Impulse Response Capture BASICS Tutorial - How to Capture your own Guitar Cabinet as an IR
Вставка
- Опубліковано 7 кві 2019
- In today's video, we take a journey through how to capture your own guitar cabinet as an Impulse Response, in a step-by-step tutorial where we walk you through exactly what you will need, and how we went about getting results. There's more than one way to do it, but our way isn't seriously expensive - and should give you enough insight to do it your own way even if you want to do things a little differently.
To understand how to capture an IR, it might help you to understand what an Impulse Response is. As such, the first part in this series explains exactly what an IR is and where we use them. You can find that here:- • Impulse Responses 101:...
The guitar cabinet we use is a custom Zilla 4x12, and the microphones we use are a Shure SM57, Sennheiser 421 and sE Electronics R1 ribbon for the room (the closest available comparison now being the VR1).
Shure SM57:- amzn.to/2UAIQNY
Sennheiser 421:- amzn.to/2Ik07Ef
sE VR1 Ribbon:- amzn.to/2I5mv5c
Zilla Cabinets:- www.zillacabs.com/
------
What gear to buy? We recommend interfaces, microphones, monitors and more at our Thomas (Europe) and Sweetwater (USA) links below
www.thomann.de/gb/thlpg_loh2r...
🍻 BUY US A BEER WITH PATREON!🍻
/ hoppolestudios
Thanks to our Patrons who made this possible - your names are at the end of the video. You make a huge difference!
📱 FOLLOW US HERE 📱
Facebook: / adamsteelproducer
Instagram: / adamsteelproducer
All links are affiliate links
#ImpulseResponse #GuitarTone #IRCapture
-
I've been watching IR capture tutorials over the last couple of days, and yours is the best, by far. Thanks for sharing!
Hi buddy, yeahhhheee. As a physicist and hobby musician I am so happy that here is actually a nice application of what we studied years ago. The Fourier transformation! Sorry guys, but I WANT to write these lines now:
A Fourier transformation transformes a function from the frequency domain into the time domain, or vice versa*. What does that mean? Look at your DAW. The spectrum analyzer. It shows you the actual frequencies but what it "reads" are time dependent signals (functions), a mixure of different instruments and their overtones etc. So the analyzer is basically a Fourier transformer.
Ok, still reading this comment?
So, why using a sweep when creating an *impuls* response?
For a good, real impuls response, you need a perfect impuls, a really high, narrow spike, a delta distribution or Dirac impuls to excite you system (amp^, speakers, mics, mic preamps).
I would guess it's not easy to generate and maybe it would not be so nice for your
eqipment. Whatever. Just a guess.
So, the whole thing is done in the frequency domain, generating a pure sine sweep and record the systems frequency response, with all it's added characteristic
rattles, hums, shimmers, etc.
The voxengo deconvolver needs to do three things (I guess):
1. Fourier transform (FT, time to frequency) the test signal to find out to which frequency the system is responding to and
remember the time when this frequency was "excited".
2. go to the corresponding time in the recorded signal file and FT it (time to frq) it to calculate the frequency response.
We are still in the frequency domain.
Now we know all frequency responses from 0Hz-20kHz(?).
And then ...
3. perform another FT with all these frequency dependent informations and go back to the time domain.
And there you have your short characteristic - click or clack or gligg or ...
Maybe it helps, if you know that a perfect Dirac puls in time leads an absolutely equally distributed frequency spectrum.
I don't know if you ever noticed that: If there is a rather loud crackle in your audio file, or a long constant tone ends, you see
the whole frequency spectrum in your analyer rising up.
Since the recorded cabinet spectrum response was obviously not equally distributed, you have not
just a spike but a click like sound.
Last step:
This information is again convolved with your time dependent guitar+amp signal in real time. Meaning that your signal from the amp is smeared out, blurred, broadend, whatever, in the exact same way like a cabinet would do.
