Fathom Synth Development Thread

[PietW.]

@ Fathom Synth
My suggestion should not be a criticism. The format for Serum, Synthmaster, Icarus, Falcon has now become a quasi standard. U-He also uses this format in Hive. So. There are many wavetables in this format and it would be nice to use them in Fathom as well.
The format for serum uses 256 waveforms, Fathom uses 16 waveforms. Maybe there is the possibility to convert these wav files for Fathom?
As I said. I absolutely wanted no criticism or quarrel.

Piet

I support PietW here. Actually I wanted to write the same... Not quite sure but I think AudioTerm supports the format as well.
What I have written during the Fathom One Synth Challenge was actually the same. I'm quite into wavetables and I have found a way to import speech samples into Fathom. But the process was quite tedious and required some manual steps, you can check out the track below, it was completely done with fathom. The speech sample for "kybernetik" was 4 fathom instance because the number of WT slot is/was limitied to 16 only. So I split the words into 4 times 16 slots with 4096 samples each...
Anyway, long story short - WT synths like Icarus have a set of quite supportive tools to import arbitrary wavs into WT or edit WTs. Icarus for istance has the "resample" option that takes an sample and creates a proper WT of it. It's basically the "manual" steps I have described with some extra mojo .
Why the 2048 samples - it's very close to note F1 played at 44.1Khz...i.e. if you do the maths it's one cycle of 2048 samples for the F1.
I think you can even check out my patches - because I took 4 place in the OSC I was able to contribute them: It's the programms "kYbernetik 1.Lead.PeterH._.0.Program" - "kYbernetik 4.Lead.PeterH._.0.Program" in the "international/PeterH" folder. Import them into your DAW and play them one after another with a F1 and it should reproduce the word.

I remember that I have even edited the Fathom WT files - which are plain XML with individual samples in a list - by hand. The list output I generated from Audacity which has a textual exporter. Probably a proper spec of the file type would enable programmers (which are not lazy as I am) to write importer software .

Just one Idea - Probably a "Open Extensible Editor Interface API" where musician-programmers could add their javascript editors as "plugins" to Fathom would also be a good thing... thinking here about how Bitwig supports arbitrary hardware controllers...

Why I write this: Because Fathom is a Beast Synth - I have not yet seen any other Synth with for instance the "Leave one out" Option for the Oscillators. That shows me that Everett is a clever guy with lot's of ideas. Just wanted to contribute some things I have learnt elsewhere that could help Fathom.

https://soundcloud.com/p-e-t-e-r-h/the-kybernetic-gods

[FathomSynth]

Wow, a lot of activity lately on this wave table issue.

I will up this in priority and try to get this in the next release so Fathom will be able to read and write Serum wave tables.

The format looks like it may be as simple as just a big WAV file with 256 cycles and 2048 samples per cycle.

Could someone email me a Serum wave table so I can experiment with it.

[fmr]

The format looks like it may be as simple as just a big WAV file with 256 cycles and 2048 samples per cycle.

It doesn't need to have 256 cycles. It may have less (minimum is two, I think), and it also doesn't need to be 2048 samples (may go to just 256, and up to 4096). There is a header in the file that tells the synth what is the sample cycle, and from there, how many cycles there are in the wavetable file. Then, the synth engine should be able to interpolate and fill in the empty slots.

Here are some details posted by Seve Duda himself:

This is an example CMNT [edit: oops, not CMNT, 'clm '] chunk from a Serum-created wavetable (.wav) File:

Code: Select all

<!>2048 01000000 wavetable (www.xferrecords.com)

Serum currently assumes 2048 (samples per frame) at all times, so as of now 2048 should always be written there. If someone wants the ability to write 1024 or 4096 frame tables, I could make Serum support that with little effort - maybe drop me a line on xferrecords.com so I can update Serum to support that ability.

Only the two first flags are currently used:

• the first flag is the WT interpolation [0 = no interpolation, 1 = linear crossfades, 2,3,4 = spectral morph]

• the second flag is "Serum Factory WT" which means Serum assumes this file comes with Serum and everyone already has it - thus it will not embed in to user presets to keep file sizes down. PLEASE DO NOT ENABLE THIS FLAG IF YOU ARE CREATING WAVETABLES - please leave it to zero, thank you very much. If you want a similar flag for yourself to identify tables as factory or otherwise for your product, drop me a line and I will reserve you a flag or a different value on that flag.

[FathomSynth]

These file headers are part of the WAV format?

[fmr]

These file headers are part of the WAV format?

They are supposed to be written in the WAV file yes.

[zzz00m]

Could someone email me a Serum wave table so I can experiment with it.

Here is a link to a free Serum soundset with a folder containing a dozen wavetables in .wav format for Serum.
http://patcharena.com/downloads/comment.php?dlid=1315

And there is this for free, with your email. Probably more up to date, with 50 wavetables included. https://cymatics.fm/xfer-serum-starter-pack/

[fmr]

Could someone email me a Serum wave table so I can experiment with it.

Sent you a PM. Wavetable created directly in Serum.

[FathomSynth]

OK, thanks, that will help.
Peter:H thanks for your comments.

I read the other thread and looked at the WAV format specifically. I'm not totally clear if Serum is saving their table into the WAV file as 256 separate RIFF Chunks, or simply as one big wave with 256 cycles.

[] Peter:H []

OK, thanks, that will help.
Peter:H thanks for your comments.

I read the other thread and looked at the WAV format specifically. I'm not totally clear if Serum is saving their table into the WAV file as 256 separate RIFF Chunks, or simply as one big wave with 256 cycles.

I think the format is a simple as a wav file with up to 256 * 2048 samples. It's just interpreted as cycles by the lenght of 2048. I think you could probably grab a demo copy of Tone 2 Icarus and check out what they provide as content. I have added a screenshot of 3 different examples from Icarus. I cannot see anything special, no markers, loop regions or whatsoever. It's just a "convention" of how to interpret these.

[Rumdrum]

BTW I am only using ONE instance of Fathom. I have done all I can do according to the manual for CPU and memory. There is no difference. It still cradles heavily playing poly or fast.

[zzz00m]

BTW I am only using ONE instance of Fathom. I have done all I can do according to the manual for CPU and memory. There is no difference. It still cradles heavily playing poly or fast.

It could be the presets that you are using. There are a few that crush my system, and some that play just fine.

Try creating a preset of your own, with just one oscillator and filter and build up from there. Find the tipping point, and maybe re-structure a few presets to stay under that point.

[farlukar]

Hmmm… I assume you don't need an introduction on the Fizmo?

[carrieres]

Looks like exposure of Fathom is going to increase hugely on the 20th March! Congratulations! Really good news! I hope sales go well from it.

Why? What's happening March 20th?

Fathom CM :

[FathomSynth]

Yes, isn't that great. We are on Computer Music's radar. Very huge!

I have been in touch with them over the last two years and they have been very cool about keeping up with Fathom's development.

In terms of CPU, Yes, that's good advice.

You can also go the opposite direction and take one of the overbearing presets work backwards taking stuff out.

There's a chapter at the very start of the user manual on how to do this.

Some of the presets were made on my super system using Ableton (which is highly efficient) and you can easily simplify them within range of your system without hurting the quality at all. For instance in a very complex patch if one of the oscillators is using 5 detune voices instead of 2 you can change it back to 2 without it making any real difference in sound.

Top priority in 2019 is handling the CPU.

[zzz00m]

Congrats on the CM release!