Soundfont Creation

• Sep 18, 2024 - 00:42

Hello folks,

I am learning how to create soundfonts and I'm struggling.

I am starting off just experimenting and trying to make one soundfile with one sample and have it play back. I have not had any success so far and I have scoured the forums, the online manual and the internet looking for guidance, and while there is a lot of info about soundfonts, I do not find the key info I need. If anyone knows of a simple tutorial for creation of a very basic soundfont that tells you the essentials to make the thing work in MuseScore, I would be very grateful.

I am using Audacity to create my WAV file and Polyphone to create the soundfont. The best I have done so far is to create a soundfont with a single WAV file where the first split second of the sample plays but that is it. The WAV has a tiny sound at the beginning, a bit of silence and then a vocal tone for about 5 seconds. It plays the tiny sound at the beginning. I have tried using the loop function in Polyphone so that only the vocal tone is selected but then nothing at all is heard playing when I assign the soundfont from the mixer.

I could not attach the soundfont file. I have attached an MSCZ file with a measure where I have chosen my soundfont using the mixer. I applied it to the drum kit but I have also selected it on the same staff as the nylon string and the result is no different. If anyone has any insight on why this will not function would they please let me know what I am doing wrong?

Attachment Size
Soundfonts Testing.mscz 19.79 KB

Comments

• Sep 20, 2024 - 17:30

So, thank you for the link. I've actually seen this page before but there is so much information, I don't know where to start. But I'll take another run at it.

In the meantime, I've uploaded a link to the actual soundfont file via Google Drive below. And I also added 10 screenshots from Polyphone that I think show all the attributes of the sounfont (called Testoo).

Reply

• Sep 20, 2024 - 17:38

I have added a lot of info since my original post. Anyone who has any expertise with soundfonts can probably identify my errors in seconds and set me straight in a few words. Unfortunately, the posts and links, etc., I find in the handbook and the forums have such vast amounts of information that I have trouble drilling down to the few simple things I need to know. My goal at this point is to create any soundfont that works. Ultimately, I will create a soundfont that has a collection of Golpe sounds (percussive guitar sounds made in various ways, for example, by tapping the body of the guitar somwhere).

I will be very grateful to anyone who can help me sort this out.

Thanks,

Peter

Reply

• Sep 20, 2024 - 19:08

I know what you mean about the documentation. It took me quite a while just to get my head around the basics, (including lots of trial and error), and I haven't moved on much from there.

I'll take a look over the weekend but until then you could download my guitars soundfont to see how I've done things as it might give you some pointers.

YonahGuitars
https://1drv.ms/f/s!AtZbFhgSR71Ng6sz7rZz17WfXl-SrA
(The .txt files gives info on sources and customisations)

Reply

• Oct 8, 2024 - 23:03

Hi yonah_ag,
just thought I would reach out to you in the only way I can figure out because you seem really into soundfonts. I was just wondering if you've ever tried to create a custom drumset (file type = *.drm).

I've had some success with creating soundfonts and I'm using them in a score. I edited the drumset I was using for the soundfont within a test score and I noticed when I tried to update my real score and use the soundfont I created, it was no problem to select the soundfont but I had to edit the drum kit from scratch in order to use the soundfont file. I'll be using my soundfont on other scores and it would be nice if I could select my custom edited drum kit from the instruments tab "add" function. I see that you can create a *.drm file alright but it is not obvious where you need to save it to be able to select it from the instruments tab add function. I found *.drm files in a folder within program files\musescore4 but none of their names match with the selectable instruments and there are far fewer of them than the number of selectable instruments. I see that there is an instruments.xml file which contains the names of many (perhaps all) of the selectable instruments, e.g., tambourine. I'm not sure if this file contains the info for drum kits or if it is from this file that the data for selectable drum kits is extracted for the "add" function. I can't find anything helpful in the handbook.

If you know anything about this process and would like to share, I'm all ears. And if not, that is also fine. I will likely put up a separate post after I do some experimenting. I'm not optimistic the experiments will be successful, but you never know.

Thanks,

Peter

Reply

• Sep 20, 2024 - 22:52

Presuming that you only want the 5s of vocal sound then you will need to tidy up your wav file in Audacity by trimming the silence at the start and removing the tail after the 5s. It would also make sense to normalise the amplitude. 5s is a decent length sample so you may not need any looping.

Reply

• Sep 21, 2024 - 04:34

Hello yonah_ag. I have just created a Golpe soundfont that worked in MuseScore. I am elated! Thank you for providing info, it was immensely helpful, got me looking in the right places. Thank you for generously donating your time and effort into supplying me with info. I am extremely grateful. I think you can forgo delving into this on the weekend. I've got to a point where now I just need to refine. Having said that, I am happy to recieve any advice you have.

I noticed on one post I looked at, the poster said you need to use bank 128 for non-pitched instruments (that is pertinent because my ultimate goal is to have a soundfont with different percussive guitar sounds (Golpe)). It didn't seem like that was even possible but I could be just missing something. I just did it with the key range set to 0-127 and it did create the sound. So I'm not sure what he meant. But also worth noting is that I assigned the soundfont to a drum set and the different notes on the staff did change the pitch. I did try to assign bank 128 but it would not accept it. Anyway, just a couple of observations.