[EDIT]
lpsa.swarthmore.edu/Convolution/CI.html
I selected the "very fast exp" as h(x) and the "1 sec Puls" as f(x).
The "very fast exp", let's say it's your IR and the 1 sec puls is your signal...
The same procedure is used to create reverb IR in cathedrals, famous studios, famous places ...
Actually, each audio system has it's IR: The room, the mics, the console, the AD, the DA, the headphone amp, the headphones ... your eardrum.
*In quantum mechanics it's usefull to convert functions from the momentum to the position space.
^Actually you don't want to capture the amp characteristics, so I guess that's why the studio guys would use a "linear" responding solid state amp.
en.wikipedia.org/wiki/Convolution
en.wikipedia.org/wiki/Fourier_transform
ua-cam.com/video/spUNpyF58BY/v-deo.html
lpsa.swarthmore.edu/Convolution/CI.html
4 years after this video is made it helped me with my problem and now I'm sure i can probably make this work, looks like i was doing everything wrong like #1 using a tube power amp #2 using a white noise sample #3 not knowing about the deconvolve step at the end.. Thank you for this video my buddy sent me ❤
Awesome tutorial dude!
very detailed
Excellent tutorial. Just made my 1st IR with a Vox Pathfinder & an EV PL5. And it turned out great!
Awesome!!! Thanks for taking the time to explain this process. Can’t wait to play around with different cabs, mics, placement, etc
Amazing tutorial man. Followed this to get an IR out of my small combo H&K amp and it turned out really well. Thanks!
Best video on that topic !!!
Hey man thank you so much for this video
helps a lot !!
This was great.. Thanks
True Engineering :)
just awesome. thank you!
fantastic tutorial.
Really the best tutorial on this topic here on UA-cam, great work.
One question, is there a specific reason you used Voxengo and not the reaverb deconvolution?
Thanks for this.
Thanks for that great video, it's a will helpful 👍
great job.
Thank you so much!
Thank you!
Niiiiice peavey XXX super underrated budget tube amp
thanks man. your 2 vids on this made if much clearer and less intimidating than all the others i've seen. thanks again!
i had a similar experience with the ringing trying to capture some echo chambers. i had to add a fade to the end of sweep file to get it out.
I have a complicated chain of equalizers in Windows 7 (Sonic Studio 3, Viper4Windows, DFX, EqualizerAPO), which are all tweaked perfectly to the way I like every sound made by my computer... I can't make all of those effects work when I updated to Windows 10 so I thought "Why not create an impulse response for it and use an IRS file and use it with convolver (in viper4windows) to have the same effect much more easily?"
Recording is super easy... Just plug an AUX cable from speaker out to mic in and make sure the volume is proper, and now I think all I need to use is Voxengo Deconvolver to generate the signal and capture / deconvolve it... Hopefully it works.
Regardless, the vid was helpful :D
excellent tutorial, watched it many times. can you suggest a deconvolution software for Mac? I recently changed my gear. thanks anyways
Dude. Bad Ass.
Very good. I want to try it with my Palmer Combo FAT 50.
I use Reaper to, so it is perfect for me to try.
But i have no extra Amplifier.
I think the easiest way is to go out of Interface directly in my Amp over FX Loop.
So I have not the preamp Section from my Amp captured but only the power section with speaker. Do this work too?
Thank you for great your work 👍
I'm interested in creating Impulse Responses of the cabs I have access to in the practice space (all custom built), so I can enjoy the tone in a couple different scenarios (either thru my amp and a loadbox and/or loading the IRs into my Pod Go / etc)
I understand the software side but questions regarding the solid state power amp to create the sweep:
1) Does the wattage matter? I want to capture both guitar and bass cabs
2) Can anyone suggest a good amp for this purpose?
3) Does output level on the amp or mic gain matter?
Create a direct sound of the guitar and run a loop of it through the amp into the cab.
Once you have moved that mic till it’s exactly correct, proceed to the next step.
