Creating impulses.....

DFGeneer, in my case pulse sound was realy, realy quiter - I even need set volume of my amp to maximum...
I never run amp so loud. Sine test tone from Voxengo sound significantly louder. Both files have 0db at maximum.
 
DFGeneer, in my case pulse sound was realy, realy quiter - I even need set volume of my amp to maximum...
I never run amp so loud. Sine test tone from Voxengo sound significantly louder. Both files have 0db at maximum.
I was thinking... wouldn't then the noise of the amp translate into impulse noise (or echo/reverb as it will appear into the time-domain)? Of course you can take care of the nose with a wave editor, but that will also add some artifacts to the impulse itself. Here the convolution technique of preamp/power amp tracks has some advantage, since the noise balances itself in the convolution process, even contributing in a positive way for the impulse response.
I guess I am going to compare the different methods just out of curiosity. However, the easiest one for me will remain preamp/miced tracks convolution.
 
Noise exists in both methods - in case of deconvolution response also need to be cleaned (for "ideal" result). I want to check again and then post clip.
I think noise is not a problem, at least in my case. It just add some reality.
 
thanks, was looking for something not as expensive though ;)

Then use pulse aproach (without deconvolution) as stated above.
However this way will not give good response at high end. Highes slightly lower - see test.

It's really important if you use voxengo deconvolver to check silence and put at least a second , if not your file deconvolved is pretty useless, that's too short to be used.

Actually aditional silence just determine length of resulting impulse, so maybe you will not need impulse of length about second in case of cab.
In worst case created file will be zero length.
 
Actually aditional silence just determine length of resulting impulse, so maybe you will not need impulse of length about second in case of cab.
In worst case created file will be zero length.

I tried this many times in voxengo's deconvolver. if I didn't add a 1 sec silence I had a file of about 30 byte that didn't react as an impulse when loaded in boogex.
While on Waves IR-1 I don't have to mind about this.
 
please consider i'm new to all this. until now i just have played my guitar but would like to get into recording and all that shit. however i have read and at least tried to understand a lot of this...

An alternative (and perhaps more natural) way would be (provided that your amp has parallel fx-loop):

1. Plug your guitar and start your amp
2. Take signal directly after the preamp (fx-loop out) and record it as Track1
3. Take signal from the miced cab and record it as Track2
4. Play a bit while recording the two tracks
5. Add some silence at the end of Track2 (100 ms is perfect, this sets the length of the impulse file)
6. In Voxengo Deconvolver use Track1 as Test Tone File and Track2 as File to Process (Other recommended options are Out bit depth: 24 and check on Normalize to -0.3 dBFS)
7. You may want to fade out the tail of the impulse file in a wave editor in case it catches a lot of noise.

You may now want to apply the thus created impulse on Track1 to see how close it gets to the real thing recorded in Track2. Pretty damn close!

i've only read about the sine sweep impulse technique until now. but this approach above sounds even more interesting. wouldn't this even be more accurate (because it contains lot more frequencies/information)?

wouldn't it be best to play for an hour or so to gether as much information as possible?

wouldn't this be even more accurate than that nebula technique?
(i read the sticky about nebula but did not really get it.)

last but not least: why would you have to add the silence at the end of track 2?

thanks
 
Bumping a 3 year old thread....seriously this is the second necrothread you bumped...LOOK AT THE DATE OF THE POST PLEASE
 
yeah i know these are pretty old however i did not want to create a new thread because i could have been blamed not to use the search function

...also my questions are related to this specific post that DFGeneer did in here
 
i've only read about the sine sweep impulse technique until now. but this approach above sounds even more interesting. wouldn't this even be more accurate (because it contains lot more frequencies/information)?

Probably not. It definitely won't have "a lot more frequencies", as a sine sweep covers every frequency (0hz-20khz) while a guitar amp won't have the deepest lows or highs that a sine wave will. Couple that with any frequency spikes in the amp, yadda yadda... sounds like a cheap way of making an impulse to me.


wouldn't it be best to play for an hour or so to gether as much information as possible?

I guess, but that would lead to a MUCH bigger impulse file. It's a waste of time. Spending extra time to make the impulse in a still inefficient matter as opposed to running a sine sweep through it doesn't seem productive.


wouldn't this be even more accurate than that nebula technique?
(i read the sticky about nebula but did not really get it.)

Nope. An impulse is static, I.E. sampled at one volume, so you will always be missing out on some dynamic information. Nebula samples it as many different dynamic levels as you like. Not to mention the dynamic harmonic distortion.

last but not least: why would you have to add the silence at the end of track 2?

.... Not sure about that?
 
definitely [/I]won't have "a lot more frequencies", as a sine sweep covers every frequency (0hz-20khz)

ok, well then the rest is clear. i thought the sine sweep just would have been one frequency constantly - i.e. at 500hz.
 
that is pretty interesting deLuther. most of the time when listening and trying to distinguish real pwr amps+cabinets vs IRs in some shootouts or comparisons i pay attention to the low end and therefor can say what is what often. i wonder what sweep spans all the IR makers would have used out there.

however i was failing on the nebula vs ir vs real shootouts like most people ;)