How to make soundfont

How to make soundfont

Before starting this tutorial, be sure to review the «Sample import» and «Oscillator» sections of the Crystal user guide. The former will explain how to add soundfonts to Crystal and the latter will explain how to use them in a patch.

Introduction

Multisamples means you use multiple samples to provide the sound for an instrument. Why multiple? Because a single sample, such as you find in a WAV or AIFF is generally insufficient to provide a high-quality sound across the entire keyboard. With a single sample, a sound which sounds good on one part of the keyboard may sound horrible on another part. Further, using a single sample leads the rise of aliasing, a nasty sound which occurs when you play a single sample too far from its original range.

In contrast to single samples, multisamples provide multiple samples across the keyboard, helping to avoid the problems associated with single samples. Soundfonts provide the ability to map multiple samples across the keyboard, or even across velocity levels so that you get a different sample depending on how hard you strike the key.

Soundfonts are a very flexible format and can include many things, including descriptions of multiple instruments which may be layered together, envelopes, effects, and much else. Crystal only uses soundfonts as a source of waveforms, so this additional information is ignored by Crystal. If you have a soundfont with these extra bits and you’d like to hear the full soundfont, not just the waveform, try a sampler or dedicated soundfont player like Jeskola.

Start

If you’re beginning with a WAV or AIFF file which is already looped, you can proceed right to step 4. If you’re beginning with WAV or AIFF file which is not looped, start with step 3. If you’re beginning with an audio CD, start with step 2. And, if you’re beginning with a sound which you will record, start with step 1.

Procedure for creating a soundfont

These steps are performed with an audio editor. I’ll provide examples for both mac and windows.

Mac version

If you have a stereo file, convert it to mono by doing a File->Export Dual Mono, then open the Right version of the file you exported. Next, make a mono file for each note by selecting the audio for each note, excluding the silence before and after the note, then Edit->Crop, then File->Save A Copy As, then undo the crop. Repeat that for each note and you should have a mono AIFF file for each note. Then open each note’s AIFF file and, making sure you have «Preference->AutoSnapToZero» selected, choose a region for the loop. Find a beginning point which has a similar waveform appearance as the ending point, then choose Action->Loop This Selection. Try out your loop by playing the audio with looping turned on. A good loop has no audible artifacts when the loop jumps from the end to the beginning. You may have to try various loop points to get a satisfactory result.

Windows version

Open your WAV file which contains all the notes. Convert it to mono, if necessary by using Process->Stereo to Mono. Next, select each note in turn and use File->Save Selection As to save each note to its own WAV file. Next, open each note file and crop it using Process->Crop->Start And End. Choose Settings->Sample and click on «Use detected» to assign base note for this sample. Finally, loop each note using Loop->AutoSearch with the Whole Sample option. Play the WAV with looping turned on to ensure that the loop is indeed seamless :-).

Mac version

You’ll see two windows when you start up Polyphontics. Our first step is to import the AIFF files into the Audio Library. Click on the import button in the Audio Library window and shift-select the note AIFF files to import them.

Next, create a new instrument in the Document Edit window with the «Create New Instrument» button. Give it the name you’d like to see in Crystal’s Oscillator Type menu. Next, add the samples to this instrument with the «Add Audio» button in the Audio Library window.

Next, double click on each sample in the Document Edit window. For each one, enable looping and set the base note to the note that was originally played when this note was performed. Then assign the keyboard zone which will play this sample by clicking and dragging in the keyboard at the top of the window. Click Ok, then move on to the next sample.

The keyboard zones should be non-overlapping, and the zone for each note should be in the region of the original note that was played when it was performed. The result is a mapping across the keyboard with no gaps and each zone plays a note which was originally played in that zone.

Windows version

To create an instrument, choose File->New Soundfont. Select the soundfont which appears in the left pane and right click->Import Samples. Choose only the first note’s WAV file. This will create the instrument and load it with the first sample. Expand the tree view to show the instrument, and right click->rename to give it the name you would like to see in Crystal’s Oscillator Type menu. This picture shows the instrument renamed to «Glassboom»:

Now, right click on the instrument you renamed, choose Import Samples, and import the rest of the samples.

