IR experiments - Wetness, distant feel problem

[Cuki79]

Originally Posted by martingitdave

I think the distant sound may be a result of the room where it's recorded? The mic is still picking up reverberations from the room, no?

43 ms at 340m/s (velocity of sound) is 14.6meters. If you are at 3m from the wall, it means you get an echo every 17 ms (6m propagation). So 43ms account for maximum 2 re-bounds.

In a reverb, everything from 10 to 50ms is called early reflection.
https://valhalladsp.com/2011/05/18/v...arly-controls/



Everything beyond 10ms (480pts) goes into early reflections territory.

So my answer to Dave is yes probably but to what extend I don't know.

All I know is my new recipe sounds more direct and enables me to use shorter IRs while still providing a certain tone.

From sonible.com

First reflection
Every audio reflection that reaches our eardrums within 10ms after the direct signal is combined with the original, direct sound. First reflections can amplify the signal and add color to it, but they can also produce undesirable comb filter effects. As sound travels at a speed of approx. 343 m/s, it has therefore travelled around 3,5 meters after 10ms.

Early reflections
Audio reflections that reach our ears 10 – 50 ms after the direct signal provide us with a lot of information about the nature of a room – its size, its surfaces and indications of its shape. The temporal difference between early reflections and diffuse reverberation also contains important information about the listener’s distance from the sound source.

Below are pictures of an IR from my old process (left: response in frequency, right: impulse response)



One can see most of the IR happens before 10ms. We are re-creating the propagation and reflexion of sound in the top of the guitar with acoustic waves travelling at 3000-6000m/s (thumb guess: I did not compute the shear velocity).

[martingitdave]

Originally Posted by Cuki79

43 ms at 340m/s (velocity of sound) is 14.6meters. If you are at 3m from the wall, it means you get an echo every 17 ms (6m propagation). So 43ms account for maximum 2 re-bounds.

In a reverb, everything from 10 to 50ms is called early reflection.
https://valhalladsp.com/2011/05/18/v...arly-controls/



Everything beyond 10ms (480pts) goes into early reflections territory.

So my answer to Dave is yes probably but to what extend I don't know.

All I know is my new recipe sounds more direct and enables me to use shorter IRs while still providing a certain tone.

From sonible.com


Below are pictures of an IR from my old process (left: response in frequency, right: impulse response)



One can see most of the IR happens before 10ms. We are re-creating the propagation and reflexion of sound in the top of the guitar with acoustic waves travelling at 3000-6000m/s (thumb guess: I did not compute the shear velocity).

Fascinating explanation. That’s for taking the time to provide that!!

[Cuki79]

In this video, I show few presets I made with custom made IRs
1) Reference: External mic tone
2) 100% Piezo (Martin goldplus / Fishman Matrix)
3) 2048 pts IR with blend and parametric EQ to tame the 110Hz boom
4) 2048 pts IR in crossover configuration. The IR taking the role of the internal mic of a virtual "LR Baggs Anthem" pickup.
5) Totally different approach: EQ is heavily used (9 Parametrics + 1 Shelf filters) to shape the tone. IR is used just to tune phase shift and adjust the "feel & response". The idea is to get a clean & dry string tone more appropriate for LIVE use.




Please use GOOD HEADPHONES or monitoring speakers.


All the tones (except Reference mic) come from the HX Stomp XL and the piezo pickup.

The external mic is turned off everytime you hear the HX stomp XL presets.

0:27 Reference acoustic tone (external mic)
1:22 Piezo Pickup
2:39 HX stomp: IR + Blend (custom 2048pts IR)
4:18 HX stomp: IR + Crossover (custom 2048pts IR)
6:07 HX stomp: EQ + IR (custom 1024 pts IR)
6:53 Presets browse

[SpruceTop]

I tend to like best the HX stomp: IR + Blend (custom 2048pts IR, although the HX stomp: EQ + IR (custom 1024 pts IR) has the potential to sound the best to me except it sounds like a blanket has been draped over the sound. If the HX stomp: EQ + IR (custom 1024 pts IR) could be made to sound more upfront and present it would probably get my vote for the best-sounding of the video's demos.

[Cuki79]

Originally Posted by SpruceTop

If the HX stomp: EQ + IR (custom 1024 pts IR) could be made to sound more upfront and present it would probably get my vote for the best-sounding of the video's demos.

I changed a lot the new algorithm since this video.

I use 2048 pts on the IR and focus on the phase response.

