problem with transmission patch

[gerrit]

Hi there,
I'm a zebra user since the first version. Every upgrade i made was nothing but very pleasing. But now with version 2.2 i found at least one sound of the transmission sound set, i.e. classic synth - circuit bent cs01, having a distorted out put. I checked against version 2 on my MBP and found everything there like it should be, smooth sound.
So, my question. Can anybody reproduce this finding? ( i didn't check all patches, it's just this one and another one which i don't remember.)
I didn't do the update on the MBP so i cannot tell if it is PPC only - in fact i have this problem only on my G5 2.3 GHz.
Looking forward to hearing from anyone.

[Urs]

Oooph... I'll try to reproduce this... I'm currently eaten up by subtle bugs on the other platform though...

One thing that occurred recently was, very fast modulations from LFOs are now stronger because I have unfused the lowpass filter a bit... it may be possible that some LFO amplitudes in some sfx patches should be reduced a bit...

I'll check it out...

Later,

Urs

[beej]

Ouch - yeah, just tried it here on a PPC G4 - distorted horribleness.

It *definitely* isn't supposed to sound like that. (Unless Zebra sees the text "circuit bent" in the name and enters a special circuit bent software mode... )

It goes away if you bypess the VCF2 module (LP Exite mode).

It stays distorted if you switch to any filter mode. If you remove VCF2, and put in a different filter block, the sound is fine. If you put VCF2 back instead, it distorts again. Seems to be an issue with VCF2 - more specifically the modulation of it from MMix1 - some real oddness going on as you tweak the mod amount knob in VCF2...

[Wanterkeelt]

Problem also happens here in Windows, FL Studio 7.3 Beta, and the latest Z22_WinDebug

[gerrit]

Hi there,
i just found another patch, which sounds awfully distorted:
leads - orange sync lead. Please check it out.
Cheers

[bmrzycki]

I reported something similar to this during the beta. The thread is here:

http://www.kvraudio.com/forum/viewtopic.php?t=191094

Maybe the two are related?

[bmrzycki]

I found another preset that sounds horribly wrong: Modern Synth/Synthzilla.h2p. It maxes out my digital volume meter, the preset's compressor threshold reduction bar is also heavily used.

I'm on my Macbook, Leopard, AU Lab 2.0.0.

[bmrzycki]

I just noticed every preset mentioned with audio problems both here and in my original thread all use the MMix1 modulation source for filter cutoff. All of the presets use the same MMix layout:

Type: sum modulations
Mod1: LFO2
Mod2: ENV2

LFO2 is always set the same way:

Amp amount changes a bit, but rate is always set to 200.00.

If I use just LFO2 as the modulation source for filter cutoff it still sounds terrible. Did the LFO code change a bit from 2.1 to 2.2?

[Urs]

Did the LFO code change a bit from 2.1 to 2.2?

Yes, it's more snappy. It used to have too much lowpass/smoothing. By taking away some of the cutoff, the amplitude is higher on fast modulations (like, high frequencies are louder).

Hmmm.

Urs

[beej]

So, what's the verdict on this - is the new behaviour in 2.2 a bug that needs "fixing", or has the behaviour simply changed meaning that changes need to be made to all patches using this feature?

Have you had a chance to have a look into this yet Urs?

[Howard]

So, what's the verdict on this - is the new behaviour in 2.2 a bug that needs "fixing", or has the behaviour simply changed meaning that changes need to be made to all patches using this feature?

It's better now than it was before (more flexible). As only a few patches seem to have been affected, I reckon it's worth it (sorry beej) - even if 20 of my patches had been affected, I'd be quite willing to do the updating.

BTW: I've had several patches TOTALLY ruined by OS updates (Waldorf Q and CS80V). This change in Zebra is mild by comparison.

[Urs]

I think I'd love to keep the change. It can't be many patches that use LFOs in the audible range?

I initially had the idea to make the "smoothing coefficient" available as a knob, but I think that's too nerdy.

What do you think beej? How many patches are affected?

Urs

[beej]

I think I'd love to keep the change. It can't be many patches that use LFOs in the audible range?
What do you think beej? How many patches are affected?

Dunno, need to do some stock taking to understand what's going on and find an acceptable solution. It's probably not more than a handful of core patches, but of course there are also lots of variations based on those core patches, so it's probably not a trivial thing.

Of course, if it was a "bug" that was going to be fixed, the work wouldn't be necessary. As it seems like a change in behaviour that's basically desirable, that changes things somewhat so I need to do some investigation as how to proceed...

[bmrzycki]

Dunno, need to do some stock taking to understand what's going on and find an acceptable solution. It's probably not more than a handful of core patches, but of course there are also lots of variations based on those core patches, so it's probably not a trivial thing.

Beej,

I found one that's a bit more subtle, but still incorrect for 2.2. Acid Burns in the Soundscapes.


Urs,

This still doesn't root cause why heavily modulating cutoff in the "Enter the Tardis" preset still causes the Audio to shut down and need resetting. I don't think you'd ever want that to happen.

[Urs]

Yes, I think the new functionality is better...

I think it roughly boils down to

if (LFO speed faster than 1/32) then (LFO amplitude *= ~1/2)
if (LFO speed faster than 1/64) then (LFO amplitude *= ~1/3)

Or, alternatively use sine instead of square as lfo waveform...

I still have to find out why some filters go nuts though... that's another issue...

Urs