Hi Adam,
lately I've been thinking a lot about what you could do with IRs. Would it be possible to make IRs from the Amp aswell? or make IRs from lets say a vocal chain with a 1176, EQ and Reverb?
Love the channel btw!
Hey man! Unfortunately that’s not how IRs work, especially with things like compression- they capture a fixed point in time, not dynamic range. It’s been tried but it always sounds very fake and static because the biggest parts of amps and vintage equipment is both distortion and time-based change- IRs can’t do either of those
@@adamsteelproducer I see! i was thinking that maybe this might be the basis of Kemper because the capture process seems similar.
No the Kemper process is far more complicated and very clever, which is why only Kemper can do that right now....
It’s pretty cool, and I got a lot from it, like making the sweep longer and a higher bit rate, but I think the whole solid state amp can be skipped, you can use the tube amp for it, just bypass the preamp and use the return input in the fx loop and as far as you don’t push too hard those tubes you won’t get much coloring from the tube amp
That’s not how this works. The tube amp should be avoided, not because of distortion but because of impedance curves that interact with the speaker (something solid state amps don’t do) which will completely screw your IR capture up. I’ve tried it. And so have several others who work in the field. Don’t be lazy, it is the difference between a useable IR and one that doesn’t sound anything like you thought it should
@@adamsteelproducer I havent had troubles getting good IRs out of the EVH5150iii as power amp, if you could explain in the final product what would be the factor (sound wise) deterring from using it, cheers
I talk about non-linearities in the video. The impedance from valve amps changes the per frequency, so you will have a capture that sounds nothing like the cabinet alone. If that works for you, you do you - but there’s no way you could consider what you’ve got a faithful reproduction of the cabinet, or expect it to work as a direct replacement
As you are using Reaper, do you have any experiance with Reapers build in deconvolver?
Very interesting tutoarial, thanks Iwonder though, if you captured the response using a clap, a ballon, an sample to trigger the responce, instead of a sine sweep, do you still need to go through the deconvolution process? Cheers
Nice one, but what about the sound card latency between the played-back file and the recorded ones? Should it be taken into account when aligning?
Most good daws compensate for this after recording, and when generating IRs with minimum phase transform any latency is removed anyway
A bit wordy but I wanted to be precise please. At 2:39 on the video runtime you basically said:
BEGIN
"...step # 1 is to get a great sound out of your cabinet. We do this by adjusting the amp until we are satisfied with the resulting tone. Then we adjust the mic placement until we think we've got the right result ..... ". etc. etc.
END
But what sound?? In other words, my immediate goal with using an amp simulator and IR is to make an audio clip of myself performing a short slide guitar solo composition. My go-to sound (in the real world) is primarily a tube amp using an overdrive pedal, a tad-bit of reverb and that's it.
Now, getting back to step #1, in order to get a great sound, should I adjust the amp and mic accordingly to produce the best overdriven slide guitar solo sound or simply set the amp and mic to produce the best clean rhythm chords? I plan to utilize thr clean rhythm sounds as a backing track to my slide guitar solo audio clip. In either case, both require different amp settings and mic placements on the cab.
Therefore, when creating an IR for my cabinet, which sound should I be striving for in step #1, (the clean rhythm chord sound or the overdriven solo sound)?
OR
Or is this a case where I would need to create TWO different and distinct IR's (one for the clean rhythm and one for rhe overdriven solo)?
OR
When making an IR, we always start step#1 by playing CHORDS ONLY and adjusting the amp and mic placement for the best clean rhythm tone. We don't utilize single note solos regardless of how great we've adjusted the amp for that particular tone?
OR
Is this simply all about just getting the mic placement correct and it doesn't matter if we are playing clean rhythm chords or using an overdrive single note solo?
To conclude, what and how is the test-tone sweep being used to create a realistic representation of my cabinet? If I create an IR based on step#1 of using a clean rhythm chord sound then is that all that IR can be used for (is for recording clean rhythm chords)?
