanged some three versions via internet until we found the good balance that everyone liked.

Mastering has been done with mhwaveedit recording the L+R sum which passed throug a FMR Really Nice Compressor.

Here is the final result, all the process has been very quick; we recorded, mixed and mastered 7 tracks in little more than two weeks:

Hope you like it.

Tagged , , , , , , , ,

Some python code for calculating BPM and delay times

Some time ago I wrote this little script (I called it beatcalculator).

Pretty straightforward command line utility, use the switch –help to get help.

Here is the script, exported from emacs with htmlize-buffer: beatcalculator.

Copy and paste this code into a file called beatcalculator, put that file in your $PATH and give it execution privileges with chmod +x.

Then you can use it on the command line as simple calculator, providing the bpm you want to examine, or you can ask to analyze a wave file. you have to specify the number of bars contained in the wav file and optionally the meter (defaults to 4/4)

Here’s an example of the output you get (with a file with 2 bars of music):

$ beatcalculator -f afroeira123.6.wav -B 2

Meter is 4/4

afroeira123.6.wav is 3883 milliseconds
Bpm of afroeira123.6.wav is 123.62

123.62 beats per minute (bpm):

Note                            Delay time                       LFO freq
Whole                          1941.433 ms                        0.52 Hz
Dotted Half                    1456.075 ms                        0.69 Hz
Half                            970.717 ms                        1.03 Hz
Dotted Quarter                  728.038 ms                        1.37 Hz
Quarter                         485.358 ms                        2.06 Hz
Dotted Eight                    364.019 ms                        2.75 Hz
Quarter Triplet                 323.579 ms                        3.09 Hz
Eight                           242.679 ms                        4.12 Hz
Dotted Sixteenth                182.009 ms                        5.49 Hz
Eight Triplet                    161.78 ms                        6.18 Hz
Sixteenth                        121.34 ms                        8.24 Hz
Dotted Thirty Second             91.005 ms                       10.99 Hz
Sixteenth Triplet                  80.9 ms                       12.36 Hz
Thirty Second                     60.67 ms                       16.48 Hz

Hope you enjoy it.
Send me comments and suggestions to improve it.

Tagged , , , ,

Ardour 2.8.13 released

Ardour 2.8.13 released

New maintenance version of the stable branch of ardour, our beloved linux DAW

Tagged , , , ,

Track recorded with ardour

Sbanda Ballet CD cover

This is a traditional song from Salento, performed by Sbanda Ballet, a band from Bologna that plays lot of folk tunes from all over the world (Ireland, Romania, Italy, Israel, France, Bretagne, etc..). Their main focus is making people dance.

This song is part of a CD called “Stradelli Guelfi” and was recorded in June 2011 at Cricketstudios, Bologna (my project studios), with open source software (ardour on ubuntu).

Recording was made “live in studio” with no overdubbing. Microphones used are sennheiser md441 and md421, and oktava mk012.

The cd features 17 tracks total.


Tagged , , , , , , , , ,

New Balafon soundfont

A friend of mine lent me a fantastic african instrument: a Balafon.

I took the opportunity to sample it and make a soundfont to be used in GNU/Linux with qsynth.

The Balfon from which I take the samples for the soundfont


To record it I used my trusty yamaha pockettrack 2G (best buy of 2011), I then cropped the single notes with mhwaveedit and put toghether the soundfont with swami.

I don’t have the latest version of swami and there are some oddities in it, but at least I’ve been able to make it into a usable soundfont (I hope).

Please feel free to use it under the term of the CC BY-SA 3.0 license ( with the following exception:
The “share alike” condition is imperative if you modify the samples themselve or create new sample libraries with this licensed product. However, produced music and other non-sample-library works can be licensed at will).

The instrument itself is gorgeous, too bad I have to gave him back soon… I only sampled one layer for each note, basically there are three octaves of a C scale from G2 to F5 (I don’t know for shure those are the actual pitches, maybe I missed an octave in positioning notes).

The whole process took me about 2 hours to complete.

Hope you enjoy it!!

Click here for the soundfont balafon.sf2

Here’s some fiddling with the soundfont, recorded again into mhwaveedit:

Tagged , , , , , , , , ,
%d bloggers like this: