Mixed impressions on u-he Zebra

[TiUser]

If you don't like what you hear...

That's exactly the point I am not sure about... My perception is there shines a lot of really great sound through a "curtain" I can't explain. I'm not about certain patches or programming or tweaking them, that's not the point and that wouldn't change anything. I am about this "curtain" (stupid new word, I know)... maybe it's even not about deciding to get zebra or not, it's about understanding this effect. Zebra is not the only synth that makes me think I have this perception... Maybe once it is clear what it is Urs can make the very good zebra synth the best synth ever... or at least even greater.

But you might be right, the thread is up to a point I have to deal with it a while for myself... nevertheless I found most answers valuable and discussion in general quite interesting...

[TiUser]

Hehe, I guess you misunderstood my argument. I don't think you need different engines. I just said that different synths tend to be optimized for different effects. The default settings of mine are all about "nice but clean" (some may call it "thin"). From there on you can bend them to whatever you want. Other synths are optimized to sound "strong" - such as the Virus. I find latter synths hard to fit in a mix, due to the tricks used, one of them being the built-in pseudo-compresssion.

Maybe I misunderstood the intention of your explanation but on the other hand my experience is that even with triccy programming you can't fundamentally change the character of a synth, it's like bing fenced in it's own box. That's why all synths sound more or less different despite they all make you believe they do the same, a sawtooth for instance...

The often mentioned attenuation of upper end has many positive effects IMHO, one of which is that I find the raw oscillators more pleasant sounding, on the other hand it helps keeping aliasing out, without oversampling - which in turn is good for cpu usage. Urs

Hmmm, also I have only a basic imagination of DSP or audio code programming in general, I have mixed feelings on these statements. IMHO the perfect way to avoid aliasing is not oversampling or keeping high frequencies low but to avoid producing frequencies over the nyqist barrier in the digital domain at all, in every stupid simple operation, even a multiplication, otherwise there is always anything left that can fold back, being uncontrollable emphasized anywhere and maybe can be one reason for my perception. I have no clue so how to achieve this perfectly but I think it is CPU intensive and/or more effort to program, but I don't think it's impossible either.

Ok, you are the expert, tell me the final truth now, please...

[Teksonik]

you can't fundamentally change the character of a synth, it's like bing fenced in it's own box.

Like a Virus?



----

[TiUser]

you can't fundamentally change the character of a synth, it's like bing fenced in it's own box.

Like a Virus?

Don't understand how you mean that... something like I am fenced in a Virus type of soundbox? NO.

I just tried to comment on the technical arguments.

It's quite probable that the design in the audio engine Urs describes might be the reason for my perceptions. If you damp the high end and accept a little aliasing just for the sake of processing power (less need for oversampling) it's clear which box we are talking about... or did Urs hide the doors and windows in it anywhere...?

I was told by other not so tech oriented people that many VSTis take this approach, hence the "plasic shine" tendence, the "character", the "box" the sound is fenced in, for sake of processing power and high polyphony... I was not willing to believe it, 'til now... reading it from a develloper himself.

But anyway, we are still in the tech domain, we are talking about subtle perceptions and not about "Zebra sounds bad", but I have first hints why my perception might be as it is and possibly I don't need to see a doctor soon

[djsubject]

what sample rate are you running Zebra2 at?

have you tried it at higher sample rates to see if the "curtain" is effected in any way?

Subz

[tetsuneko]

(DISCLAIMER: A BUNCH OF 'BS' FOLLOWS - SALT IT UP GENTLEMEN - THE TEXT BELOW MAKES NO SENSE FOR A 'NON-BELIEVER')

I have a friend, who is a brilliant keyboard player. He is also very much into the works and philosophies of a certain Nikola Tesla. He says he can hear miniscule differences in 'sound quality', and claims that a mac sounds better than a PC, software downloaded via wlan 'sound better' than via cable, that certain digital mixers sound better when the LCD is turned off (this he actually proved to me in A/B, but sceptics can always invalidate this of course)

For a long time I thought that he was totally lost, but through the years I am starting to believe that he actually can detect some changes in the things (ie. is not totally making it up). What's worse, I find I am too becoming like him, and this has led me to 'being lost'.. I don't claim to understand anything about audio anymore, especially about digital audio/DSP. Electricity is more than the sum of it's parts - There's just something in it that doesn't add up 'all the way'.

These 'problems' you hear might be a similar symptom. If this is the case then I am afraid you must find the answer for yourself.. Try using stuff at different locations and with different hardware and see if any anomalies 'crop up' on a certain system but not in others.. Might give clues as on what is affecting things and what is not.

My solution so far has been to ignore these things when making music - as long as the essential elements in the songs are there, and the mix is about there, I'm happy. Perhaps the reasons for my insanity will be proven by future generations, but in the meantime, I'm still gonna try coping with what I have and not spend all my time & energy in trying to fit the pieces together.

[Urs]

IMHO the perfect way to avoid aliasing is not oversampling or keeping high frequencies low but to avoid producing frequencies over the nyqist barrier in the digital domain at all, in every stupid simple operation, even a multiplication, otherwise there is always anything left that can fold back, being uncontrollable emphasized anywhere and maybe can be one reason for my perception.

Sorry, hehehe, you can't not oversample or not attenuate high frequencies without having pitch/filter/vca modulated by envelopes/lfos etc. introduce audible aliasing. Apart from oscillators that alias due to sloppy implementation all other non-linear operations (-> modulation) create harmonics that eventually will fold back.

Whatever it is that makes you feel weird about Zebra, I'm afraid it's not going away - I highly recommend looking somewhere else because I'm very happy with its sound and thus I will not change anything.

