Firmware supplied IR's

[GiRa]

This post is for subscribing and a reminder for me to post a nearly perfect 4x12"-in-a-room redwirez recipe.

 

[AlbertA]

This means that, effectively, no matter the cab in question, 2m is the "magic distance," yes?

It's not a magic distance, but for a typical guitar cabinet yes you would be in the far field at 2m.

The comb filter effect with a 4x12 is equal no matter how one faces the cabinet, as well as from any distance (on all planes)? If so, I definitely confused directivity with that effect. I was un der the impression that the filtering effect would change depending on where and angle.

It does if you are in the near field; and it's not limited to multiple drivers. Even with a single driver, comb filtering effects show up. This is the reason tiny movements in close mic'ing have profound effects on frequency response.

I am unsure of what the "see above" has to do with my quote. I have had a fumble-y day today, and typically have communication issues, so it is not "on you," but me trying to figure where I went wrong...

In the context of your original post, I believe you were referring to not being able to capture your preferred position based on "comb filtering of multiple drivers". See above meant see above for why this should not be a variable as you thought.

What my "need the former" was referring to was this: Reason being, I want to be able to have the experience of playing through my cabs when using my FRFR system (studio monitors). Currently, it is drastically different, so much so that not using one of my cab is not an option. And that is the primary reason for my obsession with learning all I am able about IR use and capture. Unfortunately, the more I learn, the more I lean toward "not going to happen" for capturing my cabs *properly*.--------phpbb will not let me line break here-------I am always looking for new gear to make new sounds. This was one reason for buying the Axe-Fx II. Unfortunately for me, I have not found any stock IRs (or combinations) that "do it" for me. I am not a fan of playing while listening to a close mic'd cab. Never have been. FF IRs have not thus far sounded/felt like being in my room with a cab either. This is one reason I want some "perfect" IRs of my cabs: so I can find out if there is some failing with IRs, a psychological poison of some sort, or something I have not discovered/thought of yet. If an IR can theoretically provide for a nearly identical experience, then I want to experience it I already have a high quality FRFR system: Blue Sky System One 5.1 system. It is capturing the IRs that I am having trouble with (knowledge, technique, and the room). And being unable to afford to grab the 1x12 boxes I want for the captures. I figured if I had the cabs, I could hit a local studio. But, without the knowledge and technique...Once again, knowledge.

Sure it's not trivial, doesn't mean it's not possible. Also I'm still emphasizing the FRFR system. This can also make or break "the illusion". You may think you have a FRFR system, but have you measured its response?

For lowest latency operation for a process, some sacrifice would need to occur. Are we at a point when no compromise is needed for convolution in real-time? Does the Axe-Fx II have any compromises with its convolution method?

We are already there, the Axe-Fx there's no compromise - it uses all the IR samples you upload to it. Is this concern mostly coming from reverb convolution? Because there you have seconds of samples (i.e. hundreds of thousands of samples) as opposed to 1024 samples and yes inevitably depending on your application you may have to make compromises.

 

[nikki-k]

It does if you are in the near field; and it's not limited to multiple drivers. Even with a single driver, comb filtering effects show up. This is the reason tiny movements in close mic'ing have profound effects on frequency response.

This is what I thought, and what I was trying to communicate regarding a 4x12; mic placement matters. However, with specificity to 4x12s, I figured directivity and comb filtering would be *more pronounced*.. but, that does not *sound right* when I write it. Unsure how to communicate this in writing

In the context of your original post, I believe you were referring to not being able to capture your preferred position based on "comb filtering of multiple drivers". See above meant see above for why this should not be a variable as you thought.

If that is what you thought, then I indeed miscommunicated. I was only referring (solely) to my own inability to produce a *perfect* capture (IR) of any of my cabs, due to lack of knowledge, and thus lack of proper technique, and also due to not having the proper environment (room).

Sure it's not trivial, doesn't mean it's not possible. Also I'm still emphasizing the FRFR system. This can also make or break "the illusion". You may think you have a FRFR system, but have you measured its response?

I am absent of much of my equipment from "studio days." Sometimes I consider myself lucky to still have my oscilloscope.. until I realize I am unable to even lift it onto my non-existent work bench. Boo-friggity-hoo, lol.

The Blue Sky System One 5.1 system is a professional studio monitoring solution, and was known for being flat/linear (but not fatiguingly so, like some famous monitors ), with great imaging. Unfortunately, they can be quite revealing; I have listened to favorite recordings, only to end up disliking the production due to how revealing a good monitoring system can be. However, since my current situation is temporary, and a rental, minimal treatments are in place (real bass traps, some foam (ugh), minimal diffusion), and thus the room has not been "shot" and treated properly (but they are calibrated as per Dolby 5.1 standard). When I purchased them, I demoed a slew of systems, from Genelec to ADAMs; the top-shelf ADAMs won, but were just out of my price range (yes, I actually cried about that), and the ADAMs just under at that time lost to the Blue Sky.

