24 hours later... So, is the squirrel gone???

[HonestOpinion]

I highly doubt the majority or users are using IRs. In the Facebook user group, which has 11,000 members, every time it’s been asked, I believe stocks cabs have come out on top. Many people still don’t really understand what an IR even is. Short of actual data, I guess we’ll have to agree to disagree, but I think the average user is still relatively unsophisticated.

 

The Helix stock cabs are great and the ability to change the mic-type, distance, and add early reflections is fantastic. If they add any more features like off-axis to the Helix cabs it will become even less compelling to wade through IRs with so many options at your fingertips with the stock cabs.  Not to speak of the fact that you can change the Native cabs' parameters from snapshot to snapshot. You can of course also change the IR used from snapshot to snapshot.

 

Is it possible that you may be underestimating the number of people using at least a few IRs with more users coming on board as they get more familiar with the Helix? I think arguably the IMHO very limited and ponderous interface for using IRs (not enough slots, too short a namespace, lack of ability to see which IRs are in use, etc.) as well as the tedium prior to Helix Edit 2.3 involved in the IR backup and restore process required for upgrades has hampered the use of IRs in the Helix to some extent. Remember having to reconstruct according to a snapshot or alternatively having to rename and number all your IRs in order to be able to restore them to the correct locations.  A better IR implementation may eventually increase the number of IR users. Many IRs appear to still be selling like hotcakes, are a great tool for finding new sounds, and they were a big selling point for the Helix.

[jbuhajla]

Also keep in mind, Helix stock cabs ARE impulse responses. Line 6 just gave them some more tidy controls (mics/distance/ER) that change them behind the scenes rather than just scrolling through a bunch of IR files. 

[willjrock]

No offense but im not sure what point that makes other than to say "hey all you guys saying you prefer IRs to hlx cabs, are effectively using IRs, when you employ the use of a Helix cab.

i think that the bigger point here is evident in that a lot of us just dont think helix cabs sounds good nor do they hold a candle to the better 3rd party IRs. respectfully. 

[DunedinDragon]

No offense but im not sure what point that makes other than to say "hey all you guys saying you prefer IRs to hlx cabs, are effectively using IRs, when you employ the use of a Helix cab.

i think that the bigger point here is evident in that a lot of us just dont think helix cabs sounds good nor do they hold a candle to the better 3rd party IRs. respectfully. 

 

I think this really is an important point.  I use both hlx  cabs and IRs, but the HLX cabs have a long way to go to be able to capture the differences between a Celestion V30 and a Redback speaker, which can really be important for certain types of sounds.  Also increasing distance is not the same as being off axis or further out from the cap at a close distance.  IMO it's not that the helix cabs can't be made to sound decent, it's really the fact that they can't be adjusted as accurately as can a decent IR.

 

Plus given some of the really good mic and speaker mixes provided as a single IR by many of the IR vendors and you have a LOT of capability that can be accessed conveniently.

[Jos_K]

No offense but im not sure what point that makes other than to say "hey all you guys saying you prefer IRs to hlx cabs, are effectively using IRs, when you employ the use of a Helix cab.

i think that the bigger point here is evident in that a lot of us just dont think helix cabs sounds good nor do they hold a candle to the better 3rd party IRs. respectfully. 

Maybe it's not a matter of good and better, but getting you to sound and feel in your head. The Helix (stock) cabs get me where I want to be. Does that make my sound worse, 'not being able to hold a candle' to a sound with IR's? Not at all. I have some IR's, but after a while I noticed the patches I created with them, evolved into sounding almost exactly as the ones with stock cabs.

It starts with a big wow. But after a while you tweak a little, maybe add some EQ and a month later you go back to the original patch with the stock cab, noticing it sounds almost identical to the patch with the IR. I've had this with cabs, but with amp models too. Going in circles.

[gunpointmetal]

Maybe it's not a matter of good and better, but getting you to sound and feel in your head. The Helix (stock) cabs get me where I want to be. Does that make my sound worse, 'not being able to hold a candle' to a sound with IR's? Not at all. I have some IR's, but after a while I noticed the patches I created with them, evolved into sounding almost exactly as the ones with stock cabs.

It starts with a big wow. But after a while you tweak a little, maybe add some EQ and a month later you go back to the original patch with the stock cab, noticing it sounds almost identical to the patch with the IR. I've had this with cabs, but with amp models too. Going in circles.

I think this is true because everyone tweaks to their taste. For me, I can get the sound I want with either and they won't be much different from each other. The difference is I spend a lot less time getting there with an IR. I still use stock cabs on a lot of clean sounds.

