LMP Asks #8: An interview with Florian Bador

Category: 
Level: 
Tools: 

This month we talked to Florian Bador, FLOSS audio enthusiast and founder of music distribution website, Trust Music, that promotes open audio formats.

Hi Florian, thank you for taking the time to do this interview. Where do you live, and what do you do for a living?

I am located in Santa Monica, California.

It depends what you mean by “living”. Spiritually, I have the Music and God, and they both keep me very alive!

Financially I am a freelance computer programmer, I specialize in back-end, database, perl and Linux servers, but I also do some front-end.

You are the founder of the music distribution website, Trust Music. Tell us a bit about that?

First, thank you for getting in touch, I’m very excited that you decided to have a closer look at Trust Music.

I started Trust Music last year in 2014, it was a sudden inspiration. The next day I was already working on it. My motivations are quite unusual since Trust Music is a business project powered by a spiritual idea. But these two things are more connected than we use to think. Business is people and people are spirits.

I can’t really explain Trust Music without explaining my spiritual background.

I was raised atheist in France, but at 23 I moved on my own to Montreal, Canada, where I lived 5 years, and something happened there that would be hard to describe with mere words, that’s why we call this a personal experience. One day I was atheist, the next one a follower of God.

I always used to think that people seeking for God are lost, but I realized that we are lost until we find Him (or until He picks us up and that we don’t reject Him).

When we realize about God, our whole perspective of life is transformed. I started to see God in everyone, even those who are not aware of Him. And from there, I understood where the goodness in people comes from, and how every single human being has in themselves the Universal definition of the word “good” before they learn any earthly language.

Trust Music is entirely based on that.

The concept is that the listener decides the price of the music they download, and this price can be null. They can also come back later to edit the price and increase it. This is a concept of Trust without being a concept of donation because what people do on Trust Music is a purchase, I try to insist on that.

The website was designed to connect the downloader as much as possible with their actions, to make them realize what they are doing when they pay for the music or don’t. If we think about this as guilt it’s not very nice, but if we look deeper and try to reach the source of all that, why do we feel bad when we do bad, and good when we do good, then there’s something beautiful and irrational into it. That’s what Trust Music is really about, and the logo can help grasp this idea.

I truly believe that what motivates us for open-source software is the same thing, whether we are aware of it or not. It’s this Universal Goodness, not learned, unteachable because already known, that connects us all as One and makes us Free.

I wanted to use FLAC and OGG-Vorbis on Trust Music because people need to know and use these fantastic free formats. FLAC is even available in 96KHz 24bit if the music maker uploaded in a such pro format. I always do for my music of course. The servers spend a lot of time encoding everything the best way in so many formats, I carefully chose the encoding options to get the best quality.

I designed and programmed the entire platform, from graphic design, front-end to back-end, payments, the encoding of musics and videos. I even built a custom web daemon for Trust Music so the website itself is an HTTP/HTTPS server and also a database engine. No apache, no MySQL, it was all built from scratch. This makes everything extremely fast, most dynamic pages respond in 2 to 8ms, the rest is just network latency. It might be considered as over-engineering but I’m the one paying the bill for the hosting and the difference in cost is huge.

Are CC licences encouraged on Trust Music?

Music makers can add credits and thanks on the page of each music, so if there is any specific license, they can specify it there.

What's the difference between Trust Music and other music distribution sites such as Bandcamp?

First, Trust Music made “name your price” its core concept. It’s not just an option called “donation”, its the way purchases are made and that’s why it is called Trust Music.

This influenced everything in the website, especially the design that helps connecting the artist and the fan more humanly with e.g. a making-of video for each music and almost no text, so seeing eye to eye is the only way to connect.

Most of the website is visually simple, I wanted only features that truly help the artist and nothing to distract the fan from the goal: buying the music, sharing it with the world, and feeling connected to the artist. All these gadgets on other platforms are nice and appreciated, but they distract too much so it does not serve the artist in the end. You can spend hours browsing Bandcamp or SoundCloud without actually doing anything that really matters.

On Trust Music it’s the opposite, there is nothing else to do than what truly matters.

The listener can also edit the price retroactively after download to increase it. For example if you paid $1 you can click a week later the download link you received by email and edit that cart. If you change to $5 you will pay $4 and the statistics are updated in the artist’s account. You can come back anytime to edit this cart multiple times, and we remind you to do that if you entered zero as the price.