All of that is inconsequential when considering this (as it is a comparative, and is not dependent upon the above): if I record the Axe or one of my amps via mic'd cab, and then compare with similar IRs, the IR fails miserably every time. If I load the IR into a plugin on a track with a recording of a cab-disabled Axe source, I am able to get better results than with the same IR loaded in the Axe (same Amp Block patch), but the IR version is still.. the mic on the cab is still preferable. This is what led to this line of questioning.

And thus prompts me to wonder: If I place a neutral mic (TC30, via a neutral mic pre- Earthworks LAB102) less than a foot from the speaker in my cab.. say, 3".. and I record some playing; how about, create a patch in the Axe, Amp Block only, cabs disabled, and record the mic signal as well as a direct feed. Next, everything still in place, I capture an IR. By utilizing that IR, should the results not be (very) similar to the mic'd track? I am not pushing the speaker hard; I have done this with a Greenback, as well as an EVM12L (so far).

My question is: Should the IR be capable of providing a near-identical experience? Theoretically, under ideal conditions, identical? If yes, then I would wonder.. what would be the reasons for failure? Though my ultimate goal would be for FF IRs, viable NF IRs would suffice, with an "even better" being.. how far can I back the mic off before encountering *interference* from an "improper environment" (room dimensions, technique, etc)?

We are already there, the Axe-Fx there's no compromise - it uses all the IR samples you upload to it. Is this concern mostly coming from reverb convolution? Because there you have seconds of samples (i.e. hundreds of thousands of samples) as opposed to 1024 samples and yes inevitably depending on your application you may have to make compromises.

The concern arises from what I wrote above. Due to my inability to discover an IR or IR combination that will stand up to scrutiny when compared with the Axe tracked via an actual mic'd cab (at any distance/placement I have tried), I decided to discover whether an IR, if even only in theory, should be capable of such comparative. Through such investigation, questions arose.

Is the Axe of sufficient ability to provide for an uncompromised method to provide real-time convolution? If I were to compare with a convolution plugin hosted in a DAW, would the same source IR provide a different experience? Additionally, after reading about "post transient resonances" (the action of the transducer after the sound of the instrument has ceased; MLLSA measurement), I was prompted to ask: is a 1024point IR sufficient? Or, is this a matter of the "time invariant" aspect (inherent) with IRs?

And thus it comes back to my "need for the former": If no 3rd party IR(s) will provide what mic'ing my cabs provides, is it possible to create IRs myself, and if so, under ideal conditions, would they then provide such experience that is near-identical to mic'ing my cab?

I hope I came closer to communicating properly. I have serious issues with communication, and it is greatly exacerbated in non-realtime text exchanges such as this.

EDIT to add: just thought of it- it was the ADAM S3A 5.1 system I tried. Simply the most brilliant, inspiring, tasty, smooth, comfortable, beautiful monitoring system I have experienced. Well, shy of custom far field applications in a properly spec'd and constructed room I love the physical feedback you get from that... mmm....

 

[jakel]

Nikki, I think you're communicating fine. I am sucking up (some) of this great exchange into my pea-brain, benefiting from your thoughts and work (and others too). Thank you.

 

[AlbertA]

Double post...

 

[AlbertA]

All of that is inconsequential when considering this (as it is a comparative, and is not dependent upon the above): if I record the Axe or one of my amps via mic'd cab, and then compare with similar IRs, the IR fails miserably every time. If I load the IR into a plugin on a track with a recording of a cab-disabled Axe source, I am able to get better results than with the same IR loaded in the Axe (same Amp Block patch), but the IR version is still.. the mic on the cab is still preferable. This is what led to this line of questioning.

And thus prompts me to wonder: If I place a neutral mic (TC30, via a neutral mic pre- Earthworks LAB102) less than a foot from the speaker in my cab.. say, 3".. and I record some playing; how about, create a patch in the Axe, Amp Block only, cabs disabled, and record the mic signal as well as a direct feed. Next, everything still in place, I capture an IR. By utilizing that IR, should the results not be (very) similar to the mic'd track? I am not pushing the speaker hard; I have done this with a Greenback, as well as an EVM12L (so far).

Ok, let's focus on this.

