Virus TI running at 96khz

hey... no I haven't gone any further in depth with it really as I've had my hands full with other stuff. If setting up listening tests though I suppose a thing to keep in mind is trying to do it in ways where it's easy for you to focus.. ie, perhaps the same chord or note in repetition playing from a midi file. When I did the test above I recorded many hits using a spectral analyzer as reference to select the brightest ones to compare so I wasn't foolishly just comparing a bright take with a darker one. How much tonal variation there can be between hits depends on the patch though. Maybe I didn't have retrigger on. Anyhow.. dive in there and see if you can expand on that area!

Incidentally - there is a discrepancy in the behaviour in some circumstances between 44.1 kHz and 48 kHz - which I would put down to being a feature of the Virus... It's by no means a physical limitation. Also, the difficulty in reproducing it also demonstrates what a fringe issue it is.

wtf? difficulty in reproducing it? You really don't seem very good at understanding things man and I have no idea what on earth you are doing repeatedly involving yourself in the topic in this and other threads. Pick a topic you actually know about if you feel the need to comment. It keeps things focused and saves the time of people having to correct you.

Recording a bunch of notes and then taking care to select the brightest one recorded at each sample rate just allowed for accurate comparison. It might have just needed retrigger to be turned on to make that part easier but I can't remember as I recorded that ages ago. It would be better to do with retrigger on. .

I know you're slow and seem to like being corrected, but being limited by the filters at 44khz vs 48khz is a PHYSICAL LIMITATION. It's the primary limitation you are hearing on that sound which isn't even a high sound. It's a 147hz saw. I often end up having to use filters with less poles just to keep the blanket off the top end. The 4 pole sounds great but is often just too dark for me to be able to use.

I think it would be useful to check the the classic oscillators as well as the wavetable oscillators in classic wave shapes, since the classic oscillators algorithm generates some aliasing, an it could be that the brightness you experience in 48kHz is just added aliasing. Also, the method of "cherry picking" the sample is not very scientific. It would be much more reliable to average the samples, for example. I am not familiar with the term "retrigger" in the context of the Virus, could it be something that I overlooked?

No, I know how to pick 2 proper samples for comparing, which is exactly why I did it out of 15 or so hits and had a spectrum analyzer, oscilloscope, and tuner on the signal to confirm. Handling that kind of thing is a walk in the park for me. What I should have done was made it easier by turning on the OSC retrigger (osc phase init) but like I said this was a while back and I gapped on that. The brightness at 48khz is absolutely not added aliasing. Aliasing works the other way around, lower sample rates = more aliasing in the audible band. High and clarity/openness/air are the most commonly observed attributes when people compare 44khz vs higher SR. On 44khz the antialias filters are affecting the audible band. At 48khz the filters are raised enough that it's a lot less audible. At 88khz the filters & aliasing are raised so high out of hearing rangte that they disappear.

While the wavetables do alias less, the filter on the Virus might play the biggest part here since you can't open the filter as much at 44khz as you can at 48khz. It has to work within its limtations.

I'm sorry NMS, I didn't realise you were an expert on ADCs and DACs - here was me thinking that with a basic interpolation filter on a 44.1kHz DAC you could comfortably represent all audible frequencies with near-zero attenuation and phase distortion - but obviously you know better despite most likely not having any experience in DSP system design. The example you produced - the prevalence of which we don't yet know - cannot be explained simply by looking at the sample rate and MUST be due to a feature of the software on the Virus - be it the resampling algorithm to generate the waveforms at different frequencies, or downsampling after oversampling stages in the signal chain (ie an intermediate AA filter). It could also be differing rounding error in the coefficients used for any of the signal processing going on in there.

If you have audible aliasing at 44.1 kHz you are doing something wrong (which I would doubt) or you've had to make a compromise to squeeze more performance out of the system (probably) - at 48 kHz you have no excuse as the guard band is massive. Beyond that you are full of shit. But like I said - believe what you want. The more people believe this nonsense the more room there will be for competitiveness. The memory and processing power requirements of increasing the sample rate are devastating which is why you would only upsample for physical modelling if you have some awkward difference equations to implement (like the virus does, of course), but an interpolation algorithm, rather than upsampling the signal wholesale, can (and has been, for me) very effective.

