Saike's workshop [ JSFX | Synth | Creative FX | Nonlinear Filters ]

[JamesPeters]

Originally Posted by Aesaire

One feature it could use is an option to change the frequency scaling (x-axis on the graph). Would be nice to zoom in and view only 20Hz to 300Hz when analyzing the low end for example.

Right-click and drag a frequency range.

[Aesaire]

Originally Posted by JamesPeters

Right-click and drag a frequency range.

God damnit. Thanks.

[sai'ke]

Good to hear you got the spectral analyzer working. Thanks James for helping out

I've been working the past few weekends on a little mono-synth JSFX (tentantively called Yutani). It's intended mostly for leads and basses.

It's still a bit rough round the edges, but I would like to see if anyone is interested in helping me develop some interesting presets for it that I can ship with it.

Pros
- You help shape the release. I will also happily mention your name in the plugin comments if you wish.
- Your presets will be available to everyone using the plugin.
- Higher probability that if you feel something is really missing that your FR will get in. Note though that my time to work on this is somewhat limited and that I likely won't be able to do everything.

Cons
- Preset format might change, which could corrupt existing presets you made (I will actively try to avoid this though!).
- It takes time.
- There may be bugs.
- UI is not as strong as many commercial offerings.

Here's a glimpse (NOT FINAL):


Features:
- 2 minBLEP anti-aliased oscillators with various modes (saw, tri, square, sharktooth/fin, PWM) that can run in free or synced mode.
- Hard oscillator sync.
- Pitch modulation for oscillator 2 (useful with the hardsync).
- Bonus oscillators (these are direct copies of osc1 and 2 with the same settings, except detuned to a user specified degree).
- Cool (non-)linear multimode filters that you can morph between (2p/4p linear ZDF, MS-20 (NL), 2p ladder (NL), 4p Moog Ladder (NL), 303 (NL)).
- FM modulation of the filter cutoff.
- LFO for cutoff.
- LFO for morph.
- Custom assignable velocity effects.
- Filter ADS envelopes (R time constant is the same as D). The Filter envelope can be used in two ways. The first, which resets the envelope on each attack, or a second mode which accumulates (kind of like older synths like the 303 do, where each successive note gets a slightly higher cutoff).
- Amp envelopes (amplitude can be applied either pre or post non-linear filter).
- Distortion module with warmth control and two modes (normal and hi-gain).
- Oscillator feedback.
- Glide.
- Bugs probably.

[Eliseat]

I'm very excited. And of course I will share my presets as I did with Filther.

Greetings
Eli

Edit: Bonus OSC
Only every 100th buyer!

Btw: Didn't find it in your reapack?

[Eliseat]

Meanwhile I had fun with FM Filter.

https://mega.nz/file/K6IjlQ6A#3tNbeb...XGx3QfSYumX6l0

You can turn the knobs like hell and it colors, colors, colors! Like crazy. Just amazing! Many thanks Saike.

[Zapatero]

Originally Posted by Eliseat

Meanwhile I had fun with FM Filter.

https://mega.nz/file/K6IjlQ6A#3tNbeb...XGx3QfSYumX6l0

You can turn the knobs like hell and it colors, colors, colors! Like crazy. Just amazing! Many thanks Saike.

@Eliseat, well done, I thoroughly enjoyed the snippet of your track.

@Saike: Thanks a lot for your neverending efforts to provide cool tools and innovative stuff!

[Eliseat]

Originally Posted by Zapatero

@Eliseat, well done, I thoroughly enjoyed the snippet of your track.

@Saike: Thanks a lot for your neverending efforts to provide cool tools and innovative stuff!

Thank you. But its more of a nicely wrapped presentation of that one arpeggio pattern with Saikes FM-Filter to show its capabilities. At the end it really goes crazy with the FM amount. The whole mix gets leveled down.
Btw. Its just a stupid headphones mix without any mastering. Just for fun.

And actually I don't make tracks or music nowadays. But I like to be up to date because ... if the muse kisses me, I want to be prepared.

[sai'ke]

Originally Posted by Zapatero

@Saike: Thanks a lot for your neverending efforts to provide cool tools and innovative stuff!

No prob. Happy to hear you like it

Originally Posted by Zapatero

@Eliseat, well done, I thoroughly enjoyed the snippet of your track.

So did I

Originally Posted by Eliseat

Thank you. But its more of a nicely wrapped presentation of that one arpeggio pattern with Saikes FM-Filter to show its capabilities. At the end it really goes crazy with the FM amount. The whole mix gets leveled down.
Btw. Its just a stupid headphones mix without any mastering. Just for fun.

That sounded like a lot of fun! Very enjoyable to listen to.

I also really like that you made the self oscillation a creative effect rather than a problem to deal with

And actually I don't make tracks or music nowadays. But I like to be up to date because ... if the muse kisses me, I want to be prepared.

I know what that feels like

I just added Yutani to Reapack. But mind you, it's still pretty rough around the edges. And very much a mono synth. Still, I hope you get some use out of it.
There's one issue I'm still trying to iron out, which is one manifestation of the dreadful stuck notes issue. If you suddenly don't hear the envelope going again, assume a note is stuck and jump somewhere on the timeline.

Clicking "VEL" puts you in velocity mode. Then you can set what happens to the highlighted parameters depending on the MIDI velocity.

Also, if you install via reapack, there's some presets that I made in this location: %APPDATA%\reaper\effects\Saike Tools\yutani\bass_presets.rpl
But uh, it's my "mess around as I go" list, so it's pretty hit and miss. I deliberately didn't make it the default yet as I feel I ended up doing that too early with Filther.

[Eliseat]

Hey Saike,

its in the reapack and I downloaded it but unfortunately I can't find it in the JS list. Could it be because you gave it the category Yutani instead of synth? I have no idea.

I don't think Yutani could be more stubborn than Filther. And I tamed that old beast pretty well. So no warning needed. I'm ready to take the fight.

Edit: Okay, after a restart its there. Time to get my very first impression.

Edit2: There is something wrong with the feeding of MS-20. Its kind of reversed in its behavior. The lower the level of ocs the more resonance drive it gets. Also the whole synth is way to loud. I just got my heart pumping fast because the loudness shocked me at the early morning. Though I now are wide awake. But Yutani will be fun, that's for sure. And I will further test this evening.

Greetings
Eli

[danerius]

Originally Posted by Eliseat

Meanwhile I had fun with FM Filter.

https://mega.nz/file/K6IjlQ6A#3tNbeb...XGx3QfSYumX6l0

You can turn the knobs like hell and it colors, colors, colors! Like crazy. Just amazing! Many thanks Saike.

Nice track

[sai'ke]

Originally Posted by Eliseat

Edit2: There is something wrong with the feeding of MS-20. Its kind of reversed in its behavior. The lower the level of ocs the more resonance drive it gets. Also the whole synth is way to loud. I just got my heart pumping fast because the loudness shocked me at the early morning. Though I now are wide awake. But Yutani will be fun, that's for sure. And I will further test this evening.

Greetings
Eli

Ah yes, I should definitely make the gain more in line with other synths. I think I've mostly been testing it in isolation, so that's an oversight on my part. I'll add a global gain control tonight.

I think the resonance behaviour is normal. FM Filter does the same, but maybe it's harder to notice. Driving it chokes the filter and makes resonance less apparent. It's always a balance between audio feed and feed from the feedback loop that causes the resonance. When there's little room, the signal starts to bunch up to the ceiling on both sides and the resonance is lost.

The drive gain is compensated (approximately) as it was in FM Filter. Pushing it harder makes the input signal louder, but the increase of volume on the filter input is compensated post filter. This gain compensation could probably be better. Right now, it's compensated assuming that the filter does not attenuate the amplitude. This is clearly not the case when you're hitting the filter hard.

It's not possible to calculate an exact gain compensation since it'd depend on the frequency content of your input, but I might be able to get a better approximation by assuming that the filter behaves more like a tanh saturator for instance (initially providing no attenuation, but attenuating when it is being driven harder).

I will look into this tonight

[Eliseat]

Saike, I believe the OSC outputs are way to high. This shows up, if you use the distortion module. It reacts way earlier than the FM-Filter version. Also the whole internal signal path doesn't behave like FM-Filter but a lot more sensible. That's just what my first thoughts are. If I turn don the OSC Level to about -25 it seems to be okay.

[sai'ke]

Originally Posted by Eliseat

Saike, I believe the OSC outputs are way to high. This shows up, if you use the distortion module. It reacts way earlier than the FM-Filter version. Also the whole internal signal path doesn't behave like FM-Filter but a lot more sensible. That's just what my first thoughts are. If I turn don the OSC Level to about -25 it seems to be okay.

A lot more sensible is good or bad? The signal path is actually identical, but the poor initial gain staging may be screwing things up here.

That's curious. I was going to put a gain reduction at the end, but maybe that's not the best idea then. I mean, especially considering most of the non-linearities in this thing are at the front.

The oscs put out at whatever dB setting they're set to (but they add of course). Which means that if you set them to 0 dB you're hitting the filter pretty hard. I will experiment with some other softsynths with their filter bypassed to see what kind of level they actually put out pre-filter at 0 dB.

I don't quite get the distortion statement. If you're using an NL filter, that filter should saturate and block high levels from even making it to the distortion module. Once at the distortion module, the levels should roughly be between -1 and 1. It may be that you're distorting overtones generated from the filters a second time though.

I'll doublecheck tonight and play a bit with the gain staging. I'll keep you updated

[Eliseat]

Here is a small improvised example of how fat and clean Yutani can sound. Pretty cool.

https://mega.nz/file/rvxWBSgK#PaS7Fu...7bRAxSc6XG5F1M

The gain staging thing is of course depending of what I used to put into FM Filter. I never got it as hot as 0db so this is maybe the discrepancy. With more sensible I mean its not easy to get into something usable. It seems that some parameter just take over the sound and don't allow other parameter like filter follow. If I get into the same situation I will make a video record to show you, what I mean.
I guess its all about fine tuning the relations between the modules. And I have no idea how the heck you even see anything in the code especially because those relations are dynamically related to each other. Many thanks Saike!

[sai'ke]

The clip sounds cool

But sad that it's difficult to find usable sounds with it. Let's see if we can found out what is causing it and fix that up.

I'd love to see a demonstration of that.

I'm curious what is different between how we are using this thing.

I've made a fair number of presets for it and I've never found it running excessively hot at those -6 dB levels on both oscs. But I did go for pretty aggressive sounds.

The routing looks like this (negative indicates that its the inverse of the earlier stage):

n x (Osc A + Osc B) => Gain(Drive dB) => Filter => Gain(-Drive dB) => Distortion Preamp (+ dB) => Distortion De-Emphasis (warmth) => Distortion => Distortion Post Amp (- dB) => Distortion Re-Emphasis => Output

The amplitude envelope can be either before the drive or at the very end depending on how you toggle that switch. If you put it before the drive, it will result in less controllable envelopes, but it's good for bassy sounds where you want to hear the filter "decay out".

In parallel:
n x (Osc A + Osc B) => LP => Mod(Filter Cutoff)

I'm wondering if that may be an issue. Does the lack of control happen with the self modes of the FM filter? I haven't used those very much yet on the synth, so they may be undertested.

[Eliseat]

This is how I open Yutani by default.

https://mega.nz/file/LzpjTYqb#zd2xOU...VrOn7ubdCq-THU

This is how its expected after leveling down the oscs to -20

https://mega.nz/file/vj5TgKBQ#_lybBO...oLPQOIGlhKreTY

Could it be a driver problem?

[Eliseat]

Originally Posted by sai'ke

The clip sounds cool

Thanks. But the real cool thing is how one pattern gets endless drama and color with Yutani. And this was just live improvisation. You could fine tune it with automation and it would lead you thru a whole indie song.
There was a time when I downloaded lots of demo synths just to see how they perform. But I have to say, Yutani is such a playful thing, it ranges above of most of the expensive competitors. Again. (Filther was too) Its just fun to use the knowledge of sound creation and really get unbelievable reactions. Yutani as an analog hardware synth would be the burner.

[Eliseat]

Originally Posted by sai'ke

I'm wondering if that may be an issue. Does the lack of control happen with the self modes of the FM filter? I haven't used those very much yet on the synth, so they may be undertested.

No, I would say it was just the overloud OSCs which drove the signal so hot that it wasn't possible to add or change anything anymore. If you've found and maybe fixed what causes this I wwill make a further test. okay?

Greetings
Eli (sweating in my room!)

[sai'ke]

Thanks, that's helpful

I've been shopping around for inspiration a bit for better oscillator scaling.

