by Eric Keller
December 1, 2017, updated May. 11, 2018
(This article also appears on the support page of www.synthfont.com/TutorialX.html/TutorialX.html.)
I am an apprentice composer and I use SynthFont2 together with the EastWest music database in order to get my initial compositions into public view. I want people to listen to my compositions, even download the audio files, the midi files and the music sheets. All this is done on a single computer at a reasonable cost.
If you have the time, please have a look around this website and see what I consider essential for promoting my initial steps in music composition. All that you can see on this site was created on just a single, fairly powerful computer. While I do have a MIDI piano, in theory you could even do your sound inputting entirely on-screen.
Some composers do not need a studio – at least when they start up. They compose their music in a music composition program (like the freeware MuseScore2 program), and then they build a credible music output with a high-quality music synthesis program. All this can now be done at a very low cost – once you have your initial, fast computer. This is a high-efficiency approach to creating a set of initial music pieces that are needed to break into the tough music world.
Doing it alone
Can you or should you do it all alone? Before delving in, let's just check out this question.
The first issue is the level of quality required for your project. Recordings that should go on a label for commercial distribution require exceptional sound mastering. That level of quality is simply out of consideration here. In this article we concentrate on excellent quality recordings that you might put on your personal website, submit to your teacher, or send to a potential employer.
The second consideration is the computing equipment at your disposal. Speaking for Windows here, we will require for the solution proposed below, an internal acoustic line built into the computer called "Digital Video Recorder" or "Line In" / "Line out" -- or another means of recording internal sound from the computer. It so happens that the "Line in" solution has become elusive on Windows 10. I've seen smaller and lighter Intel computers that do not have this capability. So unless you organize some other means of recording the internal sound from Windows 10, the solution proposed below will simply not work. (Please check your computer before reading much further.)
Here is an article that deals specifically with this issue: http://www.betterhostreview.com/record-internal-sound-windows-10.html. Solutions at a reasonable cost are offered by the Aiseesoft Screen Recorder or the Replay audio programs from the Applian Technologies group.
And finally comes your level of technical wizardry. The solution below requires a bit of technical application. Recording and distributing your work has traditionally been done by studios, agencies and record companies. In order to make money in a diminishing record market, these places are now concentrating on musicians that can give them “big sales world wide”. Individual musicians or composers that do not fit into their mould currently find it very hard to break into a commercially interesting scene.
But there are some of us creating music in low-budget modes, and we also want to be heard.
The alternative to the “studios” is to do it all by yourself. This can be done, but it is not always easy. If you find writing music is hard, getting it “published” online in a concrete, visible manner may well prove to be yet harder. In my experience, only persons equipped to withstand long hours of banging their head against the wall are likely to succeed in going the path entirely alone. All along, technical incompatibilities can cause frustrating and often longish delays.
I have done it that way, and this is why I am writing this introductory tutorial. If you've looked around this website you understand why. My music is not at all in the popular mode, I am not overly rich, and hardly anyone knows my work. I simply had no choice.
Here is what you need:
- You need a fast and powerful computer with "Digital Video Recorder" (or some other internal sound recording) capacity to do this work. Believe me, it really needs to be fast, have lots of memory and be connected to big disks. Also you need two or three screens to work comfortably. (I have a third, portable USB screen I got for some 200 EUR). This is one place where you will need to spend some money. You'll be glad you did.
- Then you need the following programs. I'll tell you what I use, but cheers to you if you can find some other programs that can do the same or more. If so, do share. I use:
- MuseScore2 for writing my compositions (free). It has everything that a beginner composer needs. I am very happy with it. https://musescore.org
- SynthFont2. Not entirely free but one of the best long-term investments ever. A very powerful and sophisticated program from Kenneth Rundt. http://www.synthfont.com.
- loopMIDI. Microsoft has taken internal midi processing out of the last few generations of Windows platforms. Thank goodness, there has been Tobias Erichsen who has dedicated years of free programming for providing one of the best possible solutions to this huge disgrace by Microsoft. Get this free program from https://www.tobias-erichsen.de/software/loopmidi.html.
- The way I understand it, you also need to install virtualMIDI to make the loopMidi program work. This free program is maintained by Claudio Nicola, another dedicated programmer that we are very grateful to. Find it here: http://coolsoft.altervista.org/en/virtualmidisynth.
- Then you need a music synthesizer which goes way beyond a soundfont in terms of quality and variety of instruments. Here you have a certain choice. A few years ago, I chose the Garritan database as the best bang-for-the-buck program. It was a good choice at the time, but the database itself as well as the control program had their limitations. In 2016, a large portion of the EastWest database has suddenly become available to the world with their $25-per month subscription program, which makes it $300 a year. (Ok, in Europe they will charge you 25 Euros per month). EastWest is the big player on the block. At 300 somethings per year, you've got access to the most sophisticated music synthesis program under the sun. After having worked for a few years with lesser offerings, there is no question in my mind that this was a deal that I could not avoid. You can get signed up here: http://www.soundsonline.com. Look for “composercloud”.
