• Register
  • Help
Results 1 to 7 of 7

Topic: Optimal Sweep Length For Capturing Speaker Cab Responses

Share/Bookmark
  1. #1
    Member
    Join Date
    Sep 2004
    Location
    Poole, Dorset, UK
    Posts
    34

    Optimal Sweep Length For Capturing Speaker Cab Responses

    I'm about to embark on a project to capture impulses of the many guitar cabs I seem to have accumulated over time and at this point just looking at the cabs rather than the amps (i.e. not looking at anything groovy like dynamic convolution for capturing tube amp characteristics) - My ultimate goal here is to have captured sufficiently the cabs I use most infrequently to free up house space, so I'm planning on only doing this once as they'll no longer be around.

    Question for those who have done speaker cab impulses: What length sweep did you use? I'm presuming a sweep is better than a click to allow for the physical response of the speaker, and was thinking of using something like a 0.5 sec sweep to keep convolution processing effort low...

    Any thoughts or tales of experience from those who have already tried...?
    Cheers
    Paul

  2. #2

    Re: Optimal Sweep Length For Capturing Speaker Cab Responses

    Quote Originally Posted by moops View Post
    Question for those who have done speaker cab impulses: What length sweep did you use? I'm presuming a sweep is better than a click to allow for the physical response of the speaker, and was thinking of using something like a 0.5 sec sweep to keep convolution processing effort low...
    Room EQ Wizard is a great freeware that allows you record a frequency response and save the according impulse response. The sweeps that are used are between 5 and 25 seconds ... or so.

    You need to be a member of this site to download it:

    http://www.hometheatershack.com
    All your strings belong to me!
    www.strings-on-demand.com

  3. #3

    Re: Optimal Sweep Length For Capturing Speaker Cab Responses

    There is a huuuuuge difference between using a sweep and a spike. Simply summarized, the time aspect of the sweep and its amount of energy (vs the nearly zero amount of energy in the spike) translates into an acceptable or even very good SNR. Also, for your purpose, the sweep approach is fairly resistent against nonlinear distortions.
    For the sweep (of course) you need a deconvolution program, whereas the results of a spike can be used directly as IR (which IMHO is rather useless).

    But, don't forget: finite impulse responses CANNOT capture nonlinear effects! So, your IRs will capture only part of the sound of your cabinets, but not all of it. Say, you run a nice sine wave through your IRs, don't expect them to return a waveform of an overdriven cabinet sound (typically including compression and distortion, which are both nonlinear).

    For this, you'd need the much more expensive approach of using infinite impulse responses (IIRs)

    (see http://dspcan.homestead.com/files/IIRFilt/zfiltiiri.htm )

    commercially known as "dynamic convolution" (tm of Sintefex) as used in the Focusrite LiquidMix.

    HTH a bit.

    PS: for my Samplicity libs I used 30 second sweeps, which boosted the SNR significantly beyond the level of the reverbs I used.

    Cheers,

    Peter
    Last edited by PeterRoos; 04-20-2008 at 02:44 AM.

  4. #4

    Re: Optimal Sweep Length For Capturing Speaker Cab Responses

    Quote Originally Posted by PeterRoos View Post
    For this, you'd need the much more expensive approach of using infinite impulse responses (IIRs)

    (see http://dspcan.homestead.com/files/IIRFilt/zfiltiiri.htm )

    commercially known as "dynamic convolution" (tm of Sintefex) as used in the Focusrite LiquidMix.
    This is very interesting! I didn't even have any idea that there are finite and infinite IRs.
    So I'm wondering now: would it be possible to sample a guy like the Orban Optimode for instance, using IIRs? And if yes... do you plan to sample such devices someday?
    NorthernSounds.NET.
    View my profile at NorthernSounds.net
    and visit my homesite: www.henrykiwan.com

  5. #5
    Member
    Join Date
    Sep 2004
    Location
    Poole, Dorset, UK
    Posts
    34

    Re: Optimal Sweep Length For Capturing Speaker Cab Responses

    Hey guys, thanks for the responses! I've been away a while so first chance to reply. Thanks Peter: Figured that the sweep was the way to go, so I've just got to experiment with sweep length I guess - I can see that 30 sec sweeps would be king when capturing a reverberant space, but I'm guessing it could be overkill for just capturing a speaker response...plus logically I did think that the longer the IR the bigger the impact in terms of CPU overhead when convolving, so was trying to keep it a little shorter...but how much shorter? Does 3 seconds give a useable time for capturing the physical speaker response, or is the answer 6 secs, or 10, or 20? or even 30??

    I'm a Windows user rather than Mac (no opinions thanks ), so that seems to reduce the number of apps available for deconvolution - Was thinking of giving the Voxengo Deconvolver a go on trial, and then get that if its giving me the results I'm after...though again would take advice if anyone has any alternate views?

    Regarding non-linear aspects, I hear what you are saying, and for this round at least was just going to capture the speaker impulses, though had considered using a form of dynamic convolution capture using Acustic Audio's Nebula which seems to have some kind of capture mechanism (in a utility called NAT) to capture the non-linear aspects of the amp heads..and might well do amp-and-cab-together captures if I go this route in the end...though I quite like the flexibility of being able to chop and change cabs with different amps.

    Many thanks
    Paul

  6. #6

    Re: Optimal Sweep Length For Capturing Speaker Cab Responses

    Paul,

    did you read my post? Room EQ Wizard does sweeps from 5 to about 23 seconds, and does the deconvolution for you already.

    When doing the sweeps be sure to wear ear protection. One unprotected sweep can be one too much.
    All your strings belong to me!
    www.strings-on-demand.com

  7. #7
    Member
    Join Date
    Sep 2004
    Location
    Poole, Dorset, UK
    Posts
    34

    Re: Optimal Sweep Length For Capturing Speaker Cab Responses

    Hi Hannes, I had a high-speed (aka not proper ) look first time round but hadn't noticed that it said you can export the filters' impulse response as a WAV file for use with a convolver until I just popped back after having seen your post...Interesting!

    Actually, it looks like a pretty cool tool all round - Am guessing that given it's title it was originally designed to give you a mechanisms to fix room acoustics...it seems to imply that it allows you to set (or sets itself) filters to correct modal responses, so am guessing it's like IK's ARC system?

    And yes, I hear you...no-one wants to hear that 'final' sweep as the tweeters shoot past you on their way out the door taking your eardrums with them...

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
  •