Select the instrument and you’ll see a zone in the pane in the upper right for each sample. Adjust the zones for each sample by selecting a sample, then dragging the zone markers. The zones should be non-overlapping and each sample’s zone should be the vicinity of the base note that was originally played for that sample (which should be the same as the base note which was assigned in SeamlessLooper in the previous step), as in the following picture.

A tutorial on how to make awesome sounding vocals for Friday Night Funkin’

Hello, I am Avenge, a composer who knows their way around FL Studio a fair bit.

By the end of this, you should be able to make your own vocals and tune them to whatever track you want and make them sound good. Let’s get started! (or watch these videos:

how to make the vocals and use them: https://www.youtube.com/watch?v=S1D-hVmGYqs

how to make a chromatic scale with those vocals: https://www.youtube.com/watch?v=uz2fRFB39VI

how to make a soundfont with that chromatic: https://www.youtube.com/watch?v=QmcU9Bh6jUQ

2. Making the Sounds

Good question! Luckily this guy did it for me:

(watch it all, very useful stuff)

A full list of shit you can do in FL Studio would take many, MANY tutorials to fully explain, however. I will only go over what you need to know. I’m going to assume that you already have a backing track for your vocals; if not, make one!

3.1. The nitty-gritty

Open up either a new project or an already existing one if you have a backing made. Open up the browser* (View>Browser) and click this icon:

Click on “Configure Extra Folders” located at the bottom of the dropdown.

That will bring up a menu that looks like this.

Click one of the folder icons in an empty row, and find your folder with all of your vocal samples. Now, your folder should be located within the browser in FL, which will make the whole process easier. You’ll be able to drag and drop samples onto the playlist without having to go through your actual files now.

3.2. Making a melody for the vocals

Now we’ll need our vocals to have a melody. This is where making a placeholder melody will come into play, as it’s hard to think of the melody when you have monotonous “a’s, ee’s, and oo’s.” If you already have one, you can skip these steps.

3.2.1. Making a melody

3.2.2. Installing sforzando

After downloading the installer, double click it to open it and follow the steps. The default configurations should be fine.
Open up FL Studio if you haven’t already, and navigate to the Options tab and click on “Manage Plugins”

Then you’ll click on “Find More Plugins”, FL will start scanning for new plugins; this may take some time depending on how fast your harddrive is.

When it’s done, scroll down until you see sforzando highlighted in yellow. Click the star next to it.

3.2.3. Using sforzando

Open the channel rack and load sforzando as a new instrument. It should look like this:

Click on the dropdown icon on the instrument thing, and select “Import.”

Navigate to where fnf_voices.sf2 is and load it. A prompt will pop up. Press ok for the next two options.

Click on the instruments tab again, selected “Converted” and navigate to BF001 (it should be fairly straightforward.)

Congratulations, you now have a BF soundfont at your disposal! Just remember to assign sforzando to a mixer track, apply an EQ, and boost the levels by +12db; it’s really really quiet without it. If you don’t know what that means, search up “effect chains FL studio” on Google. (I REALLY recommend adding some sort of effect to BF, as without it he sounds very out of place in a lot of songs)

3.3. The nitty-gritty (part 2)

Get rid of the vocals you’ve just placed, go to the browser, and click this button:
This refreshed your browser, and the new audio file should be there. Drag and drop that into your playlist where the melody starts.

3.4. This is where the magic happens

This next part may get a little confusing. We’re going to be using the plugin NewTone (which comes with FL Studio) to manually autotune our vocals to the notes of our melody; don’t worry, it’s easy.

3.5. Changing formant

3.5.1. Edison

Next, we’ll go to the Master mixer track and load the effect Edison.*

Click on the record button, and play what you currently have in NewTone. Stop it when NewTone runs out of notes to play. You can now export this audio file with the save icon in the top left, and save it to the same folder with your audio clips. Refresh the browser again.