So essentially you have tried comparing:

Axe-Fx (Cab block bypassed)-> Real Guitar Cab -> Mic -> DAW -> Blue Sky System
Axe-Fx (With cabblock using captured IR at exact same mic position as above) -> DAW -> Blue Sky System

You can PM me, let's discuss the method used for IR capture.

But yes the result should be very similar if the capture

My question is: Should the IR be capable of providing a near-identical experience? Theoretically, under ideal conditions, identical? If yes, then I would wonder.. what would be the reasons for failure? Though my ultimate goal would be for FF IRs, viable NF IRs would suffice, with an "even better" being.. how far can I back the mic off before encountering *interference* from an "improper environment" (room dimensions, technique, etc)?

Yes it should. How far you can back of the mic depends on your nearest boundary.

Without having the source tracks to hear, it would be hard to say one way or another, but...

It's possible during your IR capture that your room reverberation is significant enough that when you use this IR at its full length (whatever that may be) in your DAW you prefer it to the "truncated" 1024 sample version (since you would loose the short reverberation tail).

The concern arises from what I wrote above. Due to my inability to discover an IR or IR combination that will stand up to scrutiny when compared with the Axe tracked via an actual mic'd cab (at any distance/placement I have tried), I decided to discover whether an IR, if even only in theory, should be capable of such comparative. Through such investigation, questions arose.

Is the Axe of sufficient ability to provide for an uncompromised method to provide real-time convolution? If I were to compare with a convolution plugin hosted in a DAW, would the same source IR provide a different experience? Additionally, after reading about "post transient resonances" (the action of the transducer after the sound of the instrument has ceased; MLLSA measurement), I was prompted to ask: is a 1024point IR sufficient? Or, is this a matter of the "time invariant" aspect (inherent) with IRs?

Yes the Axe-Fx performs convolution of up to 2040 samples in real time, it just convolves the input signal with whatever you uploaded. You can verify this yourself, but in order to explain the difference you are hearing, I would not emphasize the Axe-Fx convolution engine as much (a bunch of us have already verified correct operation) but instead the method of IR capture.

And thus it comes back to my "need for the former": If no 3rd party IR(s) will provide what mic'ing my cabs provides, is it possible to create IRs myself, and if so, under ideal conditions, would they then provide such experience that is near-identical to mic'ing my cab?

Yes they should.

 

[nikki-k]

Thank you for your time, effort, and knowledge thus far... it is truly appreciated

Though cursory, I did accumulate (and felt I understood) enough information, and already possessed a certain amount of knowledge and experience, to realize certain things. Among them, mic distance, room/surface reflections and proximity, the vital nature of the equipment involved, etc. As a matter of pure comparison being afforded with the knowledge/experience afforded, I concluded that mic distances of ~1"-3" would suffice, even if not optimum to provide the most pleasing result to my ear. ~3" is a preferred, blendable distance I have used in the past, so I am familiar with the placement, and finding a sweet spot was not difficult. 1' out or so, and reflections crept in. 1m out, and room modes were audible, albeit faintly so.

My question became: if I have a close-mic position and resultant recording, should.. and would.. a properly performed IR capture/result be capable of *replacing" the mic + cab in the recording process/mix? I put aside the "I want IRs that please me while jamming via a DI'd signal" desire, and concentrated on discovering if the answer was yes, no, or.. possibly. Should my experiment yield success, this would be indicative of high chance of similar success with a FF IR that would not only make me happy for jamming and recording, but the close-mic result(s) would provide additional, blendable IRs. Win, win. Should it fail.. well, it has, and thus I am here seeking help. Mmmm... not "fail," but rather the result was not similar enough. Thus, fault(s) on my part, or a matter of factors the provide for IRs not matching the mic'd track, except under *perfect* conditions (theoretical), or only in cases where knowledge and experience are required to even get close (meaning... out of my reach, at least at this point).

I will PM.
Thank you again

 

[barhrecords]

In my experience,

All the IR's captured in the near field, and I would count 3" as near field, that I have tried have anomolies.

For me, they tend to have a sweet spot in regards to overall SPL where they sound best. And even a sweet spot in relation to where I listen to them back in the room.

Some symptoms for me might be, a preset might sound great in my studio on nearfield monitors, but have a strange comb filtered sound at a venue, etc.

Richard

 

[nikki-k]

Just a quick update: AlbertA is being incredibly kind and working with me offline to help with all of this. Once the "grand experiments" have produced results, I will post back here.

I absolutely LOVE when forum discussion can begin with ignorance (not lazy ignorance, but rather lack of enough knowledge.. and a willingness to admit it), and move in a very positive, and thus productive, direction.

 

[Stratoblaster]

^^^ cool...

 

[GiRa]

This post is for subscribing and a reminder for me to post a nearly perfect 4x12"-in-a-room redwirez recipe.

