-
for clean/jazz tones, i am disappointed in IRs. I use a helix modeler and output a non-IR tap to a power amp running through a pair of 1x12 open-back mojotone cabs with Mesa/Celestion black shadow speakers. According to Mesa, this is the classic lead 80 modified with a voice-coil-vent to be able to handle more power.
I love the sound of this rig but have never gotten close to the actual sound of the speakers other than micing the cabs. Micing them comes very close to the sound in the room but despite trying the following IRs in open back cabs and various mic/distance combos:
JBL, EV12L, Altec, Jensen, celestion creamback, alnico 90, greenback, classic lead 80;
I am unable to even come close. The IRs just sound pale, dull and lifeless. If I goose the treble and presense in the amp, I can get a bit of the brightness back but the overall sound is missing the breath of actual speaker.
I spent 6 hours over the last couple days trying to get A/Bing my 1x12 cabs with an IR with the same speaker/cab dimensions. Regardless of mic, mic placement, etc., the IR pales in comparison to the sparkle and breath of my real speakers.
If I put an E609 mic slightly off center on one of the cabs and against the grill and record it, the tone is superior (for a jazz clean tone) than any IR i have tried.
I've concluded that - while modeling tech has evolved at a breathtaking rate - speaker modeling (IRs) are still toddling along at a a 1990 level.
This is not to say that IRs are worthless but I think they have fallen behind amp modeling in terms of coming close to capturing the real sound of an amp rig.
I will say that the line6 speaker cabs are more realistic than traditional IRs and the results I get out of the various cabs in the helix 3.5+ firmware release is better than any of the hundreds of IRs I have purchased.
-
01-12-2024 10:14 AM
-
I have no experience with IR filters but my understanding is that it digitally manipulates the signal by a transfer function found by comparing a "sound" with the input to the chain of equipment that gave that sound.
So the problem may the definition of these inputs and outputs. If for example the IRs you apply only imposes the response of the speaker but you expect it to also impose the response of the amplifier the result may be disappointing.
Technically I believe it should be possible to model a complete amp + speaker + cabinet + the room you are in + your physical listening position, and send that signal to a recording device - and the result will be like being there. That requires that you generate your own IR function.
As said my insight is only theoretical, but from the demos I have heard the results can be very convincing - at least to my ears.
-
My complaint is the size of the rabbit hole that it opens. I have hundreds of IR's that were given to me. Some are just speakers. Some are speakers plus mics. etc. Doing A/B tests becomes an endless process and trying to maintain any kind of controlled process is impossibly difficult. Paralysis of choice sets in fairly early in the process. Honestly, I'd rather have many fewer choices and easier work flow.
-
Regarding optimising micing a little bit I recommend this trick that Brian May talks about here @ 18:35:
Originally Posted by jzucker
I learned this same trick from an engineer who was recording one of my former band's CDs and who had worked with bands like Nazareth. But it works with clean sounds as well, simply turn back the gain. The trick is about the frequency response the mic picks up at a certain position. Combine that with a large-diaphragm condenser picking up the room sound (watch phase relationship between the two signals!) and compress those two signals together minimally with a vintage compressor emulation, e.g. a Fairchild 670 emulation. IIRC Rudy van Geldern would mix in room microphones aiming at the walls but I am not fully sure about that.
-
Having just dipped my toe into IR's for the first time, this comports with my impression. Originally Posted by Jim Soloway
I have what should be a simple modeler, if it can be called that. Boss ME90. Comes with one IR and has slots for three more.
There is an endless array of them available on the Internet, some for free, others not. I started with a few free ones to make sure I could get them to work, which turned out not to be a problem. The ones I saw had no description and, frankly, I'm not sure I could write a useful description of what I'm looking for. Of the four I tried, I didn't hear a lot of difference. If they were actually very different, I'd then have to optimize all the other settings within a pretty complicated signal chain and then compare the four optimized versions. And, hope they'd sound the same on widely varying gigs.
When I bought my first amp in 1964, the store, Silver and Horlan (sp?) had one amp and my choice was yes or no. It was an Ampeg Reverberocket. Volume, treble roll off, dimension and two knobs for tremelo which I never used. Yes.
Features work in the store. Simplicity rules on the gig.
-
Apologies for the long post.. seems I had a lot to say.
To the initial point about sounding the same..it's not going to sound the same. You are comparing your speaker/cabinet/space to an IR of the same speaker in a different cabinet in a different space. You also don't know the details of IR production like the tools applied to produce it. You would have to make an IR of your cabinet in your own space then do an A/B normalizing all the other stuff to consider to get any sort of comparison.
And in any case, the goal is not to replicate. The goal is good tone and that, just like everything else about guitar tone, requires experimentation. First step is to choose a good IR.