[TiUser]

...have you tried it at higher sample rates to see if the "curtain" is effected in any way?...

First to repeat, we are talking about subtle things... not about Zebra being good or bad, it is good.

Which sample rate do you suggest? 192KHz? Implies 4x CPU use as well, does it?

VSTi like NI B4 are pretty fine on 44.1KHz too, I don't think more is necessary but one can use 96KHz to be on the save side. Oversampling in the converters is another story and has not necessarily to do with the rate in the internal signal path, does it?...

Personally I rate 24bit D/C conversion (and higher internal level resolution) as more important than high sample rates because of less noise - at least theoretically S/N is quite high if one uses quality converters.

So what's the message? Workaround the intentional design of Zebra by higher sample rate? I have doubts if this is the solution...

[djsubject]

...have you tried it at higher sample rates to see if the "curtain" is effected in any way?...

First to repeat, we are talking about subtle things... not about Zebra being good or bad, it is good.

Which sample rate do you suggest? 192KHz? Implies 4x CPU use as well, does it?

VSTi like NI B4 are pretty fine on 44.1KHz too, I don't think more is necessary but one can use 96KHz to be on the save side. Oversampling in the converters is another story and has not necessarily to do with the rate in the internal signal path, does it?...

Personally I rate 24bit D/C conversion (and higher internal level resolution) as more important than high sample rates because of less noise - at least theoretically S/N is quite high if one uses quality converters.

So what's the message? Workaround the intentional design of Zebra by higher sample rate? I have doubts if this is the solution...

i was just curious,

at higher sampler rates aliasing is less noticeable,

IMO aliasing is not something that affects Z2 & yes 192 is 4x more cpu

Subz

[TiUser]

...Apart from oscillators that alias due to sloppy implementation all other non-linear operations (-> modulation) create harmonics that eventually will fold back.

I agree, but maybe there are more clever options to keep side effects much lower...? anyway.

Whatever it is that makes you feel weird about Zebra, I'm afraid it's not going away - I highly recommend looking somewhere else because I'm very happy with its sound and thus I will not change anything.

Can't remember having requested changes Just wanted to understand. So it's not only Zebra, the Virus has that a bit too and I have one...

Many things in digital domain seem to me weighting in effort and consequences against each other... maybe over time new algorithms will arise. Look at Melodyne... there are always new things hard to believe at first until some are around who make them happen.

Never mind, I still think Zebra sounds good - no basic issue with that anyway.

[zerocrossing]

Hi TiUser. I think I know your problem. The lack of tone you hear is your mind being closed. Not hard to open it, it just can often be a bit tricky.

OK, I'm flaming you, but I don't really mean to be mean. (only a little). First, let's start off with your user name. You're screaming about how much you love a product IN YOUR NAME. To me this speaks volumes. I owned a Virus C and I swore it was the best thing since sliced bread... but I did keep around a few other synths to keep things varied. Each has it's own character and I love that. Each one was easily replaced by hardware. It stung a little at first like jumping into cold water, but once used to it I realized how much fun it was to SWIM in the world of software synthesis.

A month or so ago I started a thread because I (was board) and a bit irked at the "NOTHING SOUNDS AS GOOD AS HARDWARE" posts I see so often. You remember that thread, don't you? You posted in it a lot. If memory serves you were pro hardware. Anyway, I find it interesting that the synths you are mentioning are the ones I spoke about in my post. To people who ignored the thread, it went something like this: I started playing with a guy who has a Prophet '08. Great sounding synth, but when up against my small army of software synths, Zebra 2 being one of the "big guns" it didn't have the "OMG real analog IS THE BEST!" impression that people seem to speak of. Again, it wasn't bad at all, it was just not far out in front of any of my vsts and frankly a bit behind Zebra 2 for my personal tastes.

Anyway, my guess is that somewhere deep in your head you've invested a lot of money in a piece of hardware. I great sounding one, I might add. By your nickname you're telling the world that you're emotionally invested in this piece of hardware as well. I think when you hear Zebra 2 sounding as good or better, you die a little inside because the piece of hardware that you love enough to make part of your name pales a little in comparison.

Anyway, I sold my Virus C and there's not a day that goes by that I miss it.


Mark

I am not about to replace anything but the other way round, I want something different... Usually the fancy impression with the cheap speakers always worked, if something attracted my attention when listening to them it was great on better speakers... That's the point that confuses me. Maybe it's something psychoacustical, I don't know...

[TiUser]

i was just curious, at higher sampler rates aliasing is less noticeable,...

I agree but I am not about aliasing in the usual sense... aliasing changes depending on each patch but what I meant is independent from the patch. I don't mean at all that this is a big issue, otherwise I would have sold my virus or maybe had never bought one... Let's see, I think I need some time to digest now...

[beej]

He's not saying "change all your work habits now", he's saying *try* it at 96KHz to see whether it makes a difference.

Does it make a difference for you, or not?

[djsubject]

He's not saying "change all your work habits now", he's saying *try* it at 96KHz to see whether it makes a difference.

Does it make a difference for you, or not?

thanks

i thought it was me explaining badly,

well it probably was but your just used to me as bit more

44.1, 48, 96 & 192 all (to me) sound different with all plugins,

many times i prefer 44.1 & 48 over the higher smoother sounding ones

but i like dirty sounds

Subz

[djsubject]

i was just curious, at higher sampler rates aliasing is less noticeable,...

I agree but I am not about aliasing in the usual sense

many artefacts of aliasing are subliminal (hard to describe/pin point/notice) hence why i asked you to try other sample rates

but its your choice if you try or not

Subz