Also, I never said the 'added brightness at 48 kHz' is due to added aliasing - but I might have said the difference was down to an anti-aliasing or interpolation filter, depending on whether its a downsampling stage or the DAC stage, the latter being irrelevant for USB or S/PDIF audio.

At any rate I don't appreciate the offensive language and it's not in the spirit of this place - I'd spend more time demonstrating why you are wrong and how ignorant and foolish you look by being so rude, but I have a signal processing platform to finish developing before I finish my PhD in DSP so I'll leave it where it is. You can troll more if you want, and I'm sure you will, but no amount of boasing about how plugging something into a spectral analyser is easy for you (as if it would be difficult for anyone ) will change the fact that a lot of very qualified people (not slow people who like being corrected as you so eloquently put it) chose 44.1 kHz as a sample rate for CD audio with very good reason and that 48 kHz is enough of a concession to relax the requirements of SRC by having a huge guard band (4 kHz!) whilst keeping the bandwidth low.

All the while forgetting that almost nobody can hear 18 kHz, let alone 20kHz anyway!

Zitat von AtonyB

If you have audible aliasing at 44.1 kHz you are doing something wrong

So now aliasing doesn't exist? Access nailed alias free digital synthesis (that needs no help from running the synth at a higher rate) 15 yrs ago? Are you serious? Are you aware of the statement that's been made that the wavetables are significantly more alias free than the 15 yr old classic oscillators? Do you live under a rock oblivious to how much progress has been made in the last 10-15 yrs on digital synths and that aliasing is one of the biggest areas of improvement?

Zitat

The memory and processing power requirements of increasing the sample rate are devastating

Dcam synth squad and The Glue (2 of the highest regarded plugins ever made) both offer realtime oversampling up to 32x. Welcome to 2012. And you're acting like 88khz is overkill and "devastating" to implement? Seriously, no offense intended I just genuinely want to know.. do you actually live under a rock?

Zitat

At any rate I don't appreciate the offensive language and it's not in the spirit of this place - I'd spend more time demonstrating why you are wrong and how ignorant and foolish you look by being so rude

Zitat

Beyond that you are full of shit.

Classic.

Zitat

All the while forgetting that almost nobody can hear 18 kHz, let alone 20kHz anyway!

Cool! And how about the blanketed top end of those comparison files? I can hear the difference playing off my iphone speaker.

It's obvious to anyone reading that you passed the point of annoying long ago and have genuinely no business in this thread or any thread where people are expressing their desire for 88 or 96k operation of the Virus. None at all. You want to tell people that they shouldn't want what they want, but you are absolutely no one of any status or experience level to be telling them they're wrong for wanting it. If you find me rude it's because you've been an anoyying fly buzzing around this issue for ages spouting crap and backing up your words with nothing when you have no interest in having this functionality for your own use. I like my Virus too, but I don't feel the need to go "THAT'S A FEATURE NOT A LIMITATION!" when it comes to its shortcomings as if it's the most perfect synth that will ever grace the earth. Do you think the next Virus won't be able to operate at 88 or 96k internally? Guess again! Not in today's market.

Please guys, calm down!
What a pity seeing you fighting on this forum when you obviously both have the knowledge to understand each other ;
I'm following this thread from the beginning, and I'm disappointed to have to read through your mutual insults to grab the precious pieces of knowledge you're both bringing here.

Here's my technical understanding, please correct me if I'm wrong :
1- the output sampling rate and the internal processing sampling rate are two very different things.
2- an output sampling rate high enough to allow non-audible frequencies is a waste of space
3- a wav file at 96kHz can be downsampled to 48kHz without any audible difference (for a human).
4- depending on how the internal audio processing is done, a processing sampling rate higher than the output sampling rate can be beneficial : this is what we call oversampling.

Here are my questions then :
5- Are there good reasons for wanting an output sampling rate higher than 48kHz?
6- Wouldn't running the virus at 96kHz (internally) enhance synthesis/processing quality?
7- What about processing latency? A lot of processing can't be done in parallel, so halving the polyphony may not suffice. I mean, if one of your processing units is sequential and can only process up to 60k samples every second, no matter what you do to polyphony, you can't achieve real-time performance in 96kHz.