[willjrock]

DunedinDragon, on 07 Dec 2017 - 05:52 AM, said:

I think this really is an important point.  I use both hlx  cabs and IRs, but the HLX cabs have a long way to go to be able to capture the differences between a Celestion V30 and a Redback speaker, which can really be important for certain types of sounds.  Also increasing distance is not the same as being off axis or further out from the cap at a close distance.  IMO it's not that the helix cabs can't be made to sound decent, it's really the fact that they can't be adjusted as accurately as can a decent IR.

 

Plus given some of the really good mic and speaker mixes provided as a single IR by many of the IR vendors and you have a LOT of capability that can be accessed conveniently.

Well yeah...It IS important, but it doesnt change anything. It doesnt have any impact on the matter. They are what they are. Defining their origin simply doesnt have any purpose in my view. Not to mention thats probably the third time ive heard it said exactly like that.  :)

 

Jos_K, on 07 Dec 2017 - 07:57 AM, said:

Maybe it's not a matter of good and better, but getting you to sound and feel in your head. The Helix (stock) cabs get me where I want to be. Does that make my sound worse, 'not being able to hold a candle' to a sound with IR's? Not at all. I have some IR's, but after a while I noticed the patches I created with them, evolved into sounding almost exactly as the ones with stock cabs.

It starts with a big wow. But after a while you tweak a little, maybe add some EQ and a month later you go back to the original patch with the stock cab, noticing it sounds almost identical to the patch with the IR. I've had this with cabs, but with amp models too. Going in circles.

I wont disagree with you because i dont know what youre up to over there, but most often as of late, i see guys wanting to copy sounds, and not much of going after a sound in their head,so i guess it depends on what the challenge is.  What i do know is that there are sounds that hlx cabs are simply not capable of getting. Maybe with a bunch of sussing around, a bunch of added EQ time, and/or tone matching capabilities you may get there, but who wants to spend their time doing that? Heres one that ive no luck at all in reproducing without IRs. How about you? Want to give it a shot?

 

https://www.dropbox.com/s/afwfhcue3xq3cqf/fooled%20test.wav?dl=0

 

You cant simulate the outboard gear used, the engineer, their room, their mics, or all of the other factors that make IRs sound different from one another (not effectively anyway in the sense we're talking about ) This is why there is room in your library for the same cab from redwirez and ownhammer, because they sound different.

Sounds DO vary in quality. A lot of us just happen to believe that hlx cabs are on the bottom end of the chain.

[jbuhajla]

No offense but im not sure what point that makes other than to say "hey all you guys saying you prefer IRs to hlx cabs, are effectively using IRs, when you employ the use of a Helix cab.

i think that the bigger point here is evident in that a lot of us just dont think helix cabs sounds good nor do they hold a candle to the better 3rd party IRs. respectfully. 

The point was that the on board cab in the Helix is not some sort of inferior cabinet modeling technology, they are impulse responses the same (technically) as the rest of the 3rd party IRs. Some people like the onboard, some don't. I use both. 

 

There are similar debates about the tuner, some like it and some don't. Ain't no biggie...

[brue58ski]

Also keep in mind, Helix stock cabs ARE impulse responses. Line 6 just gave them some more tidy controls (mics/distance/ER) that change them behind the scenes rather than just scrolling through a bunch of IR files. 

 

Are you sure that's true? I seem to recall a Line 6 person saying they were similar to an IR but actually worked a different way. I'm probably stating this wrong but I thought it was something like that.

[Jos_K]

 

DunedinDragon, on 07 Dec 2017 - 05:52 AM, said:

Well yeah...It IS important, but it doesnt change anything. It doesnt have any impact on the matter. They are what they are. Defining their origin simply doesnt have any purpose in my view. Not to mention thats probably the third time ive heard it said exactly like that.  :)

 

Jos_K, on 07 Dec 2017 - 07:57 AM, said:

I wont disagree with you because i dont know what youre up to over there, but most often as of late, i see guys wanting to copy sounds, and not much of going after a sound in their head,so i guess it depends on what the challenge is.  What i do know is that there are sounds that hlx cabs are simply not capable of getting. Maybe with a bunch of sussing around, a bunch of added EQ time, and/or tone matching capabilities you may get there, but who wants to spend their time doing that? Heres one that ive no luck at all in reproducing without IRs. How about you? Want to give it a shot?

 

https://www.dropbox.com/s/afwfhcue3xq3cqf/fooled%20test.wav?dl=0

 

You cant simulate the outboard gear used, the engineer, their room, their mics, or all of the other factors that make IRs sound different from one another (not effectively anyway in the sense we're talking about ) This is why there is room in your library for the same cab from redwirez and ownhammer, because they sound different.

