[Latest Beta] SubLab 1.1.1 (build date 30/8/19)

Hey everyone,

We’ve a beta of 1.1.1 ready for testing. The main changes are as follows…

Change-log since 1.1.0

  • Removed clicks from SubLab’s attack phase when “reset phase” is turned on via the main top bar settings menu. Having the oscillators phase reset on a all new notes on should now be click free.

  • Allow importing of NKS sound packs. If you install a pack pack bought from our site, it will install the correct NKS files and audio previews for use in Maschine and Komplete Kontrol.

  • Fix issue where SubLab needs to be activated multiple times. This is happening to some windows users when they activate using a removable network card. We now use a different ID when activating which is not dependent on your computer being connected to the internet or your network card plugged in.

  • Update default preset so that legato is turned on by default.

  • Add resolution setting for the spectrogram, which will reduce down the amount of painting the graphics card needs to do. Some users were experiencing crackles in SubLab, by reducing down the resolution of the spectrogram will stop this happening.

  • Added “import sample” item to sample menu, as an alternative to drag and drop of samples.

  • Fixed issue where SubLab VST was causing a crash in Cubase for some users.

Anyone who wants to beta test, just send me a mail (support@futureaudioworkshop.com) and I’ll forward on the download links. If sending a mail, make sure to use the same email address as your FAW account.

Thanks again and looking forward to getting this version checked and out!

Also, every time we post a fresh beta, we’ll delete all the older comments related to previous versions.

Gavin.

6 Likes

Early tests and observations:

In Windows 10, for the VST3 version, the “about” panel in Sublab shows 1.1.0 and not 1.1.1, but it does show the Aug 30 build date and clearly has all the new features.

Transient clicks in the X-Sub do seem to be gone when configured to restart phase with every note ON. Yay!

I got the VST3 version to crash a couple times while testing phase alignment. Granted, I was running TWO instances of Sublab in the same VST Host Process, which is unusual, but I needed two instances to be able to view the Sampler waveform separate from the X-Sub waveform in a dual-channel oscilloscope.

This was in Bitwig Studio, which has several VST sandboxing modes, and I use the mode called “By Manufacturer”. This mode runs all plugins from the same manufacturer in the same VST Host Process together, so that inter-plugin communication can occur (think iZotope stuff or Fabfilter stuff, etc.). But if any plugins from that manufacturer crash, they only crash that one manufacturer’s host process and none of the other plugins in the project are affected by the crash (because every manufacturer is sandboxed into independent host processes).

Anyway, might want to test your VST3 implementation a wee bit in Bitwig? And maybe check in with the Bitwig devs about any VST3 support guidelines they might have for 3rd party developers? I never had stability problems with the first two versions of SubLab, but then I never ran more than one instance per project, either.

I’ll try to do a similar test with VST2 Sublab tomorrow and see if they’re more robust and crash-proof with two instances active.

All great testing. We updated the version number, so the fact that it is not appearing is good to know, we’ll sort that out before release.

We’ll have a look at the VST3 in Bitwig, thanks for the heads up and the explanation of what is going on. We’ve done a good bit of internal work around an area that could be causing this crash, this is exactly what we need to know :slight_smile: Let me know how it goes with the VST2.

Thanks!
Gavin.

@Baphometrix, figured out what was up with the version number, seems our installer builder was caching an older build of the beta that didn’t have the updated version number on the about page :man_facepalming:

I’ve swapped out the beta download link with the updated installer, if you just download again and install you’ll have the version with the correct installer. I’ve been testing here in Bitwig and haven’t been able to recreate the crash that you experienced. Can you give me some exact steps you are doing in order to have it crash?
Thanks,
Gavin.

Hi Gavin. Okay I can reproduce the crash behavior on video. To reproduce the crash behavior, it seems to revolve around having TWO Sublab windows open and working in each one, then closing ONE of them and then trying to do something in the remaining window. The crash, when it happens always revolves around that attempt to interact with the remaining window after closing one of them. This happens consistently in both VST2 and VST3. Here’s a video link to the timestamp where I start making this happen consistently.

Reproducing the crash behavior:
youtu.be/aL9-Wr3INuU?t=859

(Hey since you don’t allow links in posts here, just prefix “http://” to the front of that string above an you’ll find it.)

1 Like

And during the video capture of the crash condition, I discovered that you STILL have transient clicks and pops in both the Sampler and in the X-Sub when Phase is set to restart at note ON. The phase itself is stable, but the waveform is starting at various random places well away from the zero-crossing. Also, the waveform seems to consistently start with a move into the negative DC range, which is generally a no-no if your goal is to layer sounds (you want all sounds starting with a positive DC move, in general).

In the video, shortly after the timestamp below, I demonstrate how if you make a sub in Serum it will ALWAYS start from the zero-crossing, and it will ALWAYS start with a positive DC move. Both of which are the desirable behavior, especially for subs that might need to be phase-aligned with kicks. I mean, I did an entire deep-dive producer tutorial on this subject just last week.

Transient pops because phase not restarting at 0 crossing
youtu.be/aL9-Wr3INuU?t=527

(As in the previous comment, prefix “http://” to the front of that URL to make it work.

1 Like

Very useful @Baphometrix.

Looking at the video, starting at the linked t=527, this might need a bit of deeper explanation of how the voice works. When a new note on comes in, the voice starts up and the oscillators in X-Sub and Synth engines keep running until all the envelopes are finished. The envelopes are classified as finished when the level they are at during release decreased to <0.00001. Nothing strange there. Once the envelopes are finished, we totally stop processing the voice. What this means is that if the oscillator’s phase was at say 82 degrees, it freezes at this value. Then when a new note comes in, we start up again. In the previous version 1.0, we would start the oscillator at 82 degrees as this was the value the oscillators stopped at. In 1.1.0 we reset the phase to 0 prior to starting up again. However, if a new note comes in, while the voice still hasn’t finished, we don’t reset the phase, we just re-start the envelope at the attack stage and take the last value the envelope was at as its starting point and off it goes again climbing towards one. The oscillator shouldn’t be touched.

Looking at your waveform, the switch in direction in the waveform in the image below should not be happening. It looks like you’re in either the release or decay stage, and while the volume should start to increase with the new note on as you’re back in attack, there is a reset or a glitch in the phase. I’ll need to verify this here, but it looks like there is something up, X-Sub is changing its direction, even if the voice is still playing when the new note on comes in. There are a couple of edge cases that might cause a waveform like this, say a release time of zero so the voice finishes, oscillator reset very fast and starts again, but I don’t think this is what is happening here.

I’ll verify X-Sub and will have a look at the sampler. :+1:

On the crash, we’re still not able to recreate this here, but will keep investigating. Just to make sure, can you let me know the build time and also the version number you see on the about page.

Even if the AAX plugin component is deselected during installation files at C:\Program Files\Common Files\Avid\Audio\Plug-Ins\SubLab.aaxplugin are created by the Windows installer.

I got email with latest versions yesterday. Mac download is 1.1.0 beta 1

Hi Gavin, I understand exactly what you’re describing about the intended voice handling for the oscillators in SubLab. We have a similar distinction over in Bitwig. One of the voice handling modes is called “True Mono” and works like you’ve described (each new gate ON will restart the attack at the same exact amplitude that the release is currently at).

And yeah, if the waveform weren’t suddenly changing direction at the same time, that would sound very smooth and not result in any noticeable transient “click”.

BTW the release time I was using was in the 230 ms range, and it also happens with a much longer release of 3.0 ms… Likewise, the attack phase was something like 9.6 ms.

VST3: 1.1.0 30 Aug 2019 10:45:00am
VST2: same

As for the Sampler, the video shows the same transient pop behavior there too (looks like the waveform suddenly reverses, and also tries to start at the same amplitude as the current release value. But you can see in the following screenshot that the original sample always starts from zero-crossing and starts in a positive DC direction.

Interestingly, SubLab only does this “direction reversal” some of the time, but not all of the time. Here I caught two in a row, but other times I could go 4 times in a row with no reversal. It seems to be a random glitch of some sort, at least in the Sampler osc. I think over in the X-Sub osc it’s more consistently reversing like this.

1 Like

@Baphometrix

We’ve got the VST 3 crashing issue fixed and also the phase issue with X-Sub.

I’m going to put a build together and send on the links in the next hour or so. Once we get this verified as working can dive into the sampler!

Gavin.

Hi @Gavin_FAW please send me the .exe once done, I have a week off and I can deeply test it
Thanks

1 Like

Yup, will do Andy! :+1:

Looking like tomorrow when we’re able to push out the build, hang tight there guys and really appreciate the help testing :slight_smile:

Gavin.

2 Likes

Quick update, the new beta is ready and we’ve just sent it out.

Thanks,
Gavin.

Don’t see the 1.1.1 beta?

Regression: All X-Sub C notes are now playing C0 (32.7 Hz) instead of C1. This is a bad regression. Even the mighty PK “Gravity” subs used at major festivals start losing power at 34 Hz. And your more typical subs at clubs? Not even close to that. Home/Studio subs? Even worse.

Bass music producers typically won’t use a sub note lower than E or Eb at the very lowest. With F and G being the most popular keys to write in by far because that’s the lowest you can count on the wide majority of subwoofers out there (including at clubs and many festivals) to play a clean sub frequency at full power and not come out sounding like mushy rumble.

Yes, I know there’s plenty of Trap music written in C, but they can get away with it because trap doesn’t rely on clean sub-bass sounds. Trap songs written in C are using very dirty, distorted 808s that are chock full of harmonics that hide the fact that there’s not really any power output down at C0.

If you put 1.1.1 into production with the X-Sub transposing all C notes to C0, you’re going to become essentially useless to anyone except trap producers. Why do I feel strongly about this? Because songs written in F (a very popular key for bass music) will use C a lot as the 5th of F. And we want that C hitting at C1 (65.41 Hz), just like the X-Sub has been doing all along until now. And songs written in G will also use C a lot as the 4th of G.

If you really want to truly satisfy ALL audiences, you need to make the X-Sub transposition range user-configurable. And frankly, you also need to make the transpostion behavior user-bypassable. Until such time as you can make it user-configurable at the very least, and also ideally user-bypassable, then I strongly recommend that you put the C note back to playing at C1. We can handle the D note playing down at D0. While that’s still going to be played with lower power on most subwoofers everywhere, it’s at least not rumbly mush at D0 on most subwoofers.

I’m serious about this. SubLab becomes essentially useless to me if I can’t use it mostly for fairly “clean” sub-bass sounds. But if that C is going to play down in the powerless “rumble” range, I simply can’t use Sublab.

BTW here is a diagram that the vast majority of electronic music producers would agree on. Especially those who produce more aggressive bass-oriented genres for the festival (and large club) crowds. It’s ranged for one of the DAWs that puts middle C (266 Hz ) at C4, so transpose accordingly. Notice that C, C#, and D are considered the “weak” notes in the sub range. But of the two possible Cs, the higher of the two is WAY better than the lower of the two. Same deal for the C#. On the other hand, if you really need to use a D note in your sub riff, most producers would choose to use the lower D.

On the bright side, you did get rid of the transient click on the X-Sub, and the oscillator shows it behaving in a predicable and desireable way, so yay.

However, the Sampler still does that wonky reversal of the waveform and still makes a very noticeable transient click most of the time if you retrigger while a previous note is still in its release phase.

Hey @Baphometrix,

This is all what we’re trying to do! We can infer what people want prior to release or we can release and then as people use it and give feedback adjust how things work. We spent a lot of time on X-Sub and there are conflicting ideas and feedback that comes in all the time. On the which C to play, we will make this a user setting. One idea is to also switch depending on the direction: if your previous notes are in a downward direction, use C0, if they are climbing use C1.

I’ll add a setting to setting.xml where C0 vs C1 becomes a global setting in v1.1.1

We’re going to have to release 1.1.1 in the next days and then we’ll move straight into 1.1.2 and the sampler.

Thanks Baphometrix!
Gavin.

1.1.1 Beta 3 Regression A:

X-Sub phase is drifting again, even when explicitly set to restart phase ON.

1.1.1 Beta 3 Regression B:

SubLab window will randomly resize itself to smallest dimension possible when you close the focus window for the VST and re-open it. It is semi-random and this started happening in 1.1.1 Beta 2. In Beta 3 I’ve gotten it to happen 3x within 2 minutes across two different instances of SubLab in the same project.