- Finally, you need good and trusty Audacity, your free audio processing program from https://sourceforge.net/projects/audacity/files/.
Let's put it together
My basic chain is this:
Stage 1: Composing with MuseScore2
- I write my compositions with MuseScore2. The best all-around soundfont has been FluidR3-GM2-2.sf2. I've found that the piano in that soundfont has a very clear tone that lets me hear my chords as well as well as my errors. For pieces destined just for the piano, I use a soundfont with the superb-sounding Salamander piano (either Nice-Keys-Ultimate-V2.3.sf2 or SGM-v2.01-Sal-Guit-Bass-V1.3.sf2). At the end, I print out the score as a pdf and as an MP3 and discuss the result with my teacher.
- To save and prepare the composition for the website, I set up the following chain: MuseScore2 → SynthFont2 → EW Play (EastWest) → Audacity.
- MuseScore2 → SynthFont2 is easy. Simply save your work as a standard MIDI file and read it into SynthFont2.
Stage 2: MIDI Control with SynthFont2
For the SynthFont2 → EW Play link, you need to install loopMIDI and virtualMIDI. Restart your computer after the installation. You may wish to reserve 3-4 MIDI channels at computer start-up, but don't worry, you can always change it later with the loopMIDI task tool. From now on, your computer will maintain these MIDI channels.
You may wonder whether you need to change the MuseScore2 instruments for outputting at this stage. The answer is basically no. Obviously you need to respect the tone limits of your intended instruments as you compose, but the final instrument choice will only be made at the EW Play stage. For example, you can do all of your composing with the MuseScore2 pianos (which sound a lot better than the violins), and end up using violins at the EastWest Play stage.
Load your midi file into SynthFont2. You can play it in the usual way using a soundfont. To channel the output to the EastWest Play program, select your midi channels with the “MIDI Out” button on the task bar. You will see that the loopMIDI ports are available.
Once you've made your assignments, save your setting as an “Arrangement” in the File menu. The SynthFont2 program will remember the place where your midi file resides as well as your settings.
Stage 3: Virtual Instruments with EastWest Play
- For the next part, I shall suppose that you have subscribed to EW composercloud, and that you've downloaded at least one instrument in question. I personally went the whole hog and downloaded the most immediately important databases, which were the four pianos (Bechstein, Bösendorfer, Steinway and Yamaha), the EW Symphonic Choirs Gold and the EW Symphonic Orchestra Gold Complete, plus a few other goodies. It was enough to stop the smile: more than 500 Gb on my 4 TB disk...
- Now open the EW Play program. Open the Settings (ratchet wheel) menu. The MIDI tab shows the loopMIDI ports. That is where the sound will feed into the EW Play program. If the MIDI channels don't show up, there is typically one of two problems. First, it may be that you haven't used the loopMIDI ports yet that day. When you open the loopMIDI program, that problem will be solved. Second, the MIDI channel setting in EW Play may be set to the frequently used "Omni (MIDI channel 0)" setting. Go to the "Other" tab, and select "Automatic increment" as your MIDI channel assignment setting. Now the loopMIDI channels should show up on the MIDI settings tab.
- Under the Audio tab, you will find the usual Windows audio settings. I recommend using 48 kHz, because with 41 kHz an unpleasant background noise might be generated that comes from the conversions between 48 and 41 kHz. Leave the Device Type on DirectSound and for initial testing purposes, specify your normal speaker as an OutputDevice.
- In the same tab, the ASIO option also permits you to establish a direct link with the sound output system. For years, this was the preferred solution. However on two recent I7-based computers, the ASIO link was notably slower (introduced more delay) than the DirectSound option. So your speed may vary. A description of setting up the ASIO link is given in my introduction to the SynthFont2 programme.
- Also to be sure, use a fast 64-bit Windows computer. All recent work on the EW database has been in 64 bits. Work with less recent Windows 10 databases and fewer parameter modifications is possible with 32-bit computers using EW Play 4. In my case, I use EW Play 6 on an Intel i7 with 16 GB installed RAM running at 2.9 GHz. Also, a smaller travel computer with an Intel i7 and only 8 GB running at 2.9 GHz can handle smaller jobs correctly (I've used it with 4 voices).
- Now choose your instruments – carefully. The manual will be a good guide because a perfectly good instrument can give you awful results if you use the wrong database choice with inappropriate settings. As an example, the Goodnight Hymn on my website was created with the instruments shown in the Play 5 illustration below. Quite a few other settings gave unsatisfactory results.
- Notice four midi channels, each for a different instrument or voice.