On the non-technical side, market adapts to what people will buy, not to what people will benefit from ; it's a fact that (for the same price) a 96kHz soundcard will sell better than a 48kHz one. Wether it makes an audible difference or not. Because, in everyone minds, bigger numbers are better. So I think arguments based on what recent gear do are irrelevant here.

And for chilly7: a feature you're not using can still hurt you. I'm a software developer, and I know that a lot of bugs appear when features interact, even with disabled ones. For example, a lot of Virus users face the "detuned by one semitone" problem. It's due to an output sampling rate selection bug (44.1 vs 48), although a lot of us only use one sampling rate. So now, what would it be with 3 sampling rates? AtonyB is perfectly legitimate not to want this option on his synthesizer. BTW, I think AtonyB's goal is a lot more noble : he believes you're making a mistake and tries to explain his understanding to you (after all, he may teach DSP very soon!). So why so agressive?

Cheers

Zitat von Ace17

Please guys, calm down!
What a pity seeing you fighting on this forum when you obviously both have the knowledge to understand each other ;
I'm following this thread from the beginning, and I'm disappointed to have to read through your mutual insults to grab the precious pieces of knowledge you're both bringing here.

…

I think AtonyB's goal is a lot more noble : he believes you're making a mistake and tries to explain his understanding to you (after all, he may teach DSP very soon!). So why so agressive?

Alles anzeigen

Its about the sound that comes out of the Virus at the end of the day! NOT numbers!!! Im a Tech too, but this is becoming petty & one upmanship snobbery!... jesus!...

No one notices what bleedin sample rate it comes out at or whatever! as long as IT SOUNDS GOOD!!! music remember? NOT physics or Algorithms....

Do you know, i had a great recording session with the Virus tonight, and it cut through the mix like a hot knife through butter, and i was playing around with 5 synths, a Korg M50, Roland Juno Stage, Roland Juno106, Korg X3, Korg Triton, so dont get so uptight... CCHHIIILLLL!!!!

Spend more time playing than criticising... happy days!

You know Hans Zimmer will spend days working on ONE sound sometimes? Do you know some of the lengths people went to in recording some of the most sonically impressive sounding results over history? Think Thriller was just tossed together by some guys going "bah.. forget all that technical stuff it's the music that matters!" Personally, if I can hear a difference on my iphone playing over speakerphone, I think that's probably something worth concerning myself with. Let's be realistic here.. these days people don't rave about how well the raw Virus oscillators are coded and they don't stand up to current offerings like Diva, Strobe, Massive, or Sylenth. They just don't. The Virus wins by its other merits and deep programmability combined with the user interface. I just wish I had the interface & programmability of my Virus with the raw oscillator quality of today's modern coded offerings. The #1 way a user can improve the sound of older software synths is running them at higher sample rates.

Zitat von Ace17

Here are my questions then :
5- Are there good reasons for wanting an output sampling rate higher than 48kHz?
6- Wouldn't running the virus at 96kHz (internally) enhance synthesis/processing quality?
7- What about processing latency? A lot of processing can't be done in parallel, so halving the polyphony may not suffice. I mean, if one of your processing units is sequential and can only process up to 60k samples every second, no matter what you do to polyphony, you can't achieve real-time performance in 96kHz.

5. Yes, obviously for compatibility. If you work at 88khz which is in fact scientifically proven to be the best SR to work at for digital music production for several reasons then you can't get the Virus operating at higher than 44khz internally if you have a USB plugged into it.
6. Yes
7. The 88/96k environment reduces latency, it doesn't add to it. A 44k synth oversampling to 88k then back down to 44k will have more latency than an 88khz synth going about its business at 88khz.. The 2 issues are the digital audio bandwidth through the virus digital outputs due to usb 1.1 restrictions and the processing power needed. You would be unable to host 3 stereo outs at 88/96k, but you could easily host one over USB, one over S/PDIF, plus your analog outs. As far as internal processing goes, the Ti2 is MORE than sufficiently capable of running at 88/96. People wanting a higher sample rate option for studio recording would trade half their polyphony in the blink of an eye to do it if such a mode were available.

