Newbie's experiences, VST list and questions...

Anything that doesn't fit into other topics goes here!
First, big huge thank you to LMMS and VeSTige developers and document makers! It's a huge value to Linux audio hobbyists!

To me it was a bonus that JACK is not needed... (Because I'm such a n00b I don't understand JACK.)

While I had problems ("vanished" VeSTige, needed compilation, etc.) now everything works AFAIK.

I installed *huge* mass of VSTs and to my great surprise, about 80% of them worked! Well, at least preset sounds worked, testing everything with modern synths would take ages...

Here's a list of most important VSTs I managed to get working:
Ambience, Angelina, Crystal, Dimension Pro, Diva(x), Diversion(x), Firebird, FM8(x), iloSynth, Kairatune, Massive(x), Oatmeal, Rapture, RhinoCM, String Theory, Synapse Dune, Synth1, SynthMasterCM, Triangle II, VivaldiMX.

Those marked with (x) were difficult to install or activate. (Yes, all are legal!)

Kairatune seems to be buggy, but it might be development version or something.

Those with large preset banks are quite difficult to use, it's often balance between VSTs controls and VeSTige's own controls. Also, if VSTs user interface geometry changes for some reason, Wine/VeSTige doesn't update it -> parts of UI missing, sometimes.

All in all I have probably around hundred VSTs more or less working, but those are most important.

(For those who need to know, this PC is i5-750, running Ubuntu 12.04 LTS 32-bit for Wine compability)

Couple of questions:

You really should document how to use Effect VSTs better, it wan somewhat unintuitive. :D However, I'm interested what kind of VSTs have any chance of working? It's often hard to find out by testing, because the test material needs to be suitable, and often the effect is subtle...

Another thing, about freezing track: does it somehow suspend the VST? If not, it's probably better to export a wav from the track.

But, all things considered, I'm insanely happy with LMMS now!
simple_impulse wrote:Kairatune seems to be buggy, but it might be development version or something.
Haven't gotten it to work at all myself. Just takes a huge CPU load and then shows the GUI but the GUI doesn't work, and the sound is just glitch.
simple_impulse wrote: You really should document how to use Effect VSTs better, it wan somewhat unintuitive. :D However, I'm interested what kind of VSTs have any chance of working? It's often hard to find out by testing, because the test material needs to be suitable, and often the effect is subtle...
No idea really. Problem is that the Steinberg VST SDK is proprietary, and thus LMMS relies on a reverse-engineered open source implementation to provide VST support, and even compounding this problem is the fact that the vast majority of VST's are closed source - so there's really no way to predict which VST's work and which don't.

Some VST's work flawlessly. Others don't work at all. And yet others work, but are only capable of producing 44.1khz sound, and when you try to render in a higher sample rate they sound out of tune. (You can somewhat workaround this by rendering at 2x samplerate and tuning the VST an octave down, but it may not work for every situation.)
simple_impulse wrote:Another thing, about freezing track: does it somehow suspend the VST? If not, it's probably better to export a wav from the track.
The freezing functionality isn't working very well at the moment - it ignores automation, and it's indeed better to just render the entire track if you're running low on CPU power. Although, LMMS doesn't currently work very well with sample tracks either (you can't play them from the middle), so you're left with bad choices...

Anyway it's good to remember that LMMS is a work in progress and many things are still being ironed out.
simple_impulse wrote:But, all things considered, I'm insanely happy with LMMS now!
Good to hear! Then you'll probably love the v1.0.0 release which should be out this month.
diiz wrote: Some VST's work flawlessly. Others don't work at all. And yet others work, but are only capable of producing 44.1khz sound, and when you try to render in a higher sample rate they sound out of tune. (You can somewhat workaround this by rendering at 2x samplerate and tuning the VST an octave down, but it may not work for every situation.)
That's good to know, even if Iäm using 44.1 most of the time.
diiz wrote: The freezing functionality isn't working very well at the moment - it ignores automation, and it's indeed better to just render the entire track if you're running low on CPU power. Although, LMMS doesn't currently work very well with sample tracks either (you can't play them from the middle), so you're left with bad choices...
Good Freeze in my mind would be irreversible rendering to wav. No Thaw possibility, also all plugins (native or VSTs) should be removed.

I thought the idea behind not playing from the middle was intentional... :D Anyway, if it becomes a problem, I guess Audacity is good help there.
diiz wrote: Good to hear! Then you'll probably love the v1.0.0 release which should be out this month.
That's very good news. When it comes out, I'll check that it works at least as well as now... Then I think I'll introduce it to some Linuxists. :D
simple_impulse wrote:That's good to know, even if Iäm using 44.1 most of the time.
Not a good idea if you plan on doing any post-processing in eg. Audacity... it's better to export at a higher rate and then resample/compress it later, as it leads to a better sound quality than if you just render directly at 44.1k.
But doesn't that cause high frequency problems? I'm just a n00b, but someone said that over 48 kHz there are problems which can only be solved with some super expensive systems? Sorry, but I'm constantly confused and I'm always trying only to use stuff I understand... :D
simple_impulse wrote:But doesn't that cause high frequency problems? I'm just a n00b, but someone said that over 48 kHz there are problems which can only be solved with some super expensive systems? Sorry, but I'm constantly confused and I'm always trying only to use stuff I understand... :D
Who said that? That's totally BS as far as I can tell.

Higher frequencies can cause problems on lower samplerates. 44100 hz samplerate can only play back 22050 hz sound, that means that any higher frequency sound will possibly be aliasing into a <22050 hz sound, which can cause distortions in sounds. This is why AD-converters use oversampling to mitigate aliasing.
diiz, you for sure understand these things better than I, I'm just a beginner.

However, I did see an article about ultra-high rate sampling and it said that Nyquist's Theorem predicts that 44.1 kHz is ideal sampling rate. It said that these ultrasonic sample rates introduce new forms of problems, which are expensive to solve.

Maybe you mean that these sample rates work like camera cells, taking several samples to produce one pixel concurrently? Then it would make a lot of sense, I think. (Wild guess... :D )

Or the alternative is that I'm misunderstanding the whole thing; wouldn't be the first time, especially with complex things... :D
simple_impulse wrote:diiz, you for sure understand these things better than I, I'm just a beginner.

However, I did see an article about ultra-high rate sampling and it said that Nyquist's Theorem predicts that 44.1 kHz is ideal sampling rate. It said that these ultrasonic sample rates introduce new forms of problems, which are expensive to solve.
Where is that article? Sounds very much like BS or misunderstanding to me. Or in the worst case scenario, some snake-oil peddler trying to sell their own expensive "solution" to an imaginary problem... after all, we have people selling HDMI cables with built-in "virus protection", and people actually buying them...
Maybe you mean that these sample rates work like camera cells, taking several samples to produce one pixel concurrently? Then it would make a lot of sense, I think. (Wild guess... :D )
That's basically how oversampling works.
simple_impulse, you list Crystal as one you have installed succesfully. Have you tried open a project where a Crystal-patch is used? Or (worse), one you have created yourself?
In my project, Crystal is then cleaned. Not a single setting is preserved. In fact all is muted. Crystal works perfectly ad a 'play-a-long' vst, but it cant be saved (Win-xp32 0.4.15)

On vsts effects - the new version looks like it support this in much better way. I can use nyquest and karma on the p.r. -I think its much improved.


(btw.. ehh Massive? is that leagal freeware now? size]