Sounds DO vary in quality. A lot of us just happen to believe that hlx cabs are on the bottom end of the chain.

 

Thanks for the challenge! I'll get to it. might take a couple of days as I'm quite busy, but i'll get back on this one.

 

Just to be sure: no double tracking or other stuff that's impossible to reproduce in Helix going on, right?

[willjrock]

Thanks for the challenge! I'll get to it. might take a couple of days as I'm quite busy, but i'll get back on this one.

 

Just to be sure: no double tracking or other stuff that's impossible to reproduce in Helix going on, right?

The guitars are double tracked but thats just part of recording any guitar tone.

[kylotan]

To get back on the original subject - just the other day, I was trying to reproduce some of my favourite Pod XT patches with the Helix - going through matching up the amps, cabs, effects, etc. And it's actually incredibly hard to do because the Helix has far, far less top end by default. So much so that I wouldn't be surprised if they deliberately added an extra high-end cut to avoid complaints about "fizz" (or indeed "squirrel"). If anything I wish there was something to bring back some of the 'air' that is missing.

 

I do think a lot of the debate is mostly cork-sniffing and confirmation bias - people listen to their Helix through high quality monitors containing tweeters pointing at ear level, and are surprised to hear more high-end than when they're playing through a floor-level cabinet or combo on the floor, and they put this down to the modelling as they expect that to be harsh and fizzy, when in fact a lot of distorted guitars are harsh and fizzy until the cabinet and the listening environment dulls it a little.

[jbuhajla]

Are you sure that's true? I seem to recall a Line 6 person saying they were similar to an IR but actually worked a different way. I'm probably stating this wrong but I thought it was something like that.

I was thinking it was DI that posted that, but I can't find the post. It could be just my middle aged brain matter conjuring up things...

[Guitarmaniac64]

Also keep in mind, Helix stock cabs ARE impulse responses. Line 6 just gave them some more tidy controls (mics/distance/ER) that change them behind the scenes rather than just scrolling through a bunch of IR files. 

No they what Line 6 call ARE HYBRID IR

That doesnt mean it is real IR,s

 

https://en.wikipedia.org/wiki/Impulse_response

 

A real iR cant be messed with unless your IR loader will have some effects or EQ or whatever but there is NO IR loader that have the option to change Mic,Distance,Early reflection on a static IR..

 

That is beacuse an IR is a snapshot of a miced up cab where the mic is set to one position..

 

You can do IR of rooms but as soon as you move the mic you have to shoot another IR as the sound will not be the same.

 

[Jos_K]

I wont disagree with you because i dont know what youre up to over there, but most often as of late, i see guys wanting to copy sounds, and not much of going after a sound in their head,so i guess it depends on what the challenge is.  What i do know is that there are sounds that hlx cabs are simply not capable of getting. Maybe with a bunch of sussing around, a bunch of added EQ time, and/or tone matching capabilities you may get there, but who wants to spend their time doing that? Heres one that ive no luck at all in reproducing without IRs. How about you? Want to give it a shot?

 

https://www.dropbox.com/s/afwfhcue3xq3cqf/fooled%20test.wav?dl=0

 

You cant simulate the outboard gear used, the engineer, their room, their mics, or all of the other factors that make IRs sound different from one another (not effectively anyway in the sense we're talking about ) This is why there is room in your library for the same cab from redwirez and ownhammer, because they sound different.

Sounds DO vary in quality. A lot of us just happen to believe that hlx cabs are on the bottom end of the chain.

 

 

https://drive.google.com/open?id=1-yIHkFoiKzjTQGlMjmEZPbjAWGUckQXr

Here's what I came up with. Like I said, I didn't have a lot of time, but it took me only 15 minutes to create this sound. Plexi bright, a little EQ (Simple EQ) and Greenback 25 cab with 409 mic. some early reflections, no hi/lo cuts (or maybe lo cut at 90 Hz, can't remember, forgot to save  :blink:).

Of course I hear (small) differences, but to say the file you provided sound 'better' or 'doesn't hold a candle', no. If I really really, wanted to nail that sound, there are many more variables I'd worry about, like guitar/pickups (I used a Yamaha Revstar, which is a darker sounding instrument), amp/settings and so on. And then there's always me vs the recorded player. But to me that is pointless because the context you play in constantly changes. I think there's no point in trying to clone a sound 100% if the band doesn't sound 100% the same too. Even the same band, playing the same thing always sounds different.