With only one oscillators active, Yutani comes in as hot as Synth1 (with two). This is as hot as Massive with the second osc on full and the first one off. With both oscillators enabled, Yutani is hotter. To get it to the same level, set them both to -9.0 dB. I've made this the new default since it fires up with two oscs. Monark is just as hot with two oscs running(seems to be a standard).

There's some other differences when it comes to oscillator volumes. The difference is that Synth1 allows you to alter the oscillator mix, rather than control the individual volumes directly. This makes it easier for them to not have to worry about this (or was it clever? )

Massive allows you set osc volume independently. The scaling on Massive (checked with oscillator two) seems closer to linear than logarithmic. Scaling on Oatmeal was logarithmic. Maybe making that scaling logarithmic on Yutani was a mistake and that may be another way to improve the synth.

Monark seems a little louder. Monark's scaling behaviour is really weird though. Over the halfway point, when you mix in more of Osc2, you see Osc1's volume drop to maintain a constant signal level of around -12 dB. Before the halfway point they behave differently and are less interconnected.

I pushed a new version out today.

I fixed some other issues as well.
- The feedback was not normalized properly, so that also nuked the input volume when you had it active.
- The bonus oscillators added to the total signal volume. Another trap for getting too much gain overall. Now it renormalizes the total output whenever you engage them, which should be more similar to what you'd expect from a unison.
- I made the defaults -9.0 dB as mentioned before (still seems quite separated from the -20-ish that I see in your demo though).
- I added a little drive indicator next to the filter selection, so at least you can see when you're going in hot.

Edit: Derp. Ignore all that, I had wet/dry in my test project set to something stupid. I was off by a factor of four. Gain staging should be a lot better now.

[Eliseat]

Originally Posted by sai'ke

Edit: Derp. Ignore all that, I had wet/dry in my test project set to something stupid. I was off by a factor of four. Gain staging should be a lot better now.

Now its perfect! Also the feedback feels way better. Its something very unique which gives some sounds a nice soft feature.

Many thanks. I will make some video examples. Maybe its also a good feedback for you to see, how Yutani behaves.

[Eliseat]

So here is an example how Yutani behaves in live mode. Everything works like expected. Sometimes you get into dead ends especially if filter envelope and filter frequency don't fit to each other. (As you can see in the video when I turned the envelope into negative.)

https://mega.nz/file/C7gxxJxQ#nlbksF...lu31451m71W9VY

Though I have no clue what accent means. Is it kind of a velocity reaction? It did actually nothing obvious to me. More feedback will follow tonight.

Greetings
ELi

[sai'ke]

Sounds cool. I like the sound you got going somewhere in the middle, and the one when you bandpassed it.

Yeah, I can see how it could've been confusing at first with the default being so heavy into the overdrive. I do like how the filters sound when you drive them to the extreme though. I find the Moog nice in that extremely overdriven mode. Especially for very low bass notes it starts to choke and gargle. No more room for resonance though.

Accent is something I'm still messing with so don't rely on it staying the same yet . Currently it applies some distortion which is triggered by the envelope, but I want to change that.

The attack when the filter envelope is in 303 mode (did you notice the filter has two modes of operation?) makes sure that within that attack frame it goes to N% of its output. In that mode, if you play rapid notes in succession you end up with a higher value than you would've with individual well spaced out notes.

With accent, I want to boost the rate at which it increases, but not reduce the time it is increasing, meaning it'll increase extra with an accented hit.

See the following graph, you see the envelope growing over time, rather than resetting for each note. Then with the accent, you temporarily really bump it and then it'll decay back down, emphasizing that part of the melody or bass.

Disclaimer, the effect is much more subtle in reality

In the end, it's supposed to be something you use on a few notes that you want to give an extra boost.

Have you given the velocity modifiers a spin? I really like playing with those. Can make really dynamic sounds with it if you combine it with velocity programming on the notes:
https://soundcloud.com/saike/test-ve.../s-4Hbql4cnOgI

[Eliseat]

Yes. Right before I read your post I just noticed this little envelope button for the first time. And its effect is amazing. As you see I was bumping the +amount to 100% because I needed a higher starting point. Now its more the way I like it with the 303 mode. Nice feature.

May I ask why you separated pulse and square? Because most monosynths just use the square as a base and let you shift the pulse width.

One thing I noticed is, that if I start a sequence it sometimes blocks the second OSC in the first loop and adds it with the beginning of the second loop.

