No posts in this section yet, so I thought id start with some easy overlooked tips on getting better sound and volume for your saber font sounds.
Most often, people assume the first place to start for getting a better saber sound is by replacing the speaker. While replacing the speaker may have good effect (depending if the speaker is of better quality), there are many other easier ways to get better sound without having to rip open your saber and grab the soldering gun.
1) Check your resonance space/chamber -
- The resonance chamber space is the empty space distance between the speaker and the pommel (if any space exists). A resonance chamber basically traps some of the audio in this space and the sound waves bounce around inside it, before exiting the pommel vents. Resonance space helps create a bit more low end and bassy sounds (bass build up is created in this chamber), and sometimes even gives more apparent volume. Too short of a resonance space makes a saber sound a bit brighter and more clarity, with the sacrifice of reduced low end. Too long of a resonance chamber will sound dark and muddy. I’ve found the optimal balance of equal parts of clarity and good low end, with the apparent effective volume increase is about a space/distance of 1" between the speaker and pommel.
2) Ensure proper pommel ventilation -
- An obvious easy fix here. If a pommel is not properly vented, the sound is choked and muted. I’ve had many of pommel styles, and have noticed all will benefit by simply drilling in more holes in your pommel to get the sound waves out, so your audio will clearly be heard. There’s no rule of how many ‘more’ holes is needed, simply keep drilling until your sound is more defined and no longer muffled. One some of my pommels that are built with thick walled materials, I’ve even gone and drilled in one major main port/hole about 1/2" in diameter directly in the bottom center of the pommel as a main sound port, and then drilled smaller holes around it, which has improved the sound tremendously.
3) Speaker Replacement -
- This is usually a last resort for me. Mostly because I’ve tested out various speakers (of the same size/type) from different companies all varying in price and found the difference to be too subtle, and the 2 above suggested fixes (resonance and ventilation) are far more apparent in noticeable & beneficial changes. Off all the speakers I’ve tested (same size/wattage), all performed relatively the same, gave the same apparent volume output, and only one speaker showed to have a difference in sound (SmugglersOutpost on Etsy). This speaker did show a touch more clarity and brightness, however with the sacrifice of slight reduced low end. Almost as if this speakers frequency spectrum was shifted slightly higher to achieve more high end clarity, but with the low end being rolled off a bit more…a trade off. So a speaker replacement isn’t the first place id tinker with if a ‘better sound’ is what I’m after, and usually only go this route if the speaker has shown to be faulty/blown.
5 Likes
Another tip is that some sound fonts sound better if you filter out the low frequencies. The small speakers we use can’t play these frequencies anyways, but they strain the speaker and the amplifier even if nothing actually comes out. Filtering out the low frequencies lets the speaker work better for the frequencies that it can play.
Of course, it would be nice if ProffieOS could do this filtering automatically. And while that is entirely possible, I’m not actually sure how much CPU power would be used up by the filtering code.
2 Likes
Well said! Even though saber speaker specs ‘claim’ they can do 50hz-18k frequency spectrum, I find that to be pretty false, lol. Super low frequencies, all it does is make the smaller speaker cone push/pull more in movement and nothing really audible comes out of it. Rolling off anything below 100-180z on a sound font generally lets the speaker breathe a bit more.
But, If you are going to update proffie to auto-filter low frequencies in future updates, you may want to have a config override for people who may have a speaker that is capable of producing such low frequencies, and dont want this feature active.
Oh absolutely.
The frequency cutoff has to be configurable too, as it depends heavily on what kind of speaker you have. A 33mm speaker may handle frequencies down to 100-150Hz while a 22mm speaker will cut off anything below ~300Hz.
2 Likes
Creating a low/high pass filter preset makes optimization of font files a breeze. If you’re really particular you may want to have one per chassis, as the resonance is different from hilt to hilt…but only the most discerning need go that far.
Another thing I’ve seen is a speaker not sitting snugly. Any vibrations introduced by the cone itself will translate to the housing. As such, it’s imperative to keep it secured in the chassis with no wiggle. Seen a few chassis where the speaker is merely held in by wire tension.
Great writeup!!
1 Like
Adding a dedicated DSP a possibility?
I’ve seen people have luck with sealing the rear cavity behind the speaker as well. Not tried that myself though.
Probably not.
I need to try it and find out how much cpu it actually uses before I do something like that anyways.