There is a second very important and unique part of the concept of Trust Music: the referrals. You cannot join Trust Music unless you are referred by an artist. This artist will get 7.5% of whatever you sell so they want to invite you.

Once you are invited you can invite other artists too and you will get 7.5% of their music.

Now, where it gets interesting is that the artist you invite will help support the artist who invited you. The same happens for you with artists that your guests invite. And it goes even further...

This referral concept is calculated like this:

  • A1 refers A2 who refers A3 and so on.
  • A8 generates $100 of music sales. A8 gets $70, Trust Music gets $15, and $15 is shared among A8’s successive referrers.
  • A7 gets half of it ($7.5), A6 half of that ($3.75), and so on until A2 who gets about $0.12.
  • A1 does not get anything because we stop at 7 deep so there is no big boss on the top, everyone has the same chance of referring a great family who support each other.

The family you referred can be quite large, for example if you and your guests all refer just 3 artists you receive an income from 37 artists plus yourself, so 2,188 music makers. If one of them is big, that can be significant.

In the end, since Trust Music only gets 15% you still get more from your own music than you would with iTunes who takes about 35%.

You get 70% instead of 65%, plus the referral income from all your successive guests up to 7 deep.

The third important difference is that Trust Music focuses on quality, of sound and of music.

It’s very hard for professionals to browse a website like Bandcamp because there is everything from the 10 year old kid to the touring band. On Trust Music, there are only artists that someone wanted to refer because they believed in them.

Concerning the sound quality the formats are amazing, 96KHz 24bit is far beyond any other platform. The focus is clearly professional, no junk on Trust Music! :)

Can you tell us a bit about any other projects you are involved in?

Besides working on Trust Music and computer programming, I am also an orchestral composer. It definitely does not pay the start-up bill of Trust Music for now, but it pays the one of the soul.

I am currently working on a music video project that you should see by the end of the year, a quite unusual thing that I will post on my Trust Music profile page.

I also train physically a lot and I do time trial cycling. I am a strict vegan which helps a lot for performance. You would be surprised to see how much a vegan diet has in common with the open-source, and that’s not even a joke. They are both about compiling everything from the source!

What is your musical background?

I started to play a tiny synthesizer when I was 2. My Mum realized I was spending most of my time on it, so she somehow told Santa who brought me a better one, and so on. Years later she regretted it because I never really opened a book from school. I was leaving the school bag at the door, was running to my bedroom to make music and I would pick the same bag the next day without opening it. My bedroom was more like a music studio with a bed as bass trap.


Florian's bedroom studio when he was a kid

Being a teacher herself she got worried but some people learn better by personal experience than with a teacher, and the sooner we realize what type of learner we are, the less time we waste.

So for music just like programming and everything else I am self-taught. My composition companion is the piano.

I composed for a few short films and documentaries but I’m not sure I want to go in the movie industry today... composing two hours of music in three weeks, getting your music over-edited so it ends up sounding like a mere background sound-design, all this seems to kill the beautiful spirit of Music. I guess it depends on the director, but for now although my style sounds typically film music, I want to stay free to make innocent Music, as innocent as a five year old child, just the way it should be.

What is your history with Linux?

Around 14 I had my first computer. A cousin custom-built it for me and I looked at him carefully to understand how it works from within. It was an extremely powerful 433MHz Celeron with 64MB of RAM and this colossal 17.3GB hard drive. I put Windows on it and started to host websites from my bedroom. I had to connect a DB-25 pin from the parallel port to the reset switch because the system would get heavier with time. I built a Visual Basic program to reset every couple hours so I could go to school.

After a while I got tired of this so I started to learn Linux and then magic happened. I discovered the real power of a hardware machine and I could finally use it fully.

Once again, only a personal experience can make us realize that. It's hard to explain to a Windows or Mac user how much computer power they waste by buying that instead of free and open-source software, they need to live it to understand.

What is your typical workflow when making music?

My setup is a little complex for most musicians, because they are not necessary programmers, but it’s really worth it in the end.

First, I use only custom built PC computers and I run Linux Slackware64 on them. I optimize the system by recompiling the kernel and the most important libraries and programs. I’m too perfectionist to use the original binaries.