[HonestOpinion]

No they what Line 6 call ARE HYBRID IR

That doesnt mean it is real IR,s

 

https://en.wikipedia.org/wiki/Impulse_response

 

A real iR cant be messed with unless your IR loader will have some effects or EQ or whatever but there is NO IR loader that have the option to change Mic,Distance,Early reflection on a static IR..

 

That is beacuse an IR is a snapshot of a miced up cab where the mic is set to one position..

 

You can do IR of rooms but as soon as you move the mic you have to shoot another IR as the sound will not be the same.

 

 

 

Also keep in mind, Helix stock cabs ARE impulse responses. Line 6 just gave them some more tidy controls (mics/distance/ER) that change them behind the scenes rather than just scrolling through a bunch of IR files. 

 

Did it get posted somewhere as to exactly how Helix cabs were implemented? I am not sure from Jbuhajla's comment exactly how the native cabs operate. If I remember correctly I have seen comments from Line6 saying the native cabs are optimized, not sure what that means. Here are a few ways that have been speculated on that Line6 could have used to implement their cabs - methods which allow mic type, mic distance, and early reflections changes. Maybe they used some other technique that Line6 has not detailed or that has not been speculated on anywhere I have seen.

• Method 1 - Helix cabs are identical to other IRs, The mic type, mic distance, and early reflections parameters are wrappers for the end user that actually transparently cycle through separate IRs under the covers for each possible combination of mic-type, distance, and early reflections. Each cab is actually a collection of hundreds(thousands) of IRs that take into account every possible parameter permutation and changing parameters simply selects the corresponding IR. This method seems highly unlikely and implausible to me. That is a lot of permutations and a lot of IRs that would have to be stored. I would say there is little to no chance this is the way things are being done but anything is possible - this method, not so much.

             Implications: This would require substantial memory dedicated to many IRs. Each cab added in an update would cumulatively add up to potentially significant additional storage space. Seems unlikely and inefficient and would limit the number of cabs that could be added to the Helix without devoting a lot of valuable memory. I think there is zero chance that the Helix cabs use this methodology.

 

•  Method 2 - I suppose a likely hybrid method could be used where, for example, there was a different IR for each mic type, and mic distance and early reflections was an applied effect after the fact. In this hybrid example, hidden from the user, a different IR would be selected for each mic type and then the mic distance and early reflections would be applied as "effects" after the fact to that IR.

             Implications: This method would reduce the number of IRs required as the number of permutations dictated by mic-type would be far less then if you required a different IR for every possible combination of mic-type, mic distance, and early reflections. Every IR gets treated with the same mic distance and early reflections simulation algorithm. 

 

• Method 3 - Helix cabs are modified "optimized" IRs with some hooks that respond appropriately to external processing algorithms such as the ones that change mic type, distance, or early reflections. Again, this seems unlikely to me as an approach. It would require taking a .wav file and customizing it into something else but who knows.

            Implications: You are now dealing with a file that is different from the conventional .wav file used for IRs which is strictly an audio format and "filter" fed to an IR engine for processing the guitar signal. You would now be dealing with either some form of modified .wav file or a completely different file format that could be modified and/or responded to by other algorithms such as ones for mic distance. Purely speculative but if you are dealing with modified "optimized" IRs that have additional data than just the standard IR/wav file does this still have to fit in the same 1024 or 2048 space as a conventional IR. If so do the bytes allotted to optimizations allowing the extra features impact the overall quality or detail of the .wav. Depending on implementation the impact to IR detail could be negligible or non-existent.

 

• Method 4 - Helix cabs are simply a regular IR created with a 'neutral' mic and then "colored" appropriately by the Helix's mic type, mic distance, and early reflection algorithms/parameters.

           Implications: This, similar to Method 2, would seem like a highly likely approach to me although Line6's comments about Helix cabs being optimized IRs does not seem to jibe with this approach which simply uses a standard neutrally oriented IR and applies post processing for the various cab parameters (mic-type, distance,  reflections).

 

 

So... in summary, I have no clue how Line6 has implemented their cabs but if they are simply "regular neutral" IRs with post processing algorithms applied for mic-type, distance, and reflections that begs the question that perhaps some of these algorithms could be applied to and opened up to third-party IRs. Applying Line6's early reflections to a third-party IR for instance. It might not make a whole lot of sense to apply a Line6 Royer mic emulation to an Ownhammer IR that had been created with an SM57 mic but I suppose it could be done. Perhaps Line6 has pioneered a whole new way of creating IRs that allows them to be "parameterized" and processed in new ways by the IR engine. If so, this technology could provide the entire IR community with a much more efficient, enhanced, and flexible method of sorting through IRs if it was applied as an industry standard.