Uberkab-T75V30-R121-NullAndWall-Room.syx,0.60
Uberkab-V30-SM7-Cap-5in.syx,0.40

 

[AlbertA]

Just a quick update: AlbertA is being incredibly kind and working with me offline to help with all of this. Once the "grand experiments" have produced results, I will post back here.

I absolutely LOVE when forum discussion can begin with ignorance (not lazy ignorance, but rather lack of enough knowledge.. and a willingness to admit it), and move in a very positive, and thus productive, direction.

I shall say the first experiment has been a success (NF IR).

 

[barhrecords]

I shall say the first experiment has been a success (NF IR).

Cool!

 

[AlbertA]

I shall say the first experiment has been a success (NF IR).

So I asked nikki to record 3 things:

1. The sweep of cabinet in question of course,
2. A recording of some guitar playing through the real physical cabinet with the microphone in the same position as when capturing the sweep.
3. At the same time record the direct sound from the Axe-FX being used for 2.

I produced the IR and made a patch in the Axe-FX containing just a single cab block in mono hi res mode, with the input set to USB.

Here's a comparison where I'm flipping back and forth between the original Mic/cabinet track and the direct signal processed with the IR.
http://guitarlogic.org/comps/Mic01-AxeCab-alternating2.mp3

I would say they sound pretty much identical

 

[barhrecords]

So I asked nikki to record 3 things:

1. The sweep of cabinet in question of course,
2. A recording of some guitar playing through the real physical cabinet with the microphone in the same position as when capturing the sweep.
3. At the same time record the direct sound from the Axe-FX being used for 2.

I produced the IR and made a patch in the Axe-FX containing just a single cab block in mono hi res mode, with the input set to USB.

Here's a comparison where I'm flipping back and forth between the original Mic/cabinet track and the direct signal processed with the IR.
http://guitarlogic.org/comps/Mic01-AxeCab-alternating2.mp3

I would say they sound pretty much identical

That turned out nice!

Is the resulting IR a combination of a match eq and deconvolution of the sweep?

Richard

 

[AlbertA]

That turned out nice!

Is the resulting IR a combination of a match eq and deconvolution of the sweep?

Richard

It's Deconvolution of the sweep plus and low shelf adjustment of -3dB at around 1kHz.

I'm waiting for a sweep of a loopback connection to see if that accounts for the adjustment or not.

 

[nikki-k]

Loopback later today.

HUGE thanks to AlbertA once again!!
With access to so much information online, it can actually become difficult to navigate it all, and finding the most recent information can be quite fun. Albert helped shine a light in many areas, but most illuminating is attempting something, discussing the results, and realizing how the math and science actually applies.

Major realization: *Proper* FF IRs require certain minimums, and most of us do not possess them. Even capturing an excellent quality IR with a mic 1' back can be... challenging.

The clip AlbertA posted was a quick mic placement, with little concern for the final result being pleasing or not; it was primarily a matter of finding out (for me) if a close(r) mic placement recording versus an IR created from a capture with the same mic in the exact same position could adequately provide the same sound and feel. In this case, absolutely. But, I would also add, once again, I would not have obtained the same result without AlbertA's expertise. And that was the purpose of my original line of reasoning: If an IR, in theory, should be able to accomplish this, will it do so in actual practice? And if so, to what do I attribute my failure to do so thus far? Even with every scrap of knowledge, IMO it comes down to experience through the application of that knowledge. In this case, to have someone who excels in the field provides more than any book could.

This has changed my world.

 

[barhrecords]

Lovin' this thread!

Nikki, armed with a matched pair of Eathworks and the matching Earthworks preamp, I set out to digging to create some FF IR's myself only to find out the same conclusion. I just don't have the expertise or final pieces to properly get a FF IR.

I still like the discussion and knowledge transfer... good stuff!

Richard

 

[nikki-k]

A big part of the capture is the proper environment. Noise floor, proximity versus room effect, ground plane measurements, etc all play a part, and the further from the speaker/cab one moves the mic, the more vital these become. Compensations, to a certain point, can provide for a decent enough result I think. In this single case, there was a ~+3dB up to about 1k, and AlbertA was able to analyze, and then *edit* the result to compensate well enough to produce what you hear in the mp3.

I am planning another round this coming Tuesday and Wednesday, and will be adding another cab into the mix. The furthest capture I will be attempting will be ~1'. I am not sure it will work well or not.

 

[Search4Tone]

If I load the IR into a plugin on a track with a recording of a cab-disabled Axe source, I am able to get better results than with the same IR loaded in the Axe (same Amp Block patch), but the IR version is still.. the mic on the cab is still preferable. This is what led to this line of questioning.