I've been using the Strymon Iridium for years. The IR's it comes with (Ownhammer I think) are not great. However, switch to York Audio IR's and given apples to apples (1X12 Deluxe Reverb, same mic, same place) and it's surprising how much better the York Audio sounds.
Next you have to experiment to see which mic and placement you like. I tend towards a Neumann U87 in various positions but many prefer the smoothness of a Royer 421 centered. And for warmth, hard to beat the SM57. Like tone settings, sometimes something sounds better on any given day. Fortunately, like tone settings, it's a knob to turn so you're not stuck with the one thing your amp does. And to those that wonder if you can hear all these subtleties.. yea.. you can. Even with my less than stellar ears.
Then there is speaker configuration.. 1X10, 1X12, 4X10.. they all sound different and you have to find what you like for your guitar, pick, strings, and the sound you're looking for. I find I end up liking the same mic and position but, as you would expect, the cabinets sound quite different from each other. On my current setup (not Strymon) I can also choose a power amp model. I usually prefer a 6L6.
Finally, assuming you're using a clean Class D sort of amplification, there's the question of speaker. The gold standard would be studio monitors but not always a practical choice. I sometimes use a Yamaha DZR10 which is a pretty good speaker. Sounds OK, not great. Go over to a Schertler Roy with across the room backfill from a Schertler Jam150 extension speaker.. that sounds great. The better the speaker set up, the better the Amp model / IR is going to sound.
My current setup is not the Strymon I've been using forever. It's good but oddly.. and maybe just a phase.. I'm liking Joyo better. I'm running an analog preamp model (Joyo American) into Joyo's newish Cabinet Modeler. The IR's I like, the 1X10 Princeton, the 2X12 Matchless, and the 4X10 Bassman are really quite good. Also lets you pick mic/placement/power amp/5 band EQ and a high and low pass filter. $150 including a proper XLR direct out. Not as quiet as the Strymon, but it's OK. Haven't tried loading the York Audio IR's yet.. that will be interesting.
My tone is not, to my ears, dull or lifeless. It's better than my '81 Rivera Fender Concert with an EV in it. It's better than my Roland Artist. It's better than my Quilter head into the Redstone cabinet. And infinitely more adaptable, adjustable, and fun than any of these excellent amps. Took a lot of trial an error but I'm happy with where things are for now. And given the market penetration of models and IR's.. I don't think it's just me.
-
I am brand new to IRs. I understand the process of recording a system's reaction to an impulse and I even know what convolution is. Originally Posted by jzucker
But, I'm still trying to figure out exactly what this amounts to.
Or, stated another way, is this nothing more than very fancy EQ? That is, lots of notches and bumps in the frequency curve to mimic a particular piece of gear? Or is it actually doing anything else, for example, adding tiny bits of delay, or distortion or something?
There are a zillion IRs available. How do I narrow them down to IRs that might help me get a warm jazz tone (as if that's specific enough!)?
-
This is one reason I like the updated speaker cabs that fractal and line6 over. Instead of hundreds of IRs with miscellaneous combinations of mics, mic positions, room size, etc., the newer speaker modeling applets allow you to select the microphone and its position programmatically. Makes it much easier. On the other hand, if the speaker you want is not in their library, you're back to the problem you describe. Originally Posted by Jim Soloway
On update to all this is i picked up a fractal FM3 again and their CAB emulation is much better than line 6. It sounds and feels much more like my actual cab in the room. Not exactly the same but close enough. I'm using their ultrares cabs which are 8k samples. I tried to load one of the bass cabs i purchased that I really like with the helix stomp and fractal gave me an error saying it wasn't high enough resolution. Other cabs I purchased through celestion are ok...
-
I don't expect it to sound the identical. However, I can mic the cab and I bought IRs from celestion of a similar size cab with the same speaker and the impulse responses sound nothing like mic'd cab. In theory, they should at least sound similar. Originally Posted by Spook410
-
"In theory there is no difference between theory and practice - in practice there is." Yogi Berra
-
here's a demo of the fractal FM3 and accompanying EV12L IR (which sounds a lot better than the IRs on the line6 IMO)
https://youtube.com/shorts/ndlQqMSA4PE?feature=share
-
I think one of the problems is that there are just too many variables that go into a mic'd cab for ir's to be a predictable representation. Which mic? is the mic straight on or at an angle? Pointed near the center or closer to the edge? cab on the floor or on a stand? And I'm betting that the group could add a whole lot of other choices ... and that none of that even begins to address the effect produced by the player's attack, the output of the pickup or the gain structure of the rig. That's an awful lot of variable to be addressed in a single snapshot. Originally Posted by jzucker
-
The Bose S1 comes with dozens of presets, apparently "optimized" 31 band EQ. Originally Posted by Jim Soloway
One preset is Les Paul Guitar. Is that Les himself, or Slash?