 

Maybe that is where we are headed while currently different manufacturers (e.g. Line6, Fractal) offer different IR parameters that can be modified. One standard IR protocol would offer many mic, distance, and potentially other parameter choices across different products instead of requiring a different IR for each parameter change. Similar to the way a Helix cab functions. One has to wonder though if "parameterized" IRs are or would be as faithful as one that is created recording the actual mic selected at the distance, axis, etc. indicated. My gut says it would take a long time to get certain parameters like mic-type in "parameterized" IRs sounding as authentic. Perhaps all the actual recorded mic-types could be combined into a single file and then parameterized instead of requiring separate files. Anyway, IRs feel to me like somewhat of a moving target and an evolving technology/art right now.

 

One conclusion I have come to while stepping through some of the possible methods Line6 might have used to create their cabs is that running a Helix cab is not "the same" as running an IR. Either it has an essentially different DNA than a standard IR .wav file or it is a standard although perhaps "neutral" sounding IR(depending on whether or not Line6 is using different IRs for different mics) that is receiving custom Line6 post processing that the Helix does not allow to be used on third-party IRs. Either way this makes Helix cabs a somewhat different animal than a typical third-party IR. Even if the only difference is that the Helix's native cabs are IRs that have access to some post-processing that third-party IRs do not. This conclusion could admittedly be negated by some method that was not listed here and changes the playing field.

[zooey]

Well said HO, agree on all counts.

 

I'm also glad you brought this up in general, with rumors going around of more specific info that AFAIK we don't actually have, and probably never will, as it's part of the Secret Sauce .

[jbuhajla]

I did find this. It sounds like the "hybrid cab" is an IR plus algorithms for the associated changeable parameters. So it is an Impulse Response plus algorithms that make the IR adjustable for mic, mic placement and room reflections. 

 

https://l6c-acdn2.line6.net/data/6/0a020a3d43095900a82ce2652/application/pdf/helix-blog-hybrid-cabs.pdf

[datacommando]

Each cab is actually a collection of hundreds(thousands) of IRs that take into account every possible parameter permutation and changing parameters simply selects the corresponding IR. This method seems highly unlikely and implausible to me. That is a lot of permutations and a lot of IRs that would have to be stored. I would say there is little to no chance this is the way things are being done but anything is possible - this method, not so much.

Wow, HO now my head hurts! That concept really would cost a small fortune to implement! Never gonna happen!

So, a serious question regarding this speaker model versus Impulse Response thing.

Would you have any ideas as to how IK Multimedia manage to create something like the 3D Cab Room in AmpliTube 4. It gives you the ability to shift the position of 2 mics (from a selection of 18), change or combine the type of (29) speakers in any cab and have these in a variety of 6 different room spaces. Plus all the additional stuff, stomp fx, this can be run on my iPad or even iPhone for less than $100.

How do they do that?

[HonestOpinion]

Well said HO, agree on all counts.

 

I'm also glad you brought this up in general, with rumors going around of more specific info that AFAIK we don't actually have, and probably never will, as it's part of the Secret Sauce .

 

Thanks and I agree, at no point am I expecting Line6 to chime in with the family recipe. I would be fascinated though to know more of the nuts and bolts of how different manufacturers are handling IRs.

[HonestOpinion]

I did find this. It sounds like the "hybrid cab" is an IR plus algorithms for the associated changeable parameters. So it is an Impulse Response plus algorithms that make the IR adjustable for mic, mic placement and room reflections. 

 

https://l6c-acdn2.line6.net/data/6/0a020a3d43095900a82ce2652/application/pdf/helix-blog-hybrid-cabs.pdf

 

Great reference article, thx! Doesn't exactly tell you how their parameters are being rendered nor would I expect that but sheds a little light on the subject nonetheless.

[HonestOpinion]

Wow, HO now my head hurts! That concept really would cost a small fortune to implement! Never gonna happen!

So, a serious question regarding this speaker model versus Impulse Response thing.

Would you have any ideas as to how IK Multimedia manage to create something like the 3D Cab Room in AmpliTube 4. It gives you the ability to shift the position of 2 mics (from a selection of 18), change or combine the type of (29) speakers in any cab and have these in a variety of 6 different room spaces. Plus all the additional stuff, stomp fx, this can be run on my iPad or even iPhone for less than $100.

How do they do that?

 

No clue but your comment definitely feeds into the potential for an evolving industry standard that would combine the innovations from multiple manufacturers into a robust and flexible IR format that allows many parameters to be easily modified across diverse devices and does not require wading through so many files. I realize there is big competition and constant innovation in music production but other industries have managed to come up with common protocols and standards although it feels a bit like a pipe dream right now in the audio world. If only...