I use 4 computers for music:

  • Main: runs the window manager where all programs show up (via ssh -Y). This means all other machines have no graphic interface and no monitor. That way, I am never afraid of troubling audio processing with graphics. I can open Firefox, multimedia or office programs in the same screen space, it will not disturb the music production at all.
  • Audio: runs the DAW
  • Sampler: runs the GIG libraries
  • Reverb: runs 8 convolution reverbs


Florian's multi computer set up

I use Ardour as my main DAW, I run a development version (from git) so I always have the latest features. In Ardour is the MIDI editing with a lot of MIDI-CC automation curves, especially for the expression of the strings. It’s really in these curves that I get a natural sounding violin or cello. Of course you need to know how a real orchestra is supposed to sound, so listening to a lot of great music helps.


Click on image for Florian's full sized dual monitor set up

That MIDI is sent via Netjack as 24 MIDI ports to the sampler machine which runs LinuxSampler and has no sound card.

LinuxSampler sends that back to Ardour via NetJack as 64 audio channels at 96KHZ 32bit float with only 5ms of latency.

In Ardour, there's an insert on every track so I can use analog gears in a single click by enabling the insert.

Then there are 8 channels of reverb sends that output in ADAT 24/96 (RME RayDat) to the reverb machine and come back the same way as returns to be mixed.

I use a Mackie Control with Ardour and mix from there.

What’s very handy is that I can still change a note even at the very last stage because everything is real-time in a single chain.

What sample libraries do you use?

I always have a huge amount of banks preloaded in RAM on the sampler machine. Then I pick the best for each instrument. I don’t even look at the names, I just switch from MIDI channel to channel and ports until I find that piano that sounds like I wanted for this music.

It is rarely the same brand because they all have their strengths and weaknesses. The VSL Epic Horns e.g. are unbeatable, but their strings are possibly the worst I’ve ever heard. In the other hand, I really love the strings of Sonic Implants.

But there’s always a lot of EQ after selecting the instruments.

How do you manage sessions? Do you use a session manager and how does this work across computers?

I spent a lot of time making templates at the beginning.

I have a LinuxSampler template loaded at boot and since I never reboot this machine I don’t have anything to do. For example, it has 3 months of uptime so far. Since LinuxSampler itself is a daemon without GUI and that it has this awesome network protocol to control it, all it takes is really a boot script. It can be totally detached without a single window. I love this idea, this is beautiful.

I also have a template for Ardour and that is where the mixer settings and all jack connections are stored. It has a lot of MIDI and audio tracks ready for each instrument, I just have to disable the ones I don’t use when I start a new music. The .ardour file is about 1MB so that’s a pretty large template.

Where I spend a lot of time is no longer in setup but in the editing of the MIDI controllers expression curves, especially for the strings. When you have 2 voices of violins, 1 of violas, 2 of cellos and one of basses, every single string chord is a lot of work on each of these 6 notes. Being very careful with these curves is IMO the only way to get a great sounding string ensemble. In fact, if you do it well, even a real orchestra of great students will not sound as good, only a pro and expensive orchestra can beat it.


Florian's not so modest connection matrix in qjackctl

Tell us a bit about your hardware set up

Nothing crazy there because when the software is good it’s amazing what you can achieve with a simple machine.

  • Main: Core2 duo E8500 3.15GHz - 6GB DDR2 800
  • Audio: Core i3-4150 3.5GHz – 16GB DDR3 2400 – 120GB SSD
  • Sampler: Pentium G3240 3.1GHz – 32GB DDR3 1600 – 2x 1TB regular WD HDD.
  • Reverb: Pentium E5700 3.0GHz – 4GB DDR2 800

I use very little analog gear for audio because although they would dramatically improve the quality of my mixes, they can be extremely expensive and the return on investment is very little, if it even exists. Just a couple basic Tube compressors / preamps.

Why do you feel open source is important, and what for you is the most important aspect of Linux audio?

To be honest, even if we forget about the cause aspect of open-source for a second, just the performance is astonishing.

With all the tracks and buses, my DAW has 120 channels at 96KHz / 32bit float with only 5ms of latency on a machine that is totally affordable, and it sends and gets back 64 of these channels over the network. There is simply no way any licensed setup could to that today. Those who do it simply buy more expensive hardware, which is totally absurd knowing that they already pay for software that makes their machine weak.