And If I create an IR based on step#1 using an overdriven single note slide solo sound then is that all that particular IR can be used for (is for recording overdriven single note slide solo sounds) ?
Thank you for your time.
It's not as simple to explain the science behind it as it is to just make the ir responses. Some people use filter sweeps (which is suppose to be better for amps) and some use transients like a clap. You are capturing the sonic qualities of the speaker itself in this case with a mic. Overdrive would be added after the fact. That's why he used a solid state amp into the speaker cab to get a clean waveform that can be used by the plugins or hardware. He also eliminated reflections that would affect the ir. For more realism some people will record from their amp head into software or hardware that will utilize the ir giving the direct sound of a close mic speaker cab with no influence of the room.
Could I make an IR of my Fender Blues Jr.? I really like this amp in the space I play it in, and would like to have an IR of it I can take anywhere. Also, could I make an IR of that same amp, but add an effect pedal of my BOSS ST-2 Power Stack pedal?
And you can also use these techniques to create reverb IRs too :)
Indeed you can! I just wanted to keep this focused on a single topic to make it easier to find for those looking for it 👍
@@adamsteelproducer I’ve got an old Alesis Quadraverb that Ive been meaning to use to create some IRs so I think that’s going to be my Xmas project :)
I have a question: i used a perfectly flat microphone (Behringer ECM 8000) to have a more precise and real IR, am i right? I just dont need to consider the microphone in my ir
I'm getting saturation when recapturing the sine back with all my FX. Will this be an issue?
Will you capture your room sound also?
Why not use a white Noise pulse?
Room sound is for different tones. And white noise pulses don’t work with guitars as the response is far too short so there is no detail.
Hey Adam, do you need a reamp box to come out of the interface and into the amp?
Not in this case, since it’s a solid state amp with a line input. Guitar amps with instrument inputs definitely do
@@adamsteelproducer alright, thanks! I ask because i shoot my IRs through a Mooer Baby Bomb 30 power amp pedal. Should i use a reamp box in that case?
Depends, does it sound right without? Power amps almost always take a line level (that’s what preamps do, they make line level signals) so it shouldn’t be a problem
I love how British people pronounce the word "tube" like "CHEWB."
I like how americans pronounce "tube" toob
@@sprenzy7936 lol no
its tyeoob in received english
I like how they pronounce tube as "valve"
“Chyewb”*
I like to use IRs from other brands. Now when I load up my amp sim (on REAPER), I like to load up one/two of these IRs into the amp’s cab section. Now when my mac is “breathing” too hard, I would sometimes get this random delay/echo effect after I stop playback. Not sure what’s causing this issue. Is it the IRs, the Cab section, idk.
So in essence I still need to utiliuze a guitar amp simulator (like Amplitube4, Bias Series FX , Overloud TH3 etc.) because an IR is added to the overall sound i.e. the IR you made in this video is added to the chain and is used as my cabinet?
Exactly yes! You can also use a real guitar amp head with a load box for silent recording and use the IR for the cabinet- but yes, the “amp” and the cabinet are separate so you need both to get a realistic sound
Ok, thank you.
what do you think of using a sample of white noise instead of a sine sweep?
Doesn’t work on guitar cabs. The impulse is so short there’s no detail captured
@@adamsteelproducer i think it's not about shortness but about frequencies, and a white noise impulse has every frequency similar to the sinesweep??
It’s about shortness. Because the impulse is so short and resulting reverb short also, you end up with a very high noise floor and lack of detail. You get round this by using a sine wave sweep and deconvolution- there are white papers on the topic
Was hoping to hear the dry track through the digital amp and IR compared to that dry track through the head and mics.