[datacommando]

No clue but your comment definitely feeds into the potential for an evolving industry standard that would combine the innovations from multiple manufacturers into a robust and flexible IR format that allows many parameters to be easily modified across diverse devices and does not require wading through so many files. I realize there is big competition and constant innovation in music production but other industries have managed to come up with common protocols and standards although it feels a bit like a pipe dream right now in the audio world. If only...

HO,

True, this IR stuff seems to be evolving at a pace.

Another product that intrigues me is Recabinet from Kazrog. It has been around for a while and I have have had this product from version one. These guys are using IRs but adding something in the way of dynamics that adds “speaker breakup†emulation. You also have either on or off axis mic positions, beside 3 distance options, near, mid and far.

Helix has the mic placement distance thing, but I think the on/off axis idea would be a great addition to the Helix cab section.

[Guitarmaniac64]

Did it get posted somewhere as to exactly how Helix cabs were implemented? I am not sure from Jbuhajla's comment exactly how the native cabs operate. If I remember correctly I have seen comments from Line6 saying the native cabs are optimized, not sure what that means. Here are a few ways that have been speculated on that Line6 could have used to implement their cabs - methods which allow mic type, mic distance, and early reflections changes. Maybe they used some other technique that Line6 has not detailed or that has not been speculated on anywhere I have seen.

• Method 1 - Helix cabs are identical to other IRs, The mic type, mic distance, and early reflections parameters are wrappers for the end user that actually transparently cycle through separate IRs under the covers for each possible combination of mic-type, distance, and early reflections. Each cab is actually a collection of hundreds(thousands) of IRs that take into account every possible parameter permutation and changing parameters simply selects the corresponding IR. This method seems highly unlikely and implausible to me. That is a lot of permutations and a lot of IRs that would have to be stored. I would say there is little to no chance this is the way things are being done but anything is possible - this method, not so much.

Implications: This would require substantial memory dedicated to many IRs. Each cab added in an update would cumulatively add up to potentially significant additional storage space. Seems unlikely and inefficient and would limit the number of cabs that could be added to the Helix without devoting a lot of valuable memory. I think there is zero chance that the Helix cabs use this methodology.

• Method 2 - I suppose a likely hybrid method could be used where, for example, there was a different IR for each mic type, and mic distance and early reflections was an applied effect after the fact. In this hybrid example, hidden from the user, a different IR would be selected for each mic type and then the mic distance and early reflections would be applied as "effects" after the fact to that IR.

Implications: This method would reduce the number of IRs required as the number of permutations dictated by mic-type would be far less then if you required a different IR for every possible combination of mic-type, mic distance, and early reflections. Every IR gets treated with the same mic distance and early reflections simulation algorithm.

• Method 3 - Helix cabs are modified "optimized" IRs with some hooks that respond appropriately to external processing algorithms such as the ones that change mic type, distance, or early reflections. Again, this seems unlikely to me as an approach. It would require taking a .wav file and customizing it into something else but who knows.

Implications: You are now dealing with a file that is different from the conventional .wav file used for IRs which is strictly an audio format and "filter" fed to an IR engine for processing the guitar signal. You would now be dealing with either some form of modified .wav file or a completely different file format that could be modified and/or responded to by other algorithms such as ones for mic distance. Purely speculative but if you are dealing with modified "optimized" IRs that have additional data than just the standard IR/wav file does this still have to fit in the same 1024 or 2048 space as a conventional IR. If so do the bytes allotted to optimizations allowing the extra features impact the overall quality or detail of the .wav. Depending on implementation the impact to IR detail could be negligible or non-existent.

• Method 4 - Helix cabs are simply a regular IR created with a 'neutral' mic and then "colored" appropriately by the Helix's mic type, mic distance, and early reflection algorithms/parameters.

Implications: This, similar to Method 2, would seem like a highly likely approach to me although Line6's comments about Helix cabs being optimized IRs does not seem to jibe with this approach which simply uses a standard neutrally oriented IR and applies post processing for the various cab parameters (mic-type, distance, reflections).

 

 

So... in summary, I have no clue how Line6 has implemented their cabs but if they are simply "regular neutral" IRs with post processing algorithms applied for mic-type, distance, and reflections that begs the question that perhaps some of these algorithms could be applied to and opened up to third-party IRs. Applying Line6's early reflections to a third-party IR for instance. It might not make a whole lot of sense to apply a Line6 Royer mic emulation to an Ownhammer IR that had been created with an SM57 mic but I suppose it could be done. Perhaps Line6 has pioneered a whole new way of creating IRs that allows them to be "parameterized" and processed in new ways by the IR engine. If so, this technology could provide the entire IR community with a much more efficient, enhanced, and flexible method of sorting through IRs if it was applied as an industry standard.