-
There are all sorts of variables that could create a difference in sound. Variation between same model mics, variation in cabinet construction, variation in same model speakers speakers, subtle mic placement differences, subtle room contributions, preamps/interfaces, even something as simple as having an open back cabinet up against a wall or not can obviously drastically change things.
I don't think it's fair to say the technology doesn't work tho. The goal of an IR is to reproduce the recorded response of a cabinet/microphone/possibly room system, which it measurably does. But that doesn't mean it will recreate the feel of a miced cab, or that it will sound good when applied in a totally different context to that which it was created in.
-
I think they actually have value. I don't use them as a replacement for any specific piece of gear. I just try to find something that might work better than the default speaker settings for a given plugin. I have a few fall back IR's that I think sound decent and I generally use on of them no matter what else I'm using in the chain. It's one of those situations where the big challenge is to not let the perfect be the enemy of the good (or in this case, good enough). Originally Posted by BreckerFan
-
he's talking about IRs not presets. Originally Posted by rpjazzguitar
-
This isn't a subtle difference. It's off by a HUGE amount. Anyway, this is like a religious discussion. I've said my piece. I'll bow out. Take care. Originally Posted by BreckerFan
-
Yes. The point is about how detailed the variables must be. Originally Posted by jzucker
-
I'm not really sure what that's supposed to mean; it's a pretty straightforward principle that in order to make a valid comparison you need to control all the variables. I listed a number of variables that need to be controlled. Any one of them could make a subtle difference, or could make a big difference, or any combination of them could add up to make a big difference. The true test is comparing the ir to the setup that was used to make it, and I've seen plenty of comparisons where the two sound identical, which is to say the technology works just fine. Originally Posted by jzucker
But it's not necessarily enough to just pick same cabinet size, speaker, and mic and expect it to sound pretty much the same as your cab, because there are all sorts of other variables to be accounted for. It may take trial and error, or maybe even creating an IR of your setup, but there are plenty of good sounds to be had.
-
Again: An impulse response is a static representation so it is not really suited well to reproduce a system of dynamically interacting components (e.g. instrument, cable, amp, cable, speaker).
They are OK for e.g. reverbs or delays. The first plugin convolution reverb was Altiverb by Audio Ease which came out around 2000. Before that there was a hardware unit by Sony doing IR reverb IIRC.
Emulating dynamic systems with IRs is possible by interpolating between different IRs but that "dynamic convolution" is patent protected by a patent developed by Sintefex in the 90ies that was bought later by Focusrite and used in their Liquid Mix hardware.
-
I posted this separately yesterday but i just picked up a fractal FM3. The cab technology it uses sounds considerably better than any of the IRs I have been using with the line 6. This is a fender amp model and one of the EV12L cabs. With the fractal cabs, I can get very close to the same sound and feel out of the actual amp (running in an output without the IR) and the direct sound going from another output *WITH* the IR going into my audio device and through my studio monitors.
https://www.youtube.com/shorts/ndlQqMSA4PE
-
that’s all very interesting …. Originally Posted by Bop Head;[URL="tel:1311382"
On the face of it , I would expect
any system for recreating a realworld
guitar amp/cab would need to be ‘dynamic’
ie to work differently for quiet , medium and Loud , and all points in between ….
I mean that’s the whole point
of a guitar amp innit ?
a static system is bound to fail
-
What exactly do you mean by a static representation? It's not just like an EQ curve applied to the sound, it's a frequency sweep of the speaker/mic system, so it can include resonances. And speakers and mics are fairly linear I think unless you're really driving them hard. Originally Posted by Bop Head
-
In a tube amp, the speaker motor and output transformer interact dynamically and the imedance fluctuates with the frequency and amplitude. This doesn't happen with an IR. Originally Posted by BreckerFan
-
I would really like to understand this. Originally Posted by BreckerFan
What is a "frequency sweep" and how does an IR work beyond being a very detailed EQ? Isn't a resonance just a bump in the EQ graph?
Or stated another way, if an IR isn't just EQ, exactly what else is it doing to the signal? Clipping it? Delaying it? What?
Not an argument. I really don't know the answers and am curious.
Convolution is, sort of, solving an equation where the elements are curves rather than numbers. It's like Input times X equals Output, where Input and Output are curves. So my question is, curves reflecting what? Seems like it could be just EQ, but is it?
Reply With Quote
Samick Jz4 update/upgrade
Today, 03:41 PM in Guitar, Amps & Gizmos