Any chance of detuning LFOs? Would be nice because of syncing. And any chance of shifting the LFO in time and allow more than sinus? And by the way, I could go on with feature requests.
But I know its a lot of work for you and if its really needed ...? I think Yutani already has its own character and possibilities.

And the accent feature seems a bit mystical to me. But I'm excited when I'm allowed to test the final version of it.

Thanks
Eli

[sai'ke]

In the 303 mode you need much higher amounts. That will be the mode that has most use of the accent feature. It's meant for acid basslines where you often have a rapid succession of notes.

There's no real reason I couldn't put PWM under the square option. I figured it might be nice to quickly toggle to square just to check the contrast with PWM when auditioning. I also thought it might be somewhat confusing, since I don't support PWM for saw or tri. Although, you can technically already get it by setting osc1 to pwm and hardsyncing a saw or tri with it on osc2. You think it'd be better to have it under square and remove PWM?

In PWM mode you can modify the pulse width by dialing the phase knob. You can also automate this to velocity for instance.

What do you mean by detuning LFOs?

For the LFO, yes, I think starting phase, shape and toggle-able reset on note would be nice to haves. They will probably be added at some point. I can fairly easily nab these from Filther.

I was also thinking of adding a noise "oscillator", to be able to add some brightness when needed.

Right now, you have an amp and cutoff envelope, cutoff and morph LFOs. I was thinking of also adding one more "modifier" like VEL for a subset of sliders that works from an envelope. The problem with these more "dynamic" envelopes is that they're not free CPU-wise, since they require parameter updates per sample, so I want to limit how many can be done.

Is there anything you feel is really sorely missing?

[Eliseat]

Important! A double click reset for the vel settings is missing. I used the OSC tuning knob in vel mode and it wasn't possible to reset to zero. So it was out of tune forever.
But I'm actually testing the velocity mode, and it is - like you said - a very cool feature. Video example will follow. Of course with a nice new bass pattern. Be excited. This time it gets really hard.

The square OSC is kind of weird to me, because it doesn't sound very musically. I have no idea why. But i guess you use a 50/50 pulse width as default which sounds more like a humming cable. If you compare it with a 303 or even synth1 you have a more bassy and softer pulse width as default. As you also did with the PWM OSC. So only for me it is not necessary. I won't use it.

The fin OSC has in my opinion a way to low level compared to the others. But it sounds very cool.

A mode to use the tuning OSC for 1 semitone course steps would be nice. (I like your tiny green knobs.)

With tuning LFO (okay, that was misleading) I meant LFOs for the OSC tuning. But I keep my mouth closed because this would end in an feature orgy.

But noise ... would be cool too.

Meanwhile I'm really addicted to Yutani. And again its just a very cool name. Is there a meaning to it? I found several companies and names with "Yutani". Anyway. For me it is pure fun to play with it.

[mplay]

Just putting it out there, a multiband compressor based on tight compressor would be such a great tool to have. Recently I've been watching how EDM mixing and mastering legend Luca Petrolise uses a multiband compressor (fabfilter pro mb, which I can't aford atm) and have made a similar setup in Reaper. It works ok, but required quite a bit of hassle to set up and I had to use TDR Nova as tight compressor, nor reaxcomp have a sidechain and reacomp doesn't have ratio below 1. TDR Nova has no knee control or really fast attack, so it's a bit of a compromise, but I'll be happy to share the setup. Anyway, his workflow is as follows:

Using upward compression for the low and high frequencies, and actually using the low end signal as a sidechain for the high frequencies. This creates a really nice movement. Both high and low frequencies where mixed in with the dry signal. The mids compressed downward, again mixed in with the dry signal. The mid and high frequencies where also processed in mid side, being much more aggressive on the mids only a touch of compression on the sides.

It's in the second half of this video:
https://www.youtube.com/watch?v=dVmbgTFaPtI

[Linspace]

Hello,
I am trying out your Analyzer as it is something like I was looking for, but can't quite figure out how to use it. Am I correct in thinking that you should be able to show multiple tracks on a single analyzer window? If so how can I do this?
Thanks.

[Win Conway]

I haven't installed it, but first thing that jumped out as soon as I looked at the UI, why is the filter selector in the drive section ?