 

Maybe that is where we are headed while currently different manufacturers (e.g. Line6, Fractal) offer different IR parameters that can be modified. One standard IR protocol would offer many mic, distance, and potentially other parameter choices across different products instead of requiring a different IR for each parameter change. Similar to the way a Helix cab functions. One has to wonder though if "parameterized" IRs are or would be as faithful as one that is created recording the actual mic selected at the distance, axis, etc. indicated. My gut says it would take a long time to get certain parameters like mic-type in "parameterized" IRs sounding as authentic. Perhaps all the actual recorded mic-types could be combined into a single file and then parameterized instead of requiring separate files. Anyway, IRs feel to me like somewhat of a moving target and an evolving technology/art right now.

 

One conclusion I have come to while stepping through some of the possible methods Line6 might have used to create their cabs is that running a Helix cab is not "the same" as running an IR. Either it has an essentially different DNA than a standard IR .wav file or it is a standard although perhaps "neutral" sounding IR(depending on whether or not Line6 is using different IRs for different mics) that is receiving custom Line6 post processing that the Helix does not allow to be used on third-party IRs. Either way this makes Helix cabs a somewhat different animal than a typical third-party IR. Even if the only difference is that the Helix's native cabs are IRs that have access to some post-processing that third-party IRs do not. This conclusion could admittedly be negated by some method that was not listed here and changes the playing field.

Thanks for that

The best post ever about this subject.

 

If we take some ampsim makers in this guessing like Guitar Rig/Amplitube/Softube i have used those since they started some 15 years ago and they are known to have "stock cabs" instead of IR's like Revalver,S-Gear.TH3.

 

The more they added function to their cabs like how you move the mic add two mics blended and even use of off axis etc, the bigger the size of plugin would be now that have also alot to do with the GUI files but i am sure there is alot of space require for the cabs aswell as it is of course with added effects and amps and other features.

 

My guess is that Line6 used method 4..

 

The most stupid thing i've read was when someone said they had many many IR's inbedded in the stock cabs to choose from when we move parameters in the stock cabs?

But hey why didn't they include off axis and dual mics then and moving sideways towards the edge of the cap?

 

If you look at the size of the whole Helix FW its about 3.7 mb today it was about 3.5 for FW 2.2.1 i know that is not the same coding as on the computer but anyway it has everything in those 3.7 mb including stock cabs amps effects etc.

One folder of miced up IR's on my Ownhammer DLX Reverb V30 1x12 ported for Helix 16 bit 48 khz is 2.7 mb and has 187 IR in it.That is all mics end every move including the edge of the cap and fredman micing technique and room.

[HonestOpinion]

Thanks for that

The best post ever about this subject.

 

If we take some ampsim makers in this guessing like Guitar Rig/Amplitube/Softube i have usewd those since they started some 15 years ago and they are known to have "stock cabs" instead of IR's like Revalver,S-Gear.TH3.

 

The more they added function to their cabs like how you move the mic add two mics blended and even use of off axis etc, the bigger the size of plugin would be now that have also alot to do with the GUI files but i am sure there is alot of space require for the cabs aswell as it is of course with added effects and amps and other features.

 

My guess is that Line6 used method 4..

 

The most stupid thing i've read was when someone said they had many many IR's inbedded in the stock cabs to choose from when we move parameters in the stock cabs?

But hey why didn't they include off axis and dual mics then and moving sideways towards the edge of the cap?

 

If you look at the size of the whole Helix FW its about 3.7 mb today it was about 3.5 for FW 2.2.1 i know that is not the same coding as on the computer but anyway it has everything in those 3.7 mb including stock cabs amps effects etc.

One folder of miced up IR's on my Ownhammer DLX Reverb V30 1x12 ported for Helix 16 bit 48 khz is 2.7 mb and has 187 IR in it.That is all mics end every move including the edge of the cap and fredman micing technique and room.

Thx for the kind words and your observations on firmware update file sizes. That pretty much answers any speculation about Method 1. They are definitely not doing it that way, never sounded plausible to me anyway. Not sure exactly how they are doing it, it remains a mystery. From the few clues they have provided it seems like it may be either a modified .wav or other format file that provides additional cues or markers to the cab/IR engine on how to process the parameters or it could be straight up post processing on a (neutral?) IR. Could be some other entirely different approach though. Maybe they will offer us some additional details on how they pull off their custom cabs at some time in the future although I am not holding my breath on that. Any IR experts on the forum who can chime in about how IRs or proprietary cabs gain additional parameter processing?