3.6. The nitty-gritty (part 3)

Now you have auto-tuned vocals! Drag and drop that into the playlist and get rid of your placeholder melody. You can assign the audio file to a new mixer track and add various effects to it, such as Chorus and Reverb, to make it sound cool. I put a 3rd-party bitcrusher on mine and some reverb; here’s what that sounds like: cool_vocals_thingy.mp3
Sounds like a mix between Garcello and Tankman; cool!

4. What now?

You now do this for all of the vocals you have in your song (except for BF of course, there are various options for that which don’t require this method).

I’d recommend not doing all the vocals at once in NewTone, but rather in sections. If you make a mistake, you don’t have to risk doing it all over again.

Thank you for reading my tutorial, and I hope this helps you out.

Below, you’ll find audio of a finished example using this method, made by me. a_cool_thing.mp3

4.1. Chromatic stuff too

I have learned about SliceX, which can be used to turn a sample into an instrument you can play in the piano roll. BBPanzu has a tutorial on it.

That should make this process much easier, but also removes the flexibility of being able to place specific samples. In most cases, though, this shouldn’t be an issue.

Here’s the powerpoint I’ve made, in case you wanted it.

Creating and using custom SoundFonts in MuseScore

Summary

In this tutorial we’ll attempt to make a night-time forest soundtrack using some bird noises we record as samples and then include in a custom SoundFont that we’ll use to make a score in MuseScore. We begin with brief instructions on the use of Audacity to record the samples, followed by a short guide to using PolyPhone to create the SoundFont. We then discuss using the custom SoundFont in MuseScore to create a sound effects track.

The tutorial MuseScore ( ForestBirdiesTrack.mscz ) and SoundFont file (ForestBirdies.sf2, from zip file ForestBirdies.zip ) that we create are available to download as part of this tutorial.

Note that both Audacity and PolyPhone have detailed usage instructions online, so we try to avoid duplicating this information by focussing on just the key features required to run through the tutorial. The methods however are applicable to the creation of any custom SoundFont that can be used in MuseScore.

References (section 4 below) list the links you can use to get hold of and use Audacity and PolyPhone.

1. Quick start guide for playing demo Bird Noises soundtrack without following tutorial

If you just want to quickly hear the bird noises SoundFont in action without following the tutorial, follow these instructions.

1.1. Install the bird noises SoundFont file

1.2. Download and open the score file

1.3. Add the new SoundFont to your synthesizer list

1.4. Play the track

You should be ready to go: click ‘Play’ and you should hear a short C scale on a piano, followed by the bird noises.

2. Detailed instructions for creating and using a custom SoundFont

2.1. Recording the wooden instruments and creating sample files.

2.1.1. Instruments used to create the SoundFont

To create the SoundFont in this tutorial I recorded some bird noises using a set of wooden bird noise instrument (whistles etc.) purchased from a market in Arles, near Avignon, France a few years ago. They are packaged in a wooden box with a glass lid. The maker’s details are listed at the end of this tutorial.

2.1.2. Sampling instruments using Audacity

The list of sample files, their size and names are listed in a table at the end of this document. The attached SoundFont for this article ( ForestBirdies.zip below) contains the original samples, which can be extracted from PolyPhone if need be (although this is not required for this tutorial).

2.2. Creating a custom SoundFont using PolyPhone

The SoundFont file format architecture includes the concepts of Samples, Instruments and Presets.
The table below briefly discusses the 3 main components of a SoundFont file. In our example, we are only using the most simple properties of a SoundFont – the format allows for far more customization, but our bird noises example does not require the use of these additional settings.

2.2.1. Create your new SoundFont file

Click New SoundFont and set Product to your chosen name for the SoundFont. You can also add copyright information and free format comments.

2.2.2. Add samples to your SoundFont

2.2.3. Add Instruments – one per sound

Once the samples have been added, you then add a new instrument by clicking the Speaker icon.

Give your new instrument a name. In the ForestBirdies.sf2 SoundFont, make an Instrument for each sample, matching the name to the sample name (e.g. Cuckoo). Then, bind the samples to your Instrument by right clicking your sample file (e.g. Cuckoo) and select Bind to… to attach it to the corresponding instrument (Cuckoo).

