• Register
  • Help
Page 1 of 2 12 LastLast
Results 1 to 10 of 16

Topic: Should I increase the volume of sample-based VIs before bouncing to audio?

Share/Bookmark
  1. #1

    Question Should I increase the volume of sample-based VIs before bouncing to audio?

    Ok, here's the deal: I have an orchestral project in Cubase SX using several instances of EWQLSO. Now I'm about to bounce all MIDI/VST tracks to audio tracks to do the mix. One audio track for each instrument. The kompakt players are all set to 0.0 dB and I haven't touched the Cubase mixer. The volume of the different instruments is fairly balanced by means of orchestration.
    Now, when bouncing my MIDI/VST tracks to audio tracks, some audio tracks will be loud (4 trumpets playing f) while some will be very soft (violins playing pp). Of course the softer bounced audio tracks will use up less bitdepth.
    Should I increase (normalise) the volume of the softer instruments in the kompakt player (e.g. setting it to 12.0 dB) before mixing down that track, using more bit depth, and then attenuate when mixing the audio tracks?
    Or would I just be amplifying the original EWQLSO samples playing back at their original bitrate, in which case I could just as well keep the kompakt players at 0.0 dB, bounce at low volume and turn up the faders during mixing?

    I hope this question makes sense to anyone .

    Roy

  2. #2

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    In an analog recording one would always try to hit the tape as hot as possible for every take.
    Cant see why it should be any different for digital recordings.

    Hans

  3. #3

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    Hi,

    I am no expert but I would bounce the tracks as close to their later level as possible. First it saves you time, second it saves the data recalculation steps.

    Actually the data could loose quality if it is too often amplified and then decreased again. It is more than a digital addition and subtraction - the waveform is recalculated two times. Yes it is digital but there are rounding errors.

    I think of pictures, they tend to blur if they are resized (the analogy is not 100 % correct but may also not be entirely wrong).

    Just my 2c.

    Hannes

  4. #4

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    Thanks for the input, guys. Hans: of course when recording digitally myself I will always record as hot as possible. But the samples played back by a soft sampler or ROMpler are already recorded at a certain volume/bitdepth..

    The key question here is: when my sampler plays back soft samples (e.g. pp violins) and I increase the output volume of the sampler, will I get more bitdepth in my bounced audio? Or will this just amplify the samples and add no real bitdepth, in which case it's easier to bounce to a soft audio track and add gain later via the mixer when needed..

    Anyone for a definitive answer on this?

  5. #5

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    I am no expert. But, my impression is you want to bounce to max down between -6 and -3 db (max peak volume) for each track (but some people say you need for headroom to give you room to procees each track further).

    Then, once you have the audio of each track close to the ceiling, you can then reduce the volume of tracks selectively to get the right mix.

    Did that make any sense at all?

    [Oh, nevermind, I think I did not respond to your question because I do not have a definite answer]

    jeffn1
    For original progressive electronic rock influenced by J.S. Bach and (old) Rush, check out: www.soundclick.com/jeffreynaness.

  6. #6

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    I think it depends on the bit depth of the project. If you're using 32Bit float, you don't need to normalize at all, at least from a technical point of view.

    But I have to admit that some of the discussions about bit depth sound like voodoo to me. "More is better" doesn't help everywhere

  7. #7

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    I bounce to 24-bit wav files and then normalize. If there's any degradation, I can't hear it.

    So, why do I normalize? Simply so I can see the waveforms on the tracks in Vegas/ACID. It also forces me to re-mix the levels. It takes more time, but I like having a fresh perspective on the way the instruments blend. It's too easy to leave things the way they were, simply because they become comfortable during the composing process.

    I might not take this approach if I were doing pure orchestral mock ups. I'd probably find the basic mix for the orchestra and leave it be. However, I've been scoring to picture lately, and have allowed myself to be creative with orchestrations. I mix for what sounds good, rather than what it would sound like live on stage.

  8. #8
    Senior Member Richard Berg's Avatar
    Join Date
    Jul 2002
    Location
    Chapel Hill, NC
    Posts
    232

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    In an analog recording one would always try to hit the tape as hot as possible for every take.
    Cant see why it should be any different for digital recordings.
    Because you're not recording anything. The data is already digitized.

    Actually the data could loose quality if it is too often amplified and then decreased again. It is more than a digital addition and subtraction - the waveform is recalculated two times.
    No it isn't.

    I think of pictures, they tend to blur if they are resized (the analogy is not 100 % correct but may also not be entirely wrong).
    Not comparable. Changing the gain of an audio stream does not involve resampling.

    Besides, most of the time when people see an image degrade after resampling, it's due to JPEG artifacts rather than the sampling filter. Try resampling up & down several times (never below the original resolution) using PNG or other lossless codec during the intermediate stages -- the difference will be very hard to see with the naked eye.

    Yes it is digital but there are rounding errors.
    Totally insignificant so long as you've got at least 1-2 more bits than your final output, ie 17-18 bits. 24bit and floating point make this a non-issue.

    The key question here is: when my sampler plays back soft samples (e.g. pp violins) and I increase the output volume of the sampler, will I get more bitdepth in my bounced audio?
    No, unless you're bouncing to a track with lower resolution than the samples.

  9. #9

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    You seem well versed in this matter, Richard.

    Quote Originally Posted by Richard Berg
    No, unless you're bouncing to a track with lower resolution than the samples.
    So turning up the volume in the Kompakt player is really just the same amplification as would be applied when moving up the fader in your sequencer's mixer, right? So the method of choice would be to just record your low volume p violins to a low volume audio track and add gain in the mixer..

    Of course when talking soft synths (actual synthesis), you would always try to bouce as hot as possible, since your actually recording new audio, right?

    Roy

  10. #10
    Senior Member Richard Berg's Avatar
    Join Date
    Jul 2002
    Location
    Chapel Hill, NC
    Posts
    232

    Re: Should I increase the volume of sample-based VIs before bouncing to audio?

    So turning up the volume in the Kompakt player is really just the same amplification as would be applied when moving up the fader in your sequencer's mixer, right?
    Yup, so long as the volume in Kompakt is not actually turned down. If it's below unity gain and then you bring it back up with a fader you'll lose a bit of resolution but nothing you could ever detect (even at 16-bit).

    As for softsynths -- it won't hurt to record them as hot as possible, but it doesn't matter nearly as much as it does with analog recording. With analog, the useful dynamic range = peak levels - noise floor. With softsynths, the noise floor is 0 so you'll get plenty of dynamic range regardless (assuming you don't record @ -60dB or something).

Go Back to forum

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •