A basic example of the different sound of the virus generating at 44.1khz vs 48khz. Want brighter sound with more top end clarity?

  • Here's a sample for you guys illustrating the differences between running the virus at 44 vs 48. Just a basic saw waveform but it proves an easily audible difference. If you want brighter sounds and more top end clarity and sizzle, you want the Virus generating at 48k internally! In order to make for perfect comparison on a level playing field the samples are seamlessly looped with levels matched and the 48k render is downsampled to 44k just as you would do if you are producing in 44.1khz natively.


    First half is rendered at 44.1khz, second half is 48khz downsampled to 44.1k: http://www27.zippyshare.com/v/86783886/file.html **NOTE - Do NOT listen to it off the zippy player! Download it.


    Sample Rate Conversion done by the best SRC I know of: Izotope 64bit SRC. If you do decide to run the virus at 48k make sure you use a top notch offline SRC to convert it to your project's sample rate. Letting the host do that on the fly is not the best way to go.


    The Virus normally allows you to choose between running 44 or 48k internally but once you plug in the USB it goes on autopilot. The Virus will render at 44k internally if your audio interface is set at 44k or 88k. Likewise it will generate at 48k if your interface is running at either 48k or 96k. Those rates remain the same regardless of if you're using the analog or digital outputs.


    The great thing about standalone mode is you get to choose to generate at 44 or 48k and it gets generated as such then sent out the analog outputs and recorded at whatever your project is running at, so you don't need to convert the sample rate.
    I wonder if Access will ever let us pick internal clock rates while USB is connected? Less aliasing is a known benefit to going above 44.1khz since more of the aliasing (high end non-harmonic digital dirt) is moved above the audible range.

  • Just offering another view so it doesn't look like everyone on the board is agreeing with you by not responding.


    I would prefer Access prioritize other issues than this.



    Thx

  • Just offering another view so it doesn't look like everyone on the board is agreeing with you by not responding.
    I would prefer Access prioritize other issues than this.
    Thx


    Hmmm.. You may want to re-read the original post? This is about the difference in sound you can get with the virus generating at 44 vs 48k internally! As in.. with the unit as is currently available.
    As for one day allowing us to retain control over the selection of clock rates once usb is plugged in, well that's just a slight alteration, not developing a new feature or hunting down bugs.

  • Yeah I read the original post. I maintain my stance. If the choice is between allowing users to experience the difference between 3.9 kHz of barely-audible sound fidelity or having a Virus TI plugin mode without audio (as discussed elsewhere), I choose the latter. I think a lot of other people would too.


    As well, you have no idea if it's a slight alteration or a major undertaking, because you don't work at Access. Only they know.


    I do think the level of effort you're putting into this is impressive though.

  • Yeah I read the original post. I maintain my stance. If the choice is between allowing users to experience the difference between 3.9 kHz of barely-audible sound fidelity or having a Virus TI plugin mode without audio (as discussed elsewhere), I choose the latter. I think a lot of other people would too.


    As well, you have no idea if it's a slight alteration or a major undertaking, because you don't work at Access. Only they know.


    I do think the level of effort you're putting into this is impressive though.


    If the difference is barely audible for you and appears only at 3.9khz it's probably time for a new pair of monitors my friend! Haha.. what did you do, listen to it off the zippyshare player through laptop speakers?
    The difference is clear as day. In that example the difference is 4db at 15khz and only gets larger. By the time it reaches 19khz it's an 8 db difference. Ever hear people gripe about the dark sound of the virus? That would be a potential cure for that. You can always filter stuff down, but you can't open a filter more than full. The Analog 4 pole would benefit here for sure as it's a lot darker itself. I like the sound of it but it's too dark for me to use a lot of times.
    Also, where I suggested you went terribly wrong was that my post is primarily to illustrate the difference so that those who want more top end clarity and sizzle can make use of the existing ways you can operate the Virus at 48k. It's not a new development whatsoever. It's simply making use of the different clock rates in the existing functionality regardless of if usb is connected. The footnote to all that is that yes, it would be nice to be able to also have that with usb running. How do I know it wouldn't be a big undertaking? Well, because it's already implemented in the virus and happens currently in VC. The only prob is you don't have the option to choose it since it runs on auotopilot. So yes, one day a little box where you can choose between generating at 44 & 48k would be GREAT but until then you know how to get that sound if you want it.

  • Admittedly I did play it off zippy share off some laptop speakers! Just seems to fly in the face of years of industry studies showing no human being on earth ever passes the blind listening test between 44.1 kHz and 48 kHz...but I see you're talking about something else: filters behaving differently due to sampling rate.


    Personally I love the dark sound, but I can see how you'd want to have the option.


    Let's hope you're right about it being a simple change!

  • Admittedly I did play it off zippy share off some laptop speakers! Just seems to fly in the face of years of industry studies showing no human being on earth ever passes the blind listening test between 44.1 kHz and 48 kHz...but I see you're talking about something else: filters behaving differently due to sampling rate.


    LOL! classic. Man you aren't supposed to do critical listening in such a manner!
    It sounds like you are confused about studies and people not being able to tell 44 vs 48k though. I don't know that I've heard of these studies but if they are legit they are only true for particular circumstances.. like the recording of an analog source, not digitally generated sound via digital instruments & fx.