- Now click the arrow in SynthFont2, right above the wave display. You should hear the synthesized sound through the EW Play program. This is where you will appreciate having a fast computer. On current mid-speed to slow computers, you might get some unavoidable hickups in sound production.
- Study and explore the various options on the Player page. You can for example add some reverb and shape the note response curve for excellent effects. You may also wish to lower the voice limit and the velocity to reduce the dynamics of your production.
Stage 4: Record and Use Effects with Audacity
- You got that? Perfect. Now we want to save the music to a file. For that, launch your internal sound recording system.
- If your computer has the Digital Video Recorder audio line, launch Audacity and select “Digital Video Recorder” as your input audio channel. Also in EW Play, change the Audio setting Output Device to “Digital Video Recorder”. The audio from EW Play will now be transmitted in real time (streamed) to Audacity via that channel.
- The price to pay: you won't hear a thing while the sound is being recorded. But that's no great loss, because you'll see the cursor move in SoundFont2 and you'll see the sound arrive in Audacity.
- And voilà. In Audacity, you can apply any of the further effects you need: amplification, fade-in/out, etc. Finally, you simply save the file in one of the many ways that Audacity offers.
- If you use another internal recording device, you can usually leave the EW Play output on its regular output channel, which will generally let you hear the sound while recording.
The EW VST
When you install the EW Play, it also deposits a 64-bit VST in this location: C:\Program Files\vstplugins. It's called play_VST_x64.dll. Wouldn't it be a great idea to load this into SynthFont2 or into some other Digital Audio Workstation (DAW) program like Ableton Live? This would avoid all the internal re-cabling, and it would let us work with a single, powerful DAW program. The EW VST has all the features of the EW Play program, minus some input-output capacity.
Sounds good, but reality has taught me otherwise. First SynthFont2 can currently only work with 32-bit VSTs, so the SynthFont2 idea is excluded as a VST host for the moment.
Then, my own experiences with plugging the EW VST into Ableton Live 9 (which supports 64-bit VSTs) have been consistently disappointing. With the help of support at EastWest, I was able to get their VST working within Ableton Live, but Ableton's input-output support of the VST is very limited: you have to set up the EW VST before loading in the midi file and then you can no longer change the setup parameters while the audio is playing.
Also, Ableton Live is not only relatively costly, but also rather complex. It took quite a bit of time to learn to set up everything correctly, with sound going in and coming out in the expected fashion.
Furthermore, the MIDI file handling in Ableton Live is, shall we say, "circumscribed" -- no trace of the explicitness of SynthFont2.
Altogether, using EW inside Ableton Live has been very time-consuming and has not been a very flexible and agreeable experience for me.
The setup I describe here using the standalone Play is not only much cheaper but also much better, because you can change the instrument parameters in EW Play easily without changing the basic setup -- even while the audio is being played. This makes for fast adjustments. There is a mixer and there are loads of parameter adjustments in EW Play, so you can experiment for hours with with different instrumentation for the same composition, till you get the best possible sound. For us picky and cheap all-in-one composers, this is just the perfect setup.
Finally, I asked EW if they were planning to add a MIDI input and an audio output to EW Play, and the answer was negative. They said “Play is just a software built to load up the instruments. We are continually working to improve it, but the focus is more on the functionality of it as a software instrument to be integrated with a DAW for recording/mixing/mastering, rather than trying to do everything. DAWs just have so much more development/staff/time/money invested in that side of things.” (answer received in November 2017).
Absolutely, but then the VST – DAW interface must also be able to support strong interactive work, and so far, I haven't seen that.
Processing Effects Caution
With all the wonderful processing effects that are now available, one is tempted to “perfect” one's creations with compression, high/low filtering and reverberation. However, the expert consensus opinion suggests great caution, particularly when dealing with high tones in flutes, piccolos and voices.
For example, the very high soprano singing found in my “Goodnight Hymn” did not permit any effect processing because of the digital sound resampling performed in the subsequent m4v conversion needed for the video version. Sometimes I've found it useful to “fatten” lower tones of pianos a bit, but I leave high tones unaffected. The EastWest engineers have done a good job of providing high-quality sound directly from their database, even for the hard-to-handle high tones.
For the Lazy all-in-one Composer
And here is a script to launch all the three processing programs in one fell swoop. Save it as a batch program, e.g. as “LaunchSF_EW_AUD.bat”:
cd "C:\Program Files (x86)\SynthFont2\"
cd "C:\ProgramData\East West\"
cd "C:\Program Files (x86)\Audacity\"
I am very pleased with the current situation of all-in-one composing. After years of feeling excluded from the best possible music production possibilities, just because I could not afford certain top-notch programs, the current situation has suddenly changed. As a simple composer, equipped with a few key programs, I can now build truly credible, possibly even impressive pieces of music with the very best virtual orchestras in actual existence. It's a great moment to do some serious composing.