Once again, thank you for your comments. You got me thinking about the right things. I think the comment about the midi key range was critical as was the suggestion to trim silence. I think that is why it worked.

Anyway, I just want to, again, say thanks. You know what it is like when you finally make a breakthrough on something you've been trying to figure out for days. Well, I am there. Your comments helped me a lot. Thanks!

Peter

Reply

• Sep 24, 2024 - 20:22

Glad to have helped.

There is a standard called General MIDI which defines which preset number each sound should be allocated. If you stick to this then your soundfont will have its presets in the expected positions. My guitars soundfont uses bands to have different sounds on the same preset number. You can set this in Polyphone and it is helpful in the MS3 mixer.

Reply

• Sep 21, 2024 - 04:43

One more thing that is interesting: I only used one sample in my soundfont. This one sound was used on each of the notes presented by the drum kit but just in a different pitch. I will be using about 6 Golpe sounds in my soundfont, so I think I need to adjust the key ranges for each one so they will only be used by one of the notes. I will post a pic that I hope will clarify what I am saying.
Drum Kit Notes.png

Reply

• Sep 21, 2024 - 04:50

All of the notes you see above used the single sample from my soundfont but just changed the pitch. I'm guessing each of the notes here have a specific range, so if I figure out what the ranges on the notes are then I can set the key ranges in each of my golpe sounds to correspond with specific notes. So as long as the key range on the note is similar to what the actual key range of the sound is then I should be able to have a functioning, multi-golpe soundfont.

Reply

• Sep 21, 2024 - 05:01

I've just looked at the chart at this link and this was info I kept wondering about as I started to get an inkling of what needed to be done, so thank you for this chart. This will be very helpful - perhaps essential to the soundfont creator. I just posted a pic of your link to the chart, so, of course, the link isn't functional here but it is above at your oringinal post. Essential.

Reply

• Sep 24, 2024 - 20:07

Well, I have now created soundfonts that work in MuseScore. Of course, now I have a new problem. I use a multi-track digital recorder to create a stereo soundfile and then I record that into an Audacity file and use that recording to export a WAV file to my PC. From there, I open Polyphone and use the WAV files I've created for samples. I add an appropriate key range (midi value) and all that.

My Problem: once I have created the soundfont file and then selected it in MuseScore, I find that sound has lost a lot of volume and fullness. It sounds fine in Audacity but then gets really quiet in MuseScore.

I have seen a post in the MuseScore forum referring to MuseScore synthesizer which can be used to change settings on the soundfont file. But of course in MuseScore 4 there is no synth that I can find. Where is the functionality of the synth now located?

If anyone knows what I am talking about and the best way to adjust the sound of the sample once in MuseScore, I'd love to hear about it. I did turn up the volume on the soundfont file using mixer and that improved things, but I feel I am missing something.

Any comments, suggestions, hints, etc., welcome as always.

Reply

• Sep 24, 2024 - 23:16

Here is what the WAV sounds like in Audacity and what it sounds like in MuseScore. Audacity seems to have a richer sound. Why?

Here is the Audacity Sound, shared via Google Drive: https://drive.google.com/file/d/1b_Hkse19ZJ9eE9rMYNmbF584n85MWYZ8/view?…

And here is the MuseScore sound, shared via Google Drive: https://drive.google.com/file/d/1E8zJLR4zDug3V1Whc5BXJVDTPHZ6nvKU/view?…

As you can hear, the MuseScore sound loses a lot of quality.

Reply

• Sep 25, 2024 - 11:51

In terms of sound quality, I also find that samples sound better in Audacity, (and in Polyphone), than they do in MuseScore. I am not sure why but maybe it's related to the MS synthesiser playback, or maybe manipulating the samples into soundfont instruments and presets is the cause.

Normalising may also improve your MuseScore playback.

I have made various tweaks to EQ and sustain with my guitar soundfonts, (within polyphone) and this can help but it's a bit trial and error for me as I don't really know what all the settings are actually intended for.

I found this blog helped me understand some of the fundamental parameters:
https://stickz.co/blog/adsr-explained-sound-design-basics/

Reply

• Sep 25, 2024 - 11:45

Yes, that is the setting.

I normalise my samples to somewhere between -3 dB and -1 db (but not all the way up to 0 db). Normalising matches your samples to those used by other presets in your soundfont and also in other user's soundfonts, so your sound at ff will sound at a similar volume to other sounds at ff, (and at f, mf, mp, etc.) Without this it is more difficult to balance volume levels from different presets.

Polyphone also has a function for normalising samples on the tools menu.

Reply

• Sep 25, 2024 - 15:17

ok, I'm going to experiment with normalisation. Now, unfortunately, I am so new to this, everytime you post something I see something I am unfamiliar with. In this case, it is the terms ff, f, mf & mp. I think I have seen these referenced somewhere in my reviews of the documentation but I kind of ignored them because I didn't know what they were and if they would affect my soundfont creation. I hate to ask, since you have already supplied so much helpful info, but, can you briefly tell me what these parameters relate to and where would I see them presented?

Reply

Do you still have an unanswered question? Please log in first to post your question.