Asio Latency - Does the Virus Output Latency Matter ?

    Hi guys,


    I noticed that the Virus Asio configuration window shows a much higher output latency than input latency. If I set the Virus to 256 samples latency it shows roughly 8ms input latency and 28 ms output latency. My main soundcard is set to 512 samples latency and has 11ms input and 11ms output latency.


    My question: Does the virus add 36 ms latency ? Does the output latency actually matter ?


    Background of the Question:


    I assume:


    If I press a note on my keyboard then


    1. it gets into the daw (no asio latency added I guess)
    2. through the DAW out to the Virus (no asio latency added I guess)
    3.The Virus plays the sound and sends it into the DAW via its Driver. ( 8 ms Input Latency in my case added I guess)
    4. The sound then "flows" into the main audio card audio stream. (11 ms Audio Card Input Latency added I guess)
    5. Without any plugins the sound get send out of my soundcard to my speakers (11 ms Audio Card Output Latency added I guess)


    But where is the Virus Output Latency added ?


    p.s. I found this thread but it's quite vague:
    http://virus.info/forum/index.…ighlight=Latency#post8907

    It depends on what you mean by "Virus plays the sound and sends it into the DAW via its Driver". There are 2 possible ways:
    Not using VC:
    The Virus driver just acts as a MIDI device and the sound comes back through the analog and/or SPDIF outs. The Virus' response to MIDI is pretty snappy and is lower than the soundcard's input latency.
    Using VC:
    The sound data is sent back using the USB connection and is gathered in a buffer inside VC, to avoid jitter and other problems caused by USB not being a realtime oriented interface. This buffer has 2 modes from which you can select: in live mode it is 512 samples long and in non-live mode it is 4096 samples long. When playing back your project from your DAW, the DAW is expected to compensate for this latency, since VC reports it as standard plugin latency. But when playing live into VC (e.g. tracking MIDI), the DAW cannot foresee your playing and therefore no compensation is taking place and you experience the latency very clearly. In this case, live mode can give you latency that's low enough, while you may hear some audio artifacts that might not matter if you are just tracking MIDI.
    Hope this makes thing clearer.

    Thanks for your answer.


    Yes, I was speaking of using VC. The Virus ASIO Latency has no effect if you play the Virus like a "normal" hardware synth.


    If played like a normal hardware Synth, the only ASIO Latency that matters is the ASIO latency of the soundcard. In my case, with the setup described in post 1, it takes 11 ms until the sound arrives inside my daw and an additional 11 ms to get out and send out to my speakers. So it would take roughly 22 ms to hear the sound (plus maybe some mini latency by the midi device and driver),


    However, we are talking about using VC:


    And here is a quote from Mark: "the latency you see in the inspector is not what you will hear. [...] if your system is configured right, you can roughly expect the virus to have a latency which equals the buffer size of your soundcard."


    If you open up the Virus ASIO Window you can see the latencies. The input latency seems to be somewhat in line with the common soundcards but the outpust latency is totally off.


    Maybe both latencies only play a role if you send audio from the DAW into the Virus (Atomiser or so) ? I don't know, because my DAW, Cubase, does not allow sending audio to the Virus.


    The more I think about it the more consused I am lol.

    Zitat von flabberbob

    As I said, VC has its own plugin latency which is separate and is not affected by the ASIO latency, whether the Virus itself or a separate soundcard is the DAW's main soundcard.


    Sadly this answer does not solve my question :) Plugin latencies usually get added to the soundcard asio latency, yep. But as if things were not confusing enough, the plugin latency is apparantly wrong, as even Mark stated.

    if you are referring to the following statement:

    Zitat von Superscale

    And here is a quote from Mark: "the latency you see in the inspector is not what you will hear. [...] if your system is configured right, you can roughly expect the virus to have a latency which equals the buffer size of your soundcard."

    I think what it means that when you play your project in the DAW, the DAW will compensate for the plugin, and you will hear everything on time. But, when you play MIDI live from your controller through VC there will be an added latency, with live mode off, 4096/44.1kHz=92.9ms or 4096/48kHz=85.3ms, and with live mode on: 512/44.1kHz=11.6ms or 512/48kHz=10.7ms. Anyway, we will have to ask Marc if we really want to know what he meant.

    Zitat von flabberbob

    I think what it means that when you play your project in the DAW, the DAW will compensate for the plugin, and you will hear everything on time. But, when you play MIDI live from your controller through VC there will be an added latency, with live mode off, 4096/44.1kHz=92.9ms or 4096/48kHz=85.3ms, and with live mode on: 512/44.1kHz=11.6ms or 512/48kHz=10.7ms. Anyway, we will have to ask Marc if we really want to know what he meant.


    most sequencer have a feature which lowers the buffer size of the active track. that's why the we feel less latency that what you mention above. "live mode" is something similar and designed for sequencers which don't have a "fast lane". the latency reported by sound cards to the OS is the supposed throughput latency of the entire system. the grand total, at least in VST land, can be set to anything including values which are to small to make the product appear in a different light.
    with most sequencers, plug-ins can add latency to a channel which is compensated in playback. this latency is independent from above stated latencies. as a result, in a realworld situation with most sequencers, you will never face the full latency reported by a plug-in but a fraction of it.
    hth, marc
    ps. here's the dirty laundry: live users (yes, the sequencer it is called LIVE and not ABLETON) need to use the "live mode" because this sequencer shifts midi events by the reported latency. their idea is that a skilled keyboarder predelays his play to manually compensate for the latency he hears, similar to the guy playing the organ in the church. unfortunately, with a massive church like mentioned above, it becomes counterproductive. Ableton has promised to improve the situation since years.
    one other company worth mentioning is Image Line: they managed to rewrite the latency compensation system in Fruity Loops in a way that is counterproductive for Virus TI users. their explanation on why things stopped working properly which worked before i don't want to share here. as a result, this sequencer might be listed incompatible in the future - so fingers crossed that they find a way to make things work again.