[kylotan]

An impulse response is just an audio file, measuring a very specific type of audio, so you can apply all your usual audio treatments to it. On a very trivial level, you can approach cabinet modelling as an EQ matching exercise; capture the IR with the mic at 0", then capture it again at 0.5". Use an EQ matching algorithm to measure how the response has changed and generate an EQ curve to approximate it. Do that for your range of distances, with the same number of EQ bands available each time. Then, to smoothly model any mic distance from the cab, you can blend between 2 adjacent sets of EQ settings, apply that to the impulse, and then use that impulse as the cab sim.

 

You don't even necessarily need to store all the impulses, just enough so there's a small enough error between the blended/simulated impulses and the ideal values. You might even find you don't need to store the impulses at all if your EQ curve is sufficiently accurate. (The people using the 'OTB' presets might agree with this.)

 

Generally speaking you would spend a fair amount of time tuning the hyper-parameters here (e.g. how many EQ bands is needed for an accurate model? what filter type works best? how many different mic positions? how short can we truncate the IR and still get a good match?) to find out what works best. But it's certainly not impractical for Line 6 to have captured 30 cabs x 16 mics in maybe 10 different positions each and then leave the computers chewing over how best to blend them.

[amsdenj]

Another possibility is that mic and speaker IRs are composed into the resulting cab model to reduce the number of IRs. Its also possible that part of what Line6 did is to find a way to reduce the DSP load to do the convolution.

[Guitarmaniac64]

An impulse response is just an audio file, measuring a very specific type of audio, so you can apply all your usual audio treatments to it. On a very trivial level, you can approach cabinet modelling as an EQ matching exercise; capture the IR with the mic at 0", then capture it again at 0.5". Use an EQ matching algorithm to measure how the response has changed and generate an EQ curve to approximate it. Do that for your range of distances, with the same number of EQ bands available each time. Then, to smoothly model any mic distance from the cab, you can blend between 2 adjacent sets of EQ settings, apply that to the impulse, and then use that impulse as the cab sim.

 

You don't even necessarily need to store all the impulses, just enough so there's a small enough error between the blended/simulated impulses and the ideal values. You might even find you don't need to store the impulses at all if your EQ curve is sufficiently accurate. (The people using the 'OTB' presets might agree with this.)

 

Generally speaking you would spend a fair amount of time tuning the hyper-parameters here (e.g. how many EQ bands is needed for an accurate model? what filter type works best? how many different mic positions? how short can we truncate the IR and still get a good match?) to find out what works best. But it's certainly not impractical for Line 6 to have captured 30 cabs x 16 mics in maybe 10 different positions each and then leave the computers chewing over how best to blend them.

"An impulse response is just an audio file, measuring a very specific type of audio, so you can apply all your usual audio treatments to it."

 

Hmm not really plz redo your homework about what an IR is

 

OTB cant be reproduce an IR of a miced up Cab with off axis micing and a miced up cab that has the mic on the back of the cab or multiple mics for that mather

Two Notes offer those some of those options on the Wall Of Sounds plugins but to me they dont sound identical to what it sounds in real life at all.

But hey they are close but even there an IR that capture those micing techniques beats W.O.S by far..

[kylotan]

"An impulse response is just an audio file, measuring a very specific type of audio, so you can apply all your usual audio treatments to it."

 

Hmm not really plz redo your homework about what an IR is

 

This is exactly what an impulse response is. It (in theory) is a recording of the audio response to an infinitely short 'impulse' of sound. (There are other ways to generate it, but mathematically the idea is to produce this same output.) Then this audio is convolved with your other audio to produce the output. Convolution is essentially a multiplication operation and is commutative so you could, in theory use the impulse as your performance and your performance as the IR and get exactly the same output! This is why impulses can be represented as simple audio files.

 

An impulse response function - and thus, a file that accurately contains that data - captures the way a linear system responds to incoming audio in both frequency and time, which is why IRs can simulate reverb and echo as well as speaker and microphone response. However, when capturing mic/cab responses, typically there are no room reflections being captured, just the direct signal from the speaker, whatever frequency characteristics the speaker and mic impart, and phase cancellations resulting from position. These are all things that, theoretically, can be modelled with parametric equalisers. If that wasn't obvious enough, microphone manufacturers publish the frequency response of their mics and you can replicate those curves in your EQ.

 

Would I expect to adequately model a cab/mic combination with the Helix's 5-band Parametric EQ block? Hell no. Would I expect to adequately do it with maybe 10 bands, the option of minimum or linear phase EQ, and a computer learning algorithm to do the hard work of performing all the adjustments and measurements? Definitely.