    The reason for the audible difference is that with the 44khz audio format the filter that bookends the available frequency limit stretches down within audible range whereas with 48k and higher that bookend reaches well above audible range which is also where a lot of the digital aliasing gets pushed.. so you end up with a clearer signal and more open top end.


    Anyway.. I absolutely love my Virus but one weakness for me has been the raw oscillators in some instances. Generating at 48k gives me the top end sizzle and strength I was missing on a lot of occasions. It's better to have the option to make a sound darker rather than to have only the choice of using the darker sound. Check the sample on your studio monitors.. it may catch your interest.

  • I do loads of reading, the downside of that is that I can't always remember where I read something.


    Just in the last week or so I read an article that covers this very subject, from memory it was saying that the benefit of upsampling to 48k took the aliasing up to levels beyond the capabilities of the human ear - which would explain why you are getting more clarity to your examples. The aliasing is still there, just you can't hear it anymore.


    I'll try & find this article & post it here, I nearly did post it in another thread. I wish I had done now.

  • There are most definitely benefits of reduced aliasing by going much further up in the frequency range than just 48k. In the jump from 44k to 48k you'll hear a difference from the filter at the top of the available frequency range raised higher to a level where it's no longer audible as well as reduced aliasing from some, but not all, aliasing being moved out of the audible range. From that point the filter is inaudible but aliasing can be reduced much further still. Try taking a softsynth that hasn't been rendered yet and using bounce a part or just a saw tone to audio at 44k, 48k, 88k, and 192k using your DAW's offline bounce/export feature and you will hear the results of that. Just get your hands on a top notch SRC like the Izotope 64bit SRC and downsample all the results to 44k after they're rendered (simulating the final output of the public release format). Plugins work better at high sample rates. This is why I work at 88.2khz normally.

  • How frustrating, I've spent the last hour or so trying to find this article for you. My main computer has just had OSX re-installed so there's no history to look through & I mustn't have read this on my Macbook as it's not in any history there. I've found a few things on google but not the article I'm thinking about.


    It was saying something along the lines of some synths work behind the scenes at much higher sample rates pushing the nyquist frequency higher to avoid audible aliasing, then convert the signal back to the required sample rate at the outputs. Say for example you set your sample rate to 48khz, internally the processing would be happening at 192khz. I was going to post it in the forums here in response to a post Marc had made about sample rates, I'm not sure but I got the impression that the Virus was maybe a synth that did this.


    I'm sorry that I can't give you anything more solid to go on than a half recalled article I read, the paragraph I have written above may be complete rubbish - but I think I've got the main idea behind the article. If I do come across it again I will definitely post it here.

  • How frustrating, I've spent the last hour or so trying to find this article for you. My main computer has just had OSX re-installed so there's no history to look through & I mustn't have read this on my Macbook as it's not in any history there. I've found a few things on google but not the article I'm thinking about.


    It was saying something along the lines of some synths work behind the scenes at much higher sample rates pushing the nyquist frequency higher to avoid audible aliasing, then convert the signal back to the required sample rate at the outputs. Say for example you set your sample rate to 48khz, internally the processing would be happening at 192khz. I was going to post it in the forums here in response to a post Marc had made about sample rates, I'm not sure but I got the impression that the Virus was maybe a synth that did this.


    I'm sorry that I can't give you anything more solid to go on than a half recalled article I read, the paragraph I have written above may be complete rubbish - but I think I've got the main idea behind the article. If I do come across it again I will definitely post it here.


    No need.. I've actually written pages worth on the topic of sample rates and tested and discussed it to death so it wouldn't be anything new ;)
    What you're talking about is oversampling.. where a digital signal is processed at a rate several multiples of itself above to avoid digital aliasing and then downsampled to working sample rate.



    it does, at least where it really counts. please understand that i will not get into details.
    marc

    Why aren't you able to get into details on this Marc to clarify what it does? This is definitely relevant information! I had assumed it didn't since even jumping up to 48khz rendering gives an instantly noticable difference to the top end as I've shown here.


  • No need.. I've actually written pages worth on the topic of sample rates and tested and discussed it to death so it wouldn't be anything new ;)

    Fair enough


    Zitat

    What you're talking about is oversampling.. where a digital signal is
    processed at a rate several multiples of itself above to avoid digital
    aliasing and then downsampled to working sample rate.

    Yeah, I said upsampling earlier didn't I. Oversampling is what I meant ^^

  • once you plug in the USB it goes on autopilot. The Virus will render at 44k internally if your audio interface is set at 44k or 88k. Likewise it will generate at 48k if your interface is running at either 48k or 96k. Those rates remain the same regardless of if you're using the analog or digital outputs.

  • Yes. Your virus can only be sat to one clock rate at a time so if your project is at 48 or 96 then the Virus will generate at 48k internally whether usb or analog outs are used.
    Personally I avoid working at 48k and my preference is 88.2k as there are occasional benefits to working at exactly double the rate of your final product. The cleanest way for me to work it seems is switch my project to 48khz when I want to bounce audio and do it via usb, then use Izotope's 64bit SRC to convert it to my project's 88.2k sample rate.