Anyone who actually knows what they're talking about here (ie, absolutely not AntonyB) and the majority of developers who make all our best vst instruments and processing will recommend working at higher sample rates. We're talking about the people who as a matter of fact know better. Even where recording audio is concerned, take Dan Lavry for instance (one of the industry leaders in high end AD/DA conversion) who has openly stated that the ideal sample rate to use in recording is 60khz - and the ideal rate of any that are in current use would be the closest thing up from that which is 88khz.

All of this info is freely available to any of you guys on the internet. There's no excuse for living in the dark about it if you're actually interested in knowing. Go read what industry leaders like Andy from Cytomic (developer of D-CAM Synth Squad, The Glue, etc) or Dan Lavry say about it.

Zitat

I think arguments based on what recent gear do are irrelevant here.

What the current market offers is far from irrelevant. If you don't understand gear at all I can see how manufacturer offerings would be confusing or misleading but anyone who knows what they're doing knows that more isn't always better. I don't need 16 channels of conversion to run 3 synths and I don't need converters that work at 192 or 384khz in order to do my work at 88khz. Lavry for example refuses to offer AD/DA converters that work higher than 96khz. If you spent $8000 on their top of the line Lavry Gold DAC you're still not going to be able to record higher than 96k. Why? Because they know what they're doing and don't encourage recording at higher rates based on evidence showing that going beyond 96k causes converters to be less accurate and not more..

Zitat

And for chilly7; a feature you're not using can still hurt you. I'm a software developer, and I know that a lot of bugs appear when features interact, even with disabled ones. AtonyB is perfectly legitimate not to want this option on his synthesizer. BTW, I think AtonyB's goal is a lot more noble : he believes you're making a mistake and tries to explain his understanding to you (after all, he may teach DSP very soon!).

Yes and I might be teaching nuclear physics at Harvard. There is absolutely nothing noble about his involvement here whatsoever. He's trolling a thread arguing about a subject he does not understand and hasn't tested. It doesn't matter how many big words or terms he throws around or how suddenly for internet argument's sake he's finishing a PhD in dsp. It makes zero difference at all when he tries to keep up in a conversation about things he's not qualified to dispute due to not having taken the time to test it on the relevant gear in question. This is why he doesn't know any better. Some people would rather come to a debate with a lot of hot air though rather than take the time to back it up. Sorry, but that just doesn't cut it. If you're going to waste people's time you'll wear out your welcome.

A lot of people can sound like they know what they're talking about, but ask them to prove their words or back it up and it's a whole other story.

Zitat

Let's be realistic here.. these days people don't rave about how well
the raw Virus oscillators are coded and they don't stand up to current
offerings like Diva, Strobe, Massive, or Sylenth. They just don't.

I'm no scientist, but when people say things like this I'm always a bit intrigued by what they mean & am interested to know why the Virus oscillators get such a hard time.

I've done this test once before, but seeing as this has come up again in this thread I've ran it again - if anyone remembers the last test I did there's no point in searching for it as I've used different synths this time - As I say I'm no scientist so this test is by no means conclusive but I think it does show some interesting results, all 4 synths were sounding a single sawtooth oscillator, all filters were fully open & there was no further processing applied. The Virus audio is coming in via USB, the other 3 synths are audio units so there should be no discrepancy due to AD converters, all synths are from different manufacturers.

[Blockierte Grafik: http://img19.imageshack.us/img19/9692/screenshot20120515at134.png] 
[Blockierte Grafik: http://img403.imageshack.us/img403/9692/screenshot20120515at134.png] 
[Blockierte Grafik: http://img62.imageshack.us/img62/9692/screenshot20120515at134.png] 
[Blockierte Grafik: http://img268.imageshack.us/img268/7855/screenshot20120515at135.png]

So, from these pics, which oscillator is best? Why is it the best one? Which is the worst one? Why is it the worst one?

Ruari: considering that s(m)exoscope's resolution is very coarse, from all the screenshots you uploaded, the closest to the pure sawtooth y=2*(x-|_x_|)-1 is the last one. Little ripples around "jumping points" occur when lowpass filtering is applied (because the Furier series of harmonics is "incomplete"), so that's #2. I don't know what is that hybrid of filtered/unfiltered in #3, but #1 looks like a sawtooth with its bottom end going through some distortion or fold of some kind, probably in order to get more high end. My point of view is that it's easy to turn a pure sawtooth wave into something else, but it's hard to get a pure sawtooth from something that's already distorted/filtered.