I am very happy what I did "by hand", I am improving the code.

It's a work in progress

[RogerPease]

We await your update.

BTW, I liked the IR + Crossover best, but the IR + Blend was a close second.

_RP

[Cuki79]

Thanks Roger,

Here is a quick demo of what I am working on



In this case, I let the EQ do all the boost and cut in frequencies. I just manually adjust two bands for optimum balance.

I made an IR that only handle the phase. It does not alter the sound. It just delays the lows, medium, and treble so the feel is right.

1) UST goldplus
2) 2048pts IR blend
3) 2048pts IR anthem-like crossover blend
4) EQ+ 100% Phase IR

There is another idea I need to try...

[Cuki79]

Originally Posted by Cuki79

There is another idea I need to try...

Ah Ah it works... But it is so subtle, I think you need to play it to understand.

Here is the idea.

1) Quack is reported everywhere to be corrected by a notch in the 2kHz-4kHz region depending on guitars.

2) I thought about the early XX century 400,00Hz^2 rule.

From John Strohbeen on https://ohmspeaker.com

Rule of 400,000
Now comes in the “Rule of 400,000” (that I heard Western Electric developed in the 1930’s or earlier). It basically stipulates that to have a natural sounding speaker, the frequency of the low roll off is multiplied by the high frequency roll off, the results should about 400,000. If you have a low frequency response down to 80Hz, you only want treble response to 5,000Hz – neither higher nor lower. This covers the demands for voice reproduction quite satisfactorily. So, if you need only voice reproduction, you can have high efficiency; but if you want full music reproduction from a single-driver system, you will have to have low efficiency.

So I thought what is the "mirror" frequency of 2kHz and 4kHz... indeed they are 200Hz and 100Hz... Exactly where the Main mode resonance is!!!

Since I've been carefully correcting this zone with IIR (Parametric EQ), I might need to focus on the mirror area too. For those high frequencies FIR (IR file) is a better tool.

So instead of making a pure phase IR, I tuned the IR file so
* I correct the phase for EVERY frequencies
* I correct phase & amplitude in this range.

And guess what?

The two frequency boost between 100 Hz and 200Hz actually have a miror cuts in the 2kHz-4KHz area...

Did it make a difference.

I think it did. The attack is more natural and the tone less quacky.

However I am not sure you can hear it on a mp3...

Anyway I am happy! Now I need to try on other guitars.

[Cuki79]

Aaron Short posted a comment on my last Youtube video:


So I used the new process ideas I was working on (Phase & 400,000^2 rule) to try to remove the quack without changing the pickup tone.

Here is the result (Please use headphones):


1) Dry pickup (Martin Goldplus/Fishman Matrix)
2) 1024 IR file designed to tame the quack with minimal tone change.
3) The EQ+IR process last version (Work in progress)

First ingerstyle, then strumming.

Aaron Short is playing a Custom Shop Martin based on the OM28JM

[dreamingOM]

Hi, I apologize for the intrusion of this post to this experiment.

I just recently came across Cuki79's IRs from the internet.. and I do have a CabM+ that I use with my electric guitar rig.

I recently did a gig with my acoustic equipped with a not-so-desirable Fishman sonitone.

I have no understanding of IR's other than using them as my cab sims to my electric guitar rig.. If I'm understanding this correctly, a shorter 42ms IR would lead to little to no quack tone? as it's my main problem that I couldnt solve with just eq'ing with my CabM+.

currently I load everything at 200ms.. so should I load everything at shortest 42ms?

although it sounded "better" than having my acoustic go straight to the PA. I'm planning to upgrade my pickups, but I wanted to see if I can make it better using my existing gear so I can hold off spending money!

[Cuki79]

I have no understanding of IR's other than using them as my cab sims to my electric guitar rig.. If I'm understanding this correctly, a shorter 42ms IR would lead to little to no quack tone?

Not really.

The shorter the IR, the less definition in the low end.

The shorter the IR the closest to the pickup tone.

The shorter the IR the more "direct" the tone. the longer the IR, the more "wet" and "reverby" the IR feels.

On my website, almost every IRs are 2048pts (42ms), adding them in 200ms in your CABm+ makes no difference with 42ms. Most acoustic guitar IRs have about no content after 42ms.

If you want to try to improve your tone with IR:
1) Make a 2-track recording with a 2-channel audio interface. On the first track, record an external mic in front of the 12th frets about 10" from the fingerboard. On the second track, record the guitar pickup DIRECT (no pedal).
2) Export the 2 tracks either as 2 separate files (with proper naming) or as a stereo file (track 1, hard pan left, track2 hard pan right).
3) put them on a google /dropbox drive and send me a private message with the link.