Now you can click on your instrument (Cuckoo1) and set the keyboard notes you wish to use to play the sound – do this by setting the Key Range values (e.g. 48-51 for the Cuckoo sound), corresponding to C3 to D#3 (middle C is C4). So that the SoundFont preset knows which note a sample pertains to, it’s important that you set the Root Key for the instrument – this will be the number of the key that starts the key range (48 or C3) and reflects the ‘original’ frequency at which the sound was recorded.

The result of these edits is that for each sound there will be 4 notes on your allocated to play the sound, starting at the root key and moving up in frequency by a semitone. When playing the notes, you can choose which of the stepped semitones you wish to use – the original frequency, or slightly higher (and faster). You will usually want to loop your sample here too: double click Loop Playback in the Instruments window for Cuckoo, and select the middle icon (ordinary looping). This type of loop simply starts playing the note again after its elapsed duration, and repeats for as long as the key is held down.

Repeat this process for each of your samples/instruments.

2.2.4. Create Presets for each Instrument

Finally, you need to create Presets for each instrument: Preset settings tell the SoundFont how it will be selected in the SoundFont including the bank and preset numbers. You can also attach multiple instruments to a single preset, although this is not required in this tutorial. Click the Notes icon and enter a name for the preset (again using the same name you used for the sample and instrument e.g. Cuckoo).

To attach the instrument to the preset, right click the instrument and select Bind to…, selecting the corresponding preset. In our case, leave the preset settings blank, i.e. there is no need to override the instruments key range or root keys – these are picked up from your Instrument settings in the previous step.

Save your SoundFont file, and you should have a new ForestBirdies.sf2 file, ready for use in MuseScore. The summary screen from PolyPhone shows the results of creating the SoundFont for this tutorial.

2.3. Using your new SoundFont in MuseScore

As this tutorial is not designed to repeat detailed documentation for MuseScore, we’ll simply cover the steps require to create a new score using your custom SoundFont – refer to MuseScore’s documentation for more detailed information on these processes – we’ll just list them in summarized format, on the assumption you can refer to the original MuseScore documentation for more information.

2.3.1. Add your new SoundFont to MuseScore

Identify the location of your MuseScore’s SoundFont files by checking your Settings/Preferences and copying your SoundFont file to this directory location. Then, in MuseScore View – Synthesizer, click Add and select your new SoundFont file. Position the SoundFont at the top of list to ensure the bird sound plays, and not a sound from one of your other loaded SoundFonts.

You should now be able to use the SoundFont file in your next score. Click Save to Score if you want MuseScore to load the new SoundFont file when your new score is opened. By storing the reference to the SoundFont file in your score, you can avoid cluttering up your instrument selection with unneeded SoundFont patches. Note that SoundFont files disappear from your MuseScore SoundFont directory if you delete them in the Synthesizer window – for this reason, it’s a good idea to have all your SoundFonts backed up in another location ready to reload if you work on a different score using different SoundFonts.

2.3.2. Create a score to use your SoundFont file

In this example of a ‘bird noise’ score, we need a staff for each bird noise (i.e. instrument). The example SoundFont file has 6 bird noises, so we create a staff for each sound. The sample MuseScore file also has a piano instrument with a treble and bass staff, making a total of 8 staves.

If you’re using the template MuseScore file for this tutorial ForestBirdiesTrack.mscz (available to download below), the score has been set up with a staff for each instrument (bird noise).

2.3.3. Set Stave/Part Properties and clefs for each instrument

For each staff, select Stave/Part Properties (right click a note or rest in the staff to choose this setting) and give each staff a name matching the specific bird noise e.g. Cuckoo, Owl Long, Owl Short etc. Note that you will need to do this one-by-one, and select OK each time to update the score as pressing the up or down arrow to move through the staves does not seem to ‘take’ the previously set staff property.

Take note of the clef settings in the tutorial MuseScore file too – these have been set up to reflect the key range of the corresponding instrument (bird noise).

