TONEX 1.5 has been released

[Funkybot's Evil Twin]

Jamcat, try this simple test:

1. Using TONEX 1.2.6...
2. Put one instance of TONEX on a track - load a patch - disable the cabinet
3. Put a gain plugin like TrackControl after that instance - why? Because reverbs and other time-based plugins may not null in a test. A gain/utility plugin allows for a better test.
4.Put a second instance of TONEX on a track - load THE SAME patch - disable the amp block
5. Render to audio

Then...

6. Update to TONEX 1.5.1 and repeat the same steps above.
7. Line up the two audio tracks and invert the phase

Here's my guess: they'll null completely. They didn't make any changes to the sound of TONEX. If they don't, post the audio here.

Our brains are not reliable when it comes to interpreting things we hear. This would be more of a scientific test that would rule out user error or fuzzy recollections. I think you'll find the sound didn't change. And if you did, then report it to support. Post the audio here on KVR! Lots of us are using TONEX 1.5 and you're the only complaint I've heard about the sound changing.

I haven't updated on my Mac yet, so may be able to do it there at some point over the next few days, but "the sound changed" is an easy thing to prove - also an easy notion to disavow yourself of by doing a proper null test.

[zvenx]

You had me up to 1.5.2... I assume you mean 1.5.1?
rsp

[Funkybot's Evil Twin]

You had me up to 1.5.2... I assume you mean 1.5.1?
rsp

Yes. Wishful thinking on my part.

I'll edit the post.

[jamcat]

Jamcat, try this simple test:

1. Using TONEX 1.2.6...

6. Update to TONEX 1.5.1 and repeat the same steps above.

Here's my guess: they'll null completely. They didn't make any changes to the sound of TONEX. If they don't, post the audio here.

I tried v1.5.1. The bug is still present.

So here is what I have done:

I have many songs with TONEX. I always split the amp and cab over 2 or 3 instances. It's the reason I use TONEX now and not AmpliTube anymore.

TONEX is mono, and I'm usually creating a stereo split to 2 cabs. I often have Audiority's CE-1 or PSP SpringBox in between the amp and cabs, but that doesn't matter here. It happens with no plugins between 2 back-to-back mono instances, too. I always use the amp and cab from the same capture.

I know what my songs sound like. And after updating to v1.5.x, they sounded very different. And not in a good way. I have mixdowns of songs where the only difference is TONEX v1.2.6 or v1.5.x. It is immediately obvious. The first thing you notice is how much quieter it is.

After rolling back, the songs sounded correct again.

I was easily able to pinpoint the problem was when the amp and cab are isolated to their own instances. I'm not sure if they would null when using v1.2.6 in your null test. They sound the same, but it's possible splitting them may cause some phase misalignment.

But I can tell you for sure there is no way in hell v1.5.x would null when comparing amp & cab in one instance to amp and cab split over 2 instances. And there is no way in hell the latter would null with the same setup in v1.2.6.

[jamcat]

I did some further tests and I found that it happens with the VST3 version, but not the VST2 version.
It happens specifically in the cab-only plugin.

If you use the VST3 version for the amp-only instance, and the VST2 version for the cab-only instance, the problem is not present. If you use all VST2 versions, the problem is not present. It only occurs when the cab-only plugin is VST3.

This was not hard to test for. The situation is EXACTLY as I thought: someone threw on a couple of VST2 plugins, said "I don't hear it" and dismissed it. THAT is exactly what passes for "testing" at IK.

Funkybot, if you have a few moments, could you confirm this (on Mac) with the 1.5.0 or 1.5.1 VST3 plugin?

[zvenx]

Or maybe no one uses it like you do so that scenario was never tested?

Unfortunately I can't test it on mac since Cubase Silicon native doesn't allow me to use vst2 so I can a/b vst2 vs vst3 doing what you suggest. (I only have Cubendo 12 and 13 on this computer and putting it in rosetta mode alas will cause a lot of rescanning).
rsp

[jamcat]

No one uses it it the way I do?
You mean as a VST3 plugin on Mac?

Isn't testing supposed to be thorough, and on all formats/platforms?
That's what they're getting paid for, right?

I passed it on to a tester and they were not able to recreate it.

[zvenx]

The splitting..
it is the splitting part that may be unusual.. I have never splitted it for instance and I use Tonex vst3 everyday in vst3 when I am recording or working with electric guitars.
rsp

[jamcat]

Then you only do mono guitars. That's fine.

Anyway, no, you see as I added to my post above, someone allegedly tested for the specific problem I reported, according to Peter.

I passed it on to a tester and they were not able to recreate it.

Why is that?
Because they were not thorough. Because IK doesn't have real testers.

And why wouldn't a software company test to make sure their software works correctly?
There is obviously a bug with the VST3 version, specifically.

Just off the top of my head, perhaps disabling the amp causes a volume dropout or possible phase issue in the VST3 version?

[zvenx]

Ok.
I can only test fully on the PC since on mac, my Cubendo can't run vst2 in native mode.

So this is what I did and what I found.
I took the same mono guitar track and duplicated it three times (different tracks).
Track 1 tonex split, one amp and one cab vst3
Track 2 identical to Track 1
Track 3 one tonex with cab and speaker vst3
Track 4, two tonex, one amp and one cab but vst2.

Here is what I found.
Track 1 and 2, same level (I did this to eliminate any obvious non linearity in the plugin.
Track 1 and 2 are louder (like about 5db) than Track 3
Track 4 is same level with Track 3, but about 5db softer than Track 1 and 2.

My conclusion? Indeed it seems that at least currently (didn't test with Tonex prior to 1.5.1) splitting on vst3 makes it about 5db louder.

I have saved the test project and included it below if you want to send to them.

This is the test file..in Cubendo.. if anyone wants to try it or send it to ik
https://www.dropbox.com/scl/fi/nj67t9lk ... 12l99&dl=0


rsp

[zvenx]

There is no software company that will or even can test all permutations of uses of their software on all permutations of hardware.
That is just near impossible..
So many if not most try to find a diverse set of Beta testers who work differently, different backgrounds with different hardware.
rsp

[Funkybot's Evil Twin]

I suspect what's happening is that the VST3 version in whatever DAW is handling the channels differently which is why VST2 sounds normal. And keep in mind that the CE-1 has a pretty unique stereo handling thing going on where one channel is dry and the other channel is wet by default. That could be a factor as well. Repeat the test in an empty project with just a gain plugin in between. Then repeat the test in another DAW entirely.

[zvenx]

CE-1, how did that come into play?
Unless you are doing some surround etc.. for sure Cubendo is not suppose to handle vst2 and vst3 differently in terms of audio at all. At All .
rsp

[jamcat]

It doesn't matter if you use any plugins in between instance 1 and instance 2. Either way, it sounds different when you disable the amp in VST3 in v1.5.x than it did in all previous versions.

I had already tried a variety of variations -- mono, stereo, both with or without plugins in between -- it doesn't matter. The change is in the VST3 version (at least on Mac) in v1.5.0 and 1.5.1.

I am using Studio One 6.5 native in Apple Silicon. It supports VST3, VST2, and AU.

[Funkybot's Evil Twin]

I just did this exact my test on Reaper comparing TONEX VST3 1.2.6 and TONEX VST3 1.5.1 and guess what? Both versions nulled perfectly!

Here's my exact test:

1. Empty project (don't test it "in a song" - this is science, not art)
2. Insert new empty track
3. Add the following plugins: a) TrackControl (I use this instance for consistent gain into TONEX - you can skip this), b) TONEX (amp only - cabinet block disabled), c) TrackControl at unity gain, d) TONEX (cabinet only - amp block disabled) TONEX version 1.2.6, VST3
4. In both TONEX instances, I hit the "Reset Chain" button under Advanced Parameters to make sure I didn't accidentally have the EQ's set differently or accidentally leave the reverb on
5. Recorded guitar playing into the track
6. Bounced the audio to a new empty track
7. Inverted polarity on one of the tracks (doesn't matter which)
8. Hit play with both the "active TONEX track" and the bounced down track playing (remember, one has the polarity inverted, the other doesn't)

Result #1: absolute silence. So both tracks perfectly nulled with themselves.

9. Saved project exactly as is
10. Exited Reaper
11. Updated to TONEX 1.5.1

12. Launched Reaper
13. Opened the same project
14. Hit play

Result #2: absolute silence again. Both tracks still perfectly null even though the new version of TONEX is in use on one of the tracks, and playing back against the bounced down version from version 1.2.6.

Conclusion: if they sounded different, then there wouldn't have been a perfect null. Which means there's no way the VST3 versions sound different in Reaper.

Where that leaves us: either, Studio One is doing something different. Or Jamcat's testing methodology is flawed somewhere. Maybe there's some time variant plugin (modulation, reverb) that's causing them not to null and the results are being misinterpereted - I dunno. Could be a bug in just Studio One with the VST3 version - stranger things have happened. Could be a latency compensation bug in the VST3 version and Studio One. Who knows?

My advice for next steps: follow my exact steps in Studio One and try again. No other plugins. No other tracks playing. If they sound different after updating, send your 2-track test project to IK Multimedia support. If they sound the same...then they sound the same.