So to me the performance is at least as important as the cause. All technically advanced people know that the most powerful software in the world is open-source, and there are two reasons for that:

1- Open-source can be compiled natively so each binary is specifically optimized for each machine. And of course, if you just get the binary of a distribution you miss that important point.

2- Instead of a little team of programmers that would quit their job if the salary wasn’t great, you have the entire world improving open-source projects every day, including the week-ends, by passion. They tweak this on Sunday night because someone else suggested it on Saturday from the opposite side of the globe. Nothing can beat that. In fact, you have Love VS money here. No company can reach that level of passion and drive, no matter which bikes and saunas they put on the campus, they cannot buy passion, a form of Love. I think that’s where the open-source gets its roots and strength.

From there comes the cause but these two aspects, performance and cause are really part of a whole: it’s the result of connecting people as one. And we almost get back to Trust Music here :)

Of course there’s also the responsiveness of the developers, being able to communicate with them. I often discuss with Ardour developers about new features they could implement, help them to debug something. It’s awesome to suggest them a feature they find great, and a couple hours later you download their new git, recompile Ardour and the feature is there! If you had to pay Steinberg, Avid or Microsoft to tweak their licensed software for you, that would be extremely expensive.

What do you feel is currently lacking in Linux audio?

To me, the ability to run the reverb Waves IR1, because I’ve never heard anything as close to that. The impulse responses they made are really impressive. I don’t know if that is just the impulses or if there is also something special in the plugin, but nothing can beat that. If it’s all in their impulses, unfortunately their .wir format seems encrypted or encoded in a very proprietary format. But the safest would be to run that plugin on Linux.

Waves started a project using Ardour a couple years ago, so maybe there is a bit of hope that one day IR1 will run on Linux.

I don’t see anything else missing. I just think people should encourage Ardour in their MIDI development because this step they made is giant for the entire Linux audio world. If I use Ardour today, it’s not only because I don’t want to buy Cubase, it’s because I truly believe it is the best technology for my needs. Let’s not forget there was a time when I used Cubase, and then I switched to Linux and used Ardour.

What is your favourite FLOSS plugin?

I have an unflavored, wax-free vegan plugin. It’s plant based and I use it every night before I go to bed. But besides this one, I like the 4 band parametric EQ by Nedko Arnaudov, easy and simple to quickly EQ the strings.

Are there any FLOSS projects that you are excited about at the moment?

Broccolis can be very exciting as FLOSS projects, they always find a way to compile themselves natively so that they are perfectly optimized for the architecture and the environment variables.

But besides broccolis, Ardour and Blender are two fantastic pieces of software. I know Ardour well and I love the direction it is taking with MIDI. I recently learned Blender, and of course it’s not an easy software to learn... but I now understand why they claim it’s worth the investment.

I also hope Linux Sampler will stay alive, because nothing would be possible for an orchestral composer without it, and since this is a quite specific tool, not that many people must be using it.

What changes, if any, would you like to see within the Linux Audio community?

The community itself is doing great I think. It’s hard to ask a lot from low budget.

I would say more highly successful musicians realizing what they miss. But they don’t have time for these little things. So they hire someone they know because trust is more important than technical skills for them, and this someone usually just purchases more expensive hardware to compensate for the software power.

I’m more worried about the audio hardware, especially that one day MIDI would disappear and be replaced by some dumb USB. That would be very sad because MIDI is IMO by far the greatest invention of Music production.

We loved it for over 30 years, let’s not kill it today. It’s that sort of very standard protocols that allows kids to learn and geeks to tweak anything they want. This is why something like the Raspberry Pi is so wonderful.

What advise would you give to a new Linux Audio user?

Audio or not, go for quality work. You don’t choose Linux to get an Ubuntu and go into quick fixes. The quick fix spirit is for Windows and Mac users. Linux is the technical excellence, and we don’t kinda choose excellence, if we go for excellence we take all of it.

So to me, using the distribution packages to simply copy some Linux binaries does not make any sense. I would say choose Slackware and learn the real Linux, compile things from scratch perfectly. It will take a lot more time but at least you stay in control of what you do and can understand it. The quick fix will only give the illusion to progress and learn something, until you have a problem and have no idea how to fix it.

I tried both approaches and I don’t regret my choice. Then just make a good backup! :)