One big problem with small speakers is that the sound wave coming out the back can join up with the one coming out the front. Since they are 180 degrees out of phase they cancel eachother out. The sound coming out the rear of the speaker either needs to be delayed, redirected or muffled to prevent that from happening. Lucky for us, a reasonable easy way to achive this is to put the speaker in a tube. 
I’ve been playing with this sort of thing for quite some time to get the sound I wanted.
I tend to use kyberphonics sound fonts and he adds quite a lot of low frequencies that like mentioned above just make it sound muddy of distorted as the tiny speaker tried to replicate it.
So I slowly stepped up the type of speaker and sound chambers in order to achieve what I want.
I have to admit that the built in amp is very good and gives plenty of punch if your speaker can handle it at higher volumes.
The most impressive so far in terms of quality bass and punch has been a 40mm driver 5w speaker in a ported chamber.
I know it sounds a little over the top and fitting it in will take some shoe horning but it has been fun to do.
The 33-34mm driver performed quite well too and did give a noticeable increase to the low end and also didnt really need much in ways for a chamber behind the speaker to highlight any particular frequency. So a good all rounder in my books.
I find the smaller speakers, although the do try!, not to be all that great but I am biased with this.
I did start with a 24mm speaker and I was happy at first but after hearing my sound fonts through the ear phones (it was very late and didnt want to disturb anyone) I suddenly noticed how much I wasn’t hearing!
And so the hunt began.
I will add some renders up of what I have been playing with when I’ve sorted my subscription out with fusion 360.
Pictures would be great!
Also, I feel like at this time this thread could use this video:
https://www.youtube.com/watch?v=tljbajKx-z4
2 Likes
Perhaps an option for Proffie OS: open all new WAV files, high-pass filter each file and save it back to the SD card.
I do exactly that. Batch process the whole font in one shot before it goes into the collection.
High pass 100Hz 36dB slope
Export as 44.1 16bit mono PCM wavs
Then I subfolder sounds with multiples, batch renaming as I go to all 2 digit file names, except bgn/end lock/drag/melt which get single digits to keep it to 8 char.
1 Like
Filtering can be difficult to do with looped sounds though.
Do you have an easy way to filter the hum and swing pairs without causing clicks?
It’s an interesting idea: Write code that does it beforehand instead of at runtime.
An external utility might be better though, as that wouldn’t use up any space in proffieOS.
For the V3 that might not be a problem though.
1 Like
Nope. I’ll manually go in and zero cross them. Sometimes I need to grab the end piece, reverse it and crossfade it with the front end…nothing magically easy.
Then I just end up going through the whole font; ramping swings into the ADSR envelope I feel sounds right, weed out weak clashes or mix in a blast punch to the beginning, usually need to fade some lock and drag sounds in and out of the bgn and end wavs so they butt up seamlessly, and pre WavLen, I would make all of the retraction sounds close in duration.
An audio_tools folder could have some cmdline scripts. EQ Filters, SS gen using Sox, etc…
or maybe an extension of the Style Editor page? Maybe Style Editor, WebUSB and some Audio tuner are just tabs. Audio tuner where you upload, process in the cloud, download. Nothing gets saved online.
Just brainstorming.
1 Like
Doing it through WebUSB would be pretty cool.
Probably not very fast though.
1 Like
Finally sort my fusion 360 out! (darn thing was not playing ball!)
as its now really late I’ll put this teaser one on here for now 
it is utilizing a 40mm full range driver with a 10 watt RMS coil. I’ve developed a bass reflex ported cabinet to enhance the bass down to a rough (and with my math skills that would be doubtful) 108 Hz before the 3 dB drop.
the bass reflex port on this one is spiral shaped, this design allowed me to get the port length needed in a smaller package. its important to keep the phase correct when exiting the port to not cancel out the main speaker output otherwise it’ll just sound flat.
on paper it should work quite well but we will see.
I’m going to be prototyping my designs on the 3d printer to see witch one works best.
I will post more of my rendered drawings later and also my findings on the real world tests.
3 Likes
Will you be doing frequency-response testing, or just listen to it?
Now that I would love to be able to do but I dont have the gear to do it, so yes it will be judged by my ear.
Not very scientific I know but if my calculations are correct it should have a noticable effect. Over a standard sealed chamber. If I get time I will print both and do a comparison.