And thus prompts me to wonder: If I place a neutral mic (TC30, via a neutral mic pre- Earthworks LAB102) less than a foot from the speaker in my cab.. say, 3".. and I record some playing; how about, create a patch in the Axe, Amp Block only, cabs disabled, and record the mic signal as well as a direct feed. Next, everything still in place, I capture an IR. By utilizing that IR, should the results not be (very) similar to the mic'd track? I am not pushing the speaker hard; I have done this with a Greenback, as well as an EVM12L (so far).

My question is: Should the IR be capable of providing a near-identical experience? Theoretically, under ideal conditions, identical? If yes, then I would wonder.. what would be the reasons for failure? Though my ultimate goal would be for FF IRs, viable NF IRs would suffice, with an "even better" being.. how far can I back the mic off before encountering *interference* from an "improper environment" (room dimensions, technique, etc)?


The concern arises from what I wrote above. Due to my inability to discover an IR or IR combination that will stand up to scrutiny when compared with the Axe tracked via an actual mic'd cab (at any distance/placement I have tried), I decided to discover whether an IR, if even only in theory, should be capable of such comparative. Through such investigation, questions arose.

Is the Axe of sufficient ability to provide for an uncompromised method to provide real-time convolution? If I were to compare with a convolution plugin hosted in a DAW, would the same source IR provide a different experience? Additionally, after reading about "post transient resonances" (the action of the transducer after the sound of the instrument has ceased; MLLSA measurement), I was prompted to ask: is a 1024point IR sufficient? Or, is this a matter of the "time invariant" aspect (inherent) with IRs?

And thus it comes back to my "need for the former": If no 3rd party IR(s) will provide what mic'ing my cabs provides, is it possible to create IRs myself, and if so, under ideal conditions, would they then provide such experience that is near-identical to mic'ing my cab?

I hope I came closer to communicating properly. I have serious issues with communication, and it is greatly exacerbated in non-realtime text exchanges such as this.

My experimentation with IRs I've created reveals that the IRs sound closer to the mic on cab when used in a DAW than in the AXE. This has nothing to do with the quality of the AXE convolution. It has everything to do with the length of the IR, the motion of the speaker even after the guitar signal is removed (post transient resonance), the reverberation that occurs within large poorly damped closed back cabinets, and the room reverberation that is captured no matter how much work I seem to put into attempting to dampen reflections even when placing the mic at a close distance. HOWEVER, if I truncate the IRs to 2040 points in my DAW, they sound identical whether used in the AXE or my DAW, and neither truncated IR sounds like the original mic on cab while the non-truncated IRs sound almost identical to the mic on cab. It's unfortunate in my opinion that the AXE can't use longer IRs that can capture the time domain effects of the above. However, by removing the reflections I find it's possible for captured near field IRs to come very close to replicating the mic on cab sound using the AXE as the host for the IR.

With an open back cabinet, a close mic, and a very dead room the length of the IRs becomes less than 40ms and the IRs sound very close to the mic on cab and identical whether they are hosted in the AXE or a DAW.

That's part experience and part speculation on my part. I know the reverb time for a large sealed cabinet is longer than 40ms, and a ported cabinet can have much longer decay times which will be clearly audible, but how much of this sound actually gets picked up by the microphone when such a cab is close miked is dependent on at least several variables. The post transient resonance varies with the type and size of speaker and more importantly the cabinet, with ported enclosures exhibiting long "decay" times, and again I'm not certain how much this contributes to the difference in a truncated IR vs. a non-truncated IR. It's difficult for me to separate these effects but there is something more than just the room that's affecting the length of the IR in my experience and opinion. Take it for what it's worth. Capturing IRs outside is a good way in my opinion to eliminate room reflections from the equation and get an IR having very little if any significant information beyond 2040 points using the right cabinet and speaker.

I'm not an expert by any means on speakers or enclosures as you can probably surmise from my comments but I do have a lot of practical experience with capturing and experimenting with IRs. I have yet to hear a FF IR that sounds convincingly like an "amp cabinet in the room". I've never attempted to create my own but my attempts to use the ones I've purchased have been less than satisfactory. I probably just don't know what I'm doing.

May I ask, what is the best quality software in your opinions for deconvolution and creating IRs? What are you guys and gals using? This question is directed at everyone who has contributed to this very educational thread. This is the most thought provoking thread I think I've come across in a very long time.

Thanks for the informative discussion!

And @AlbertA: Your conversion utility for AXE IR's has been of tremendous help to me over the years. Just thought I'd take this opportunity to say thanks! It rocks!