Edit*
COud you also add simple Osc FM, it only takes a two oscillator FM to create the very common vocal bass for instance (Triangle modulating a sine)

[sai'ke]

Originally Posted by mplay

a multiband compressor based on tight compressor would be such a great tool to have.

Unlikely. It's a cool video and a cool technique he shows, but making a transparent multiband compressor/expander suitable for mastering purposes is really not easy and a considerable amount of work. Both on the audio as well as the UI side. Simply splitting it into bands with crossover filters would induce really bad phase distortions, which you'd start to hear once you start compressing the individual bands. I would keep an eye out on Reeq, since I think he's planning some dynamic EQ features sometime in the future.

Originally Posted by Linspace

Hello,
I am trying out your Analyzer as it is something like I was looking for, but can't quite figure out how to use it. Am I correct in thinking that you should be able to show multiple tracks on a single analyzer window? If so how can I do this?
Thanks.

Yes, you can. Route the audio to the different input channels of the plugin. Basically make a new track with the analyzer, and then in the routing menu, send the audio from all the tracks to the track containing the analyzer.
Track 1 goes to audio 1+2, track 2 to audio 3+4 etc.

There's also a script that does this for you automatically. If you've added the github repo to your reapack list, it's called "Multi-channel spectral analyzer". It adds a script called LoadSpectrum.lua. If you run that script (it's under actions) it will add all the top level tracks to the analyzer.

Originally Posted by Win Conway

I haven't installed it, but first thing that jumped out as soon as I looked at the UI, why is the filter selector in the drive section ?

Edit*
COud you also add simple Osc FM, it only takes a two oscillator FM to create the very common vocal bass for instance (Triangle modulating a sine)

Yeah, that's a good point! It is a bit weird.

That whole section should probably be called FILTER, since it's all related to the filter.

Getting alias free FM modulation is not easy and it'd be taking the synth in an entirely different direction. For FM stuff, I think you're better off with something like Dexed. I'm not ruling it out, but for now, I want to get the stuff that's in there now polished first.

[Win Conway]

Nah Dexed is sine based, no triangles, It's no bother, I have plenty of synths that do it, Just thinking where I would go for presets for you on a two osc synth, my area is heavy Neuros and Reeses, so I will see what I can come up with when I get it installed.

[sai'ke]

Originally Posted by Eliseat

Important! A double click reset for the vel settings is missing. I used the OSC tuning knob in vel mode and it wasn't possible to reset to zero. So it was out of tune forever.

Added!

Originally Posted by Eliseat

But I'm actually testing the velocity mode, and it is - like you said - a very cool feature. Video example will follow. Of course with a nice new bass pattern. Be excited. This time it gets really hard.

Cool

which sounds more like a humming cable

Sounds like that should go in the review.

The fin OSC has in my opinion a way to low level compared to the others. But it sounds very cool.

Fixed.

A mode to use the tuning OSC for 1 semitone course steps would be nice. (I like your tiny green knobs.)

Planned...

Meanwhile I'm really addicted to Yutani. And again its just a very cool name. Is there a meaning to it? I found several companies and names with "Yutani". Anyway. For me it is pure fun to play with it.

It's a movie reference. It was a cybernetics company in the Alien universe. I also considered Weyland, but that might've been too on the nose.

Originally Posted by Win Conway

Nah Dexed is sine based, no triangles, It's no bother, I have plenty of synths that do it, Just thinking where I would go for presets for you on a two osc synth, my area is heavy Neuros and Reeses, so I will see what I can come up with when I get it installed.

Ah fair enough. I'll have a look at what's possible. I could make a "no handlebars" mode where there are some extra options that may alias and do all sorts of not so nice things. Considering you normally filter the hell out of basses, maybe the aliasing ain't so bad in those cases. I think doing a cheap implementation of it is easy. A clean one that doesn't alias much is a different story.

[mplay]

Originally Posted by sai'ke

Unlikely. It's a cool video and a cool technique he shows, but making a transparent multiband compressor/expander suitable for mastering purposes is really not easy and a considerable amount of work. Both on the audio as well as the UI side. Simply splitting it into bands with crossover filters would induce really bad phase distortions, which you'd start to hear once you start compressing the individual bands. I would keep an eye out on Reeq, since I think he's planning some dynamic EQ features sometime in the future.