Fair, but I’ve done that in loads of videos so this ones more geared as a straight tutorial rather than a comparison
@@adamsteelproducer Cool beans, just found your channel getting introduced to this subject, I'll have to binge-watch it to get caught up. {:-)
Hi, you’re using solid state, for capturing the cabinet, but what if I want to capture my whole stack?? Can i use an amp what i have? Tube amp, Or hybrid? or is there a different way to capture that? Thanks
You can’t capture a whole amp this way, physics doesn’t work like that and the results will sound very bad- I know, I’ve tried. The only way to capture a whole stack is with very clever systems like a Kemper profiling amp
@@adamsteelproducer Further to this question: What if I wanted that tube poweramp colouring in the IR? Are the results okay from not using a solid state?
@@danshathra the IR captures the frequency response of a system. This system can be a cabinet and or an amp. What an IR cannot do is to capture distortion .
In regards to amp simulators, are the cabinets offered in these products nothing more than "Impulse Responses"? Thank you.
Some of the cabinets in sims are IRs, but some are algorithms written by the plugin makers, which often (in my opinion) don’t stack up in comparisob
Not sure what's under the hood in regards to amp sims, overdrive sims and other FX sims. Are those particular products also both a combination of algorithms and Impulse Responses?
Now for the million dollar question. Which amp simulator product do you recommend (priced within reason - not over $300). I've downloaded a few demos and to be honest I'm not thrilled with the sounds. I can't even locate a realistic (ok, at least a close approximation) of a basic overdrive pedal. That is why I spent the last two days going thru your video tutorial to learn how to make IR's. I'm looking for a simple sound which is a basic Fender blackface with an overdrive pedal. The closest amp I've found that suits my needs was the '67 Blackface V2" offered in "Bias FX 2". The overdrive pedals did not suffice IMHO.
The IR is only the cab part anyway, but I would recommend the Overloud TH-U right now, especially with the American Pack. People are raving about it
Thank you kindly for your suggestion. Great tutorial video!
Did you mean to suggest the Overloud "American Classics"?
overloud.com/products/american-classics
Wow, how you set up Reaper to give color to those frequencies?
Look for “spectral peaks”
@@adamsteelproducer Thank you so much your (superquick) reply! I appreciate it :)
How do I setup Reaper to show the frequencies as colors in the waveforms?
Peak display settings
@@adamsteelproducer Ah, thanks! Never would have found that. Cheers!
Somewhat varying the idea here..... what if you sent a sine wave into (not through) an instrument like a flute, violin, acoustic guitar, pop bottle, etc. - and of course you had a mic or the right type of mics to capture the vibration of the instrument. Would that product an IR of that instrument? Maybe some physical impulse is needed?
@4:04 The "PROBLEM" is that you CAN'T capture the sound of the valve amplifier even if you wanted to because it's non-linear. The non-linearity generates an infinite number of variations of the impulse response so it becomes impossible to record...
Yes, that's right, an IR doesn't capture any non-linear behavior. I think it should be possible to capture nonlinear behavior but it would require multiple IRs, and a convolution filter that would be able to interpolate between them, which would be very CPU intensive.
All of my custom IRs are so quiet thoug, how can I really normalize them? I tried Voxengo and tubeAmp Designer - both output a low output IR
Did you select the normalise option in the voxengo deconvolver?
@@adamsteelproducer Yes, I tried that and it's still half the volume of my regular IRs ...I try to get the sweep as close to not clipping while I'm renedering it but it doesn't make a difference. The IR is always very quiet
thsnkd
Yeah... of course i want to record with my own cabinet,
but i don't have an isolation room, and someone else to run my daw....when you try it, it is like "how am i supposed to hear what i am playing along with ? "
Question: I have crappy amplifiers that are older than dirt. They have some, let's say, "colored" frequency response. Would it not be possible to capture an IR of this piece of junk, then take the inverse to negate the ...crappiness? In other words, if the response at, say, 10KHz, is -4dB, have the modeler pre-compensate 10KHz by +4dB, etc., thereby [attempting to] produce a full-range flat-response sound from a ...piece of junk? Then apply your favorite IR to that, and/or create a composite IR of the pair? (Might be cheaper to spend some money, but... I'm a dumb, cheap, engineer, who's out of money.)