Remember that each instrument (bird noise) has a playable range of only 4 notes because you assigned this key range when creating the SoundFont file. Notes outside the key range play silence. To help ensure your notes for an instrument (bird noise) match that instrument’s key range, you can optionally set the ‘Useable pitch range’ for each staff in the staff settings. To activate this feature, you’ll also need to check Color notes outside of useable pitch range in Preferences – Note Input. Now, if you enter a note outside of the playable range, the note will be colored red. This has been set up in the tutorial score file for you.
After following this process, you should have a score with 8 staves, with any notes entered outside of the instrument’s key range highlighted red (viewable only after you’ve unselected a note in the bar).

2.3.4. Set your instrument for each staff using the Mixer window

2.4. Write your bird sound ‘notes’

2.4.1. Entering notes within range limits for each instrument

Remember that when creating the SoundFont file, you set a key range e.g. Cuckoo was set to C3 to D# 3. Make sure you enter notes that fit within this range for the relevant instrument, or you won’t hear any sounds! The attached MuseScore file shows the ranges for each bird noise – there are 4 notes per sound, in sequence from C3 to B4. Each sound has a key range of 4 semitones, raising the frequency slightly higher each time.

2.4.2. Out of key note range example

With your new SoundFont working in your new score, you are then free to use MuseScore to create the rest of your ‘music’ (or in this case, a soundtrack). The sample score for this tutorial simply plays each note for each bird noise within its key range – use these bars to copy and create your own night time forest soundtrack, but make sure you stick to the key range for each instrument, or you won’t hear anything!

4. References

4.1. Files used in this tutorial

Note that these sample files were not uploaded to this tutorial, but can be extracted from the SoundFont file if you want them.

How to make soundfont

To create a soundfont, click on Create a soundfont in the home screen. Soundfonts are made from a 3-level structure, so the creation of a new soundfont involves three fundamental steps:

Sample preparation

Sample loading

Sample looping

Next, a loop may be required in the sample, if you would like it to sound longer than its normal length. An example of this would be if you only have a 1 second flute sample, but would like it to sound indefinitely. This can be done by manually or automatically assigning loop points (start + end) within the sample boundaries, so that the loop area can be repeated and thus make the sample play longer. While you can assign loop points manually, using the “Auto loop“ function to assign them automatically usually produces better results and is a lot faster.

To assign loop points manually:

Note: to hear looping in an instrument, you must select in the Loop playback parameter row of either the Global or individual notes columns in the Parameters Table. turns it On, or a blank turns it Off.

Loop illustration

Sample tuning

Finally, a tuning has to be done. To do this, a calibration tool (sine) is available in the editing page of samples. For each sample the method is as follows:

The frequency analysis can be a good indicator to start tuning.

To go further

Several tools are available to edit samples like:

Instrument creation

For this step, it is assumed that one or more samples are available for the creation of an instrument.

Creating an instrument and adding samples

First, click on New instrument in the toolbar. A name must now be entered.

Then add samples to the instrument using a “drag & drop”. In the tree you will notice that divisions appear in the instrument.

Note: When divisions are created the samples are not copied but linked. There are no constraints on the number of divisions.

Distribution of divisions

When samples are added to instruments, they appear as divisions (columns) in the instrument editor table. Each division must then be positioned on the keyboard by changing the Key range in the table. Generally, it is a good practice to have the range of a division include the root key of its represented sample. The entire surface of the keyboard must be covered (eg, key 36 to key 96 for a classic synthesizer keyboard).

The tool “Sample auto-positioning” automatically distributes the samples over the keyboard.

Setting the divisions

To enable the loop playing mode in an instrument containing looped samples, select in the Loop playback parameter row of either the Global or individual notes columns in the parameter table.

The global division, as its name suggests, allows you to enter a parameter for the entire instrument. A global parameter applies only to the divisions whose same parameter is not specified. This means that the parameters of the standard divisions take precedence over the settings of the global division.

In the context of this simple tutorial, no other parameters have to be edited for the instrument to work. It should be already possible to play with the virtual keyboard.

However, to improve the instrument it is advisable to enter a value in the Vol env release (s) row, which prevents the sound from stopping abruptly, as soon as a key is released. The other parameters are described in the description of tables.