Thanks for your reply! I am using reafir to split the bands onto three different track to minimize phase issues, but I can safely say you know much better than me how much work would have to go in a transparent mb comp

[sai'ke]

Originally Posted by mplay

Thanks for your reply! I am using reafir to split the bands onto three different track to minimize phase issues, but I can safely say you know much better than me how much work would have to go in a transparent mb comp

Just as a test, try switching out ReaFIR with my 4-pole bandsplitter in linear phase mode (enable the FIR button). If that quality is good enough (and the CPU use acceptable), then maybe it's doable.

I currently don't have an expander in my repo, so I'd also have to look into modding the tight compressor to allow for that behaviour. Something I don't have terribly much experience with.

I'll do a quick investigation on the weekend on how doable all the parts are.

[mplay]

Originally Posted by sai'ke

Just as a test, try switching out ReaFIR with my 4-pole bandsplitter in linear phase mode (enable the FIR button). If that quality is good enough (and the CPU use acceptable), then maybe it's doable.

I currently don't have an expander in my repo, so I'd also have to look into modding the tight compressor to allow for that behaviour. Something I don't have terribly much experience with.

I'll do a quick investigation on the weekend on how doable all the parts are.

CPU usage with your splitter is definitely good, goes from 0.3 to 0.5% on my old i7. My knowledge as far as linear phase vs regular phase filters is basically to listen to how linear phase is affecting the transients and if regular phase filters are causing phase issues. The dynamic processing I'm doing with Nova is with very gentle slopes, so even when mixing in the dry signal I'm getting pretty transparent results to my ears. As mentioned, its hard for me to estimate how much headaches implementing all this would be for you, but it definitely is a tool we don't have in reaper yet

[FeedTheCat]

Originally Posted by mplay

It's in the second half of this video:
https://www.youtube.com/watch?v=dVmbgTFaPtI

Really interesting! I wanted to try it out, but with dynamic EQ to avoid phase issues. Is there a drawback of doing it like this? Is it even the same thing conceptually? (The modulation input source is a low-passed version of the signal)



Btw, I'm really excited that you're making a synth Saike

[Eliseat]

Hey Saike, thanks for the fast response. I already mentioned it but now it is reproducible. If I start Reaper and let a song with Yutani play, it only plays the OSC sync but not the real OSCs UNTIL I open the GUI one time. After opening the panel or if I open the panel before starting to play, everything sounds like expected. If you like, I can give you a video example.

Greetings
Eli

[sai'ke]

Originally Posted by Eliseat

Hey Saike, thanks for the fast response. I already mentioned it but now it is reproducible. If I start Reaper and let a song with Yutani play, it only plays the OSC sync but not the real OSCs UNTIL I open the GUI one time. After opening the panel or if I open the panel before starting to play, everything sounds like expected. If you like, I can give you a video example.

Greetings
Eli

Thanks! I know where to look. That report should be sufficient to reproduce it. I will investigate tonight

[mplay]

Originally Posted by FeedTheCat

Really interesting! I wanted to try it out, but with dynamic EQ to avoid phase issues. Is there a drawback of doing it like this? Is it even the same thing conceptually? (The modulation input source is a low-passed version of the signal)


Btw, I'm really excited that you're making a synth Saike

Definitely another way to have the highs and lows pumping, but of course its only a small part of what he is doing, You really want to squash the mid-range frequencies to put them in your face for that modern pop/edm finished sound . Dynamic EQ's are also not immune to phase issues btw

[sai'ke]

Originally Posted by Eliseat

Hey Saike, thanks for the fast response. I already mentioned it but now it is reproducible. If I start Reaper and let a song with Yutani play, it only plays the OSC sync but not the real OSCs UNTIL I open the GUI one time. After opening the panel or if I open the panel before starting to play, everything sounds like expected. If you like, I can give you a video example.

Greetings
Eli

Hm actually, could you give me a preset for which this happens and maybe the video after all? I can't seem to reproduce it.

[Eliseat]

First things first. Here is the Vel mode play. Its such a nice thing. Really, Saike. So much power from clean to dirty, and everything is there. I wish I would have had this thing back in the nineties. No joke. If you ask me whats the best mono soft synth. The answer is Yutani.

Well done!

https://mega.nz/file/bmhjUZQA#-mXr4R...yzu3LjDSlZNnqA

Many many thanks. And I have no idea why nobody jumps in. Maybe you should open an extra thread for Yutani. But be aware, this would lead into a feature request battle.