Short answer, no. “Crappy” amps aren’t just limited in frequency, often they have some distortion that happens over a time period (maybe only milliseconds but that’s enough) and that’s something that can’t be removed by an IR.
That doesn’t mean your amp will sound bad, it might make a cool IR! It just won’t be “accurate”- depends what you’re going for
@@adamsteelproducer Thanks, Adam. Appreciate it. Seems like maybe an investment in better amplifiers ought to precede one in an amp/cab sim.
One of the things I've done in a class-D amplifier design (work, not audio) is compensating non-linear effects on a sample-by-sample basis. It's primarily a single frequency, and THD improved dramatically. It's basically linearizing the D/A conversion. I wonder if I could do something similar for audio. I suspect it'd help single tones and hurt music.
What kind of spec IR does this produce? I'm guessing a 96khz one!? But how long and how many points?
Depends on your setup, it’s open ended. The “points” aren’t constrained, unless you want them to be to fit a particular software or hardware unit
@@adamsteelproducer thanks for the reply. It's just that often commercial packs come in many different formats, but there seems to be very little noticeable different. I don't see an option in Deconvolver to set the length for example
Thats all done afterwards if needed to suit different output formats. Generally a higher sample rate capture makes for a better resample
@@adamsteelproducer gotcha. Thanks!
So it seems that an IR is basically:
(signal + equipment) - signal = IR
.
The computer takes the original signal and subtracts that from the signal colored by the equipment. What's left is only the equipment color.
You have the right concept!
Deconvolution is actually it's own mathematical process in signals and systems, an electrical engineering concept.
We know the input, and the output, and the program is using deconvolution to figure out the "function" in the middle that changes the input to the output.
A question...
When recording the IP at 96 KhZ do you
Ahh, I should have watched the whole video before commenting.
Yes, you do need to set your DAW to the same sample rate...lol
(Thinking of changing my Patreon support to you. More of the same and I'll do it ;-) )
is the IR too long? cuz it sounds weird
You should of shared the sweep file that was generated.
Why? You can generate your own, for free, to your own specs very easily.
@@adamsteelproducer Sorry I wasnt meaning the IR's, did my first attempt today and your tutorial really helped with that, but what I was meaning was the sweep file to put into reaper that you generated from Voxengo.
I ended up using the twonotes Blend IR which has it built in which seemed to work pretty well for my first try.
Is this technically an impulse response? Is a sweep an impulse?
A sweep is not an impulse, but using Deconvolution in software it is converted into an impulse response. This middle step is necessary with guitar cabinets to capture the detail and nuance of the cabinet.
One of the better tutorials but mannnn that sounded bad at the end. im not excited lol
If you don’t like the sound I got, that doesn’t change the method or its validity
There is a far much quicker technic than sweep sine wave. Just shot a Dirac pulse through the cab and record it as you would do with the sweep. Done. Just edit edit the start and end to 500ms, and it's done. When capturing a full cab when multpile mic on every position, it's much more time saving.
Doesn’t work with cabinets in my experience, there’s not enough detail and too high a noise floor. The sine wave sweep is designed to increase resolution and works well, so has become a standard because of its upsides
@@adamsteelproducer Make sense for room measurement but not for speaker impulse response. You don't need to excite the room as you would do for reverb impulse. You just measure speaker frequency response. The tail doesn't really matter as impulse is 200ms and even 2O or 40ms for the Helix.
I know both way to do them & did the test both way on fft that's rigoursly the same curve.
They even nulls out with sinesweep through both of them on Ozone Match eq Soundwise it's sound the same as confirmed by the FFT analysis. If am not mistaken that's how ML Soundlab do it, as Andrew wade and Joey Sturgis.
Anyway, whatever the method, it works