Preset creation

For this step, it is assumed that one or more instruments are available for the creation of a preset.

What is a preset?

A preset is the equivalent of a sound patch name found in hardware synths. It is visible from outside a soundfont and is identified by a bank number and a preset number. A soundfont can have one or many presets. Each preset contains one or more instruments, just as instruments contain one or more samples.

Creating a preset and adding instruments

First, click on New preset in the toolbar. A name must now be entered.

Then add instruments to the preset using a “drag & drop”. In the tree you will notice that one or more divisions appear in the preset. Often there may only be one, as in the case of a preset containing a single intrument.

It is possible to create as many or more presets as instruments, and each preset can contain one or more instruments.

Setting up a preset

Changing the preset parameters may not be required since:

For the scope of this tutorial, the creation of a soundfont is finished!

How to make soundfont

To create a soundfont, click on Create a soundfont in the home screen. Soundfonts are made from a 3-level structure, so the creation of a new soundfont involves three fundamental steps:

Sample preparation

Sample loading

Sample looping

Next, a loop may be required in the sample, if you would like it to sound longer than its normal length. An example of this would be if you only have a 1 second flute sample, but would like it to sound indefinitely. This can be done by manually or automatically assigning loop points (start + end) within the sample boundaries, so that the loop area can be repeated and thus make the sample play longer. While you can assign loop points manually, using the “Auto loop“ function to assign them automatically usually produces better results and is a lot faster.

To assign loop points manually:

Note: to hear looping in an instrument, you must select in the Loop playback parameter row of either the Global or individual notes columns in the Parameters Table. turns it On, or a blank turns it Off.

Loop illustration

Sample tuning

Finally, a tuning has to be done. To do this, a calibration tool (sine) is available in the editing page of samples. For each sample the method is as follows:

The frequency analysis can be a good indicator to start tuning.

To go further

Several tools are available to edit samples like:

Instrument creation

For this step, it is assumed that one or more samples are available for the creation of an instrument.

Creating an instrument and adding samples

First, click on New instrument in the toolbar. A name must now be entered.

Then add samples to the instrument using a “drag & drop”. In the tree you will notice that divisions appear in the instrument.

Note: When divisions are created the samples are not copied but linked. There are no constraints on the number of divisions.

Distribution of divisions

When samples are added to instruments, they appear as divisions (columns) in the instrument editor table. Each division must then be positioned on the keyboard by changing the Key range in the table. Generally, it is a good practice to have the range of a division include the root key of its represented sample. The entire surface of the keyboard must be covered (eg, key 36 to key 96 for a classic synthesizer keyboard).

The tool “Sample auto-positioning” automatically distributes the samples over the keyboard.

Setting the divisions

To enable the loop playing mode in an instrument containing looped samples, select in the Loop playback parameter row of either the Global or individual notes columns in the parameter table.

The global division, as its name suggests, allows you to enter a parameter for the entire instrument. A global parameter applies only to the divisions whose same parameter is not specified. This means that the parameters of the standard divisions take precedence over the settings of the global division.

In the context of this simple tutorial, no other parameters have to be edited for the instrument to work. It should be already possible to play with the virtual keyboard.

However, to improve the instrument it is advisable to enter a value in the Vol env release (s) row, which prevents the sound from stopping abruptly, as soon as a key is released. The other parameters are described in the description of tables.

Preset creation

For this step, it is assumed that one or more instruments are available for the creation of a preset.

What is a preset?

A preset is the equivalent of a sound patch name found in hardware synths. It is visible from outside a soundfont and is identified by a bank number and a preset number. A soundfont can have one or many presets. Each preset contains one or more instruments, just as instruments contain one or more samples.

Creating a preset and adding instruments

First, click on New preset in the toolbar. A name must now be entered.

Then add instruments to the preset using a “drag & drop”. In the tree you will notice that one or more divisions appear in the preset. Often there may only be one, as in the case of a preset containing a single intrument.

It is possible to create as many or more presets as instruments, and each preset can contain one or more instruments.

Setting up a preset

Changing the preset parameters may not be required since:

For the scope of this tutorial, the creation of a soundfont is finished!