That top one (the Virus) isn't actually it's raw saw tooth. You've got the analog boost on there which causes that lower dip. For aliasing you'd want to use a spectral analyzer and your ears though not s(m)exscope..

Here's one example of aliasing done with the Virus where you can hear it breaking up the higher it goes. The second half was done with Sylenth.
To recreate: init patch, Lfo section: Lfo1 rate 13, osc modulation 100%, hold down C5 and listen to it sweep through the top range.

Unfortunately this didn't improve much when tested at 48khz.

I really just wish the Virus was coded as well as well as the vsts I mentioned. I find the hardware, plugin interface, & programmability a hundred times better to use than any softsynths. This is where the Virus shines, as well as the wavetables.
I get that the average noob working away on a pair of Rokkits may not know the difference, and that the Virus internally is getting up there in the yrs now where dsp is concerned, but any upgrade to its core sound would be nice because I'd rather not have to use those other soft synths. Aside from their sound quality I find them irritating to use compared to the user interface I get with my Virus.

Zitat von nms

That top one (the Virus) isn't actually it's raw saw tooth. You've got the analog boost on there which causes that lower dip.

Good shout, my mistake - here's the Virus with Analog boost turned off

[Blockierte Grafik: http://img688.imageshack.us/im…reenshot20120516at130.png]

Zitat von nms

For aliasing you'd want to use a spectral analyzer and your ears though not s(m)exscope..

Yes, but I was responding to you saying that the Virus oscillators weren't as good as the other synths you listed & I was (perhaps wrongly) thinking that if it can make something that looks like a sawtooth in s(m)exoscope then that would be job done.

Zitat

Here's one example of aliasing done with the Virus where you can hear it breaking up the higher it goes.

It's clearly audible in your example, but is that a problem with the oscillator or is it possibly something else?

I don't know the answer & I'm genuinely interested.

Ruari: it's very hard to estimate the quality of any oscillator only by looking at the waveform. At least, look at the spectrum, which is more important here.

You do not want to generate a signal whose waveform looks close to the reference signal : you want to generate a signal whose frequency content is close to the reference signal frequency content. Because this is the only thing your ears will pay attention to.

And guess what : in a digital world, the closest waveform generally sounds like shit.

Waveforms which contain discontinuities (square wave) or sharp edges (triangle wave) have a spectrum which extends to infinity. But frequencies higher than half the sampling rate can't be represented in a digital signal.
Unfortunately, those frequencies don't magically disappear, instead, they get reflected into the range of frequencies your sampling rate can represent.

Which means you now have a signal full of alien frequencies, some of them going down as your pitch increases ; in one word, your oscillator aliases.

You have to filter all of these non-audible high frequencies, which means smoothing the sharp edges of your signal.

Take a look at the picture below.
In red, the reference square wave. In an analog world, this would probably be what you would aim for.
In blue, the filtered square wave. See what happens on the edges? This is not a bug in the synthesis algorithm, this is something you actually want! Because this is what allows you to stay close to a square wave, but without having edges in your signal.
[Blockierte Grafik: https://farm4.staticflickr.com/3389/4625477547_d974f432a2.jpg]

Zitat von nms

For aliasing you'd want to use a spectral analyzer

Zitat von Ace17

Ruari: it's very hard to estimate the quality of any oscillator only by looking at the waveform. At least, look at the spectrum, which is more important here.

Holy crap, I've ran these same synths through a spectrum analyser & I can see quite a remarkable difference between them. Here's my screenshots in the same order as before, see all the little frequencies in the top one (which is the Virus by the way) they slowly glide back & forth if I keep the key pressed.

[Blockierte Grafik: http://img515.imageshack.us/img515/411/screenshot20120516at161.png] 
[Blockierte Grafik: http://img341.imageshack.us/img341/411/screenshot20120516at161.png] 
[Blockierte Grafik: http://img163.imageshack.us/img163/411/screenshot20120516at161.png] 
[Blockierte Grafik: http://img546.imageshack.us/img546/411/screenshot20120516at161.png]

Zitat von Ace17

You have to filter all of these non-audible high frequencies, which means smoothing the sharp edges of your signal.

If a higher sample rate can't happen would better filtering be a potential solution?