I'll make you some IR files for the CAB M+


If you prefer to do it yourself, and are comfortable with computer you can use my free open source agorithm (or Jon Field's):
http://acousticir.free.fr/spip.php?rubrique15

Note that those don't correspond to my latest IR experiments.

Ask yourself the question:
1) Do I want a very mic-like tone (my guitar only louder)? The IR made with that in mind will sound good on headphones but you might struggle through a PA speaker/ amp as soon as you want to raise the volume. (Like a mic)

2) Do I just want to get rid of the quack and have a more natural attack with a piezo tone? The IR made with that in minde will provide a "forward", "aggressive", "larger than life" tone but just without the weird piezo attack.

3) Do I want something in between?

My last experiments are focused on points 2 & 3 without using the "blending" solution, that might not be satisfactory for eveyone.

[dreamingOM]

Originally Posted by Cuki79

Not really.

The shorter the IR, the less definition in the low end.

The shorter the IR the closest to the pickup tone.

The shorter the IR the more "direct" the tone. the longer the IR, the more "wet" and "reverby" the IR feels.

On my website, almost every IRs are 2048pts (42ms), adding them in 200ms in your CABm+ makes no difference with 42ms. Most acoustic guitar IRs have about no content after 42ms.

If you want to try to improve your tone with IR:
1) Make a 2-track recording with a 2-channel audio interface. On the first track, record an external mic in front of the 12th frets about 10" from the fingerboard. On the second track, record the guitar pickup DIRECT (no pedal).
2) Export the 2 tracks either as 2 separate files (with proper naming) or as a stereo file (track 1, hard pan left, track2 hard pan right).
3) put them on a google /dropbox drive and send me a private message with the link.

I'll make you some IR files for the CAB M+


If you prefer to do it yourself, and are comfortable with computer you can use my free open source agorithm (or Jon Field's):
http://acousticir.free.fr/spip.php?rubrique15

Note that those don't correspond to my latest IR experiments.

Ask yourself the question:
1) Do I want a very mic-like tone (my guitar only louder)? The IR made with that in mind will sound good on headphones but you might struggle through a PA speaker/ amp as soon as you want to raise the volume. (Like a mic)

2) Do I just want to get rid of the quack and have a more natural attack with a piezo tone? The IR made with that in minde will provide a "forward", "aggressive", "larger than life" tone but just without the weird piezo attack.

3) Do I want something in between?

My last experiments are focused on points 2 & 3 without using the "blending" solution, that might not be satisfactory for eveyone.

Thank you for detailed explanations and instructions! I appreciate it!

1. I want to try your opensource algo. I dont have a mic yet so i have to buy one first in order for me to amplify the tone of my guitar and do a combo.
2. I can record direct as I have an interface and all. Although I just use Garageband as it’s the easiest.

I’m using an affordable J-45 copy. A Sigma JM-SG45 (Laminate B&S) which sounded the closest to the Gibson J45 i tried in the store and compared it against. Unfortunately the pickup is crap. So I used your Santa Cruz Vintage Southerner, J-45 Standard and 1943 J-45 IR’s and made a few presets out of those. I did have to enable the EQ, Enhancer and Reverb in the CabM+ as I dont have any acoustic preamps to shape the tone a bit.

All three sounded very good through the PA and through my Audio Technica M50x headphone monitors. However, theres still a hint of quack that i couldnt dial out. I did try scooping the mids a little however I lost some presence as I’m not cutting through the mix enough so i ended raising it again. Do you mind sharing your EQ settings on the CabM? I saw you using one in a video. Since acoustic amplification is all new to me, I dont know which frequencies are more desirable to boost/cut.

I like a fairly big low end. I like the thump of the bass strings similar to what you’ll hear in this strum test video i did. Strung with 13’s.



I’ll record some samples of my acoustic presets using your IR’s but I’ll just private message you so as not to ruin this thread more. But let me know what I’m doing wrong.

I think with what i want at the end of the day is to get rid of the quack tone, and get a nkcer tone with whatever pickup I use. Be it an Anthem, or K&K. Then have a preset wherein I have a blend of a “mic’ed” capture/tone of my guitar.

So your experiment appeals to me at least with what you are trying to do. But given I have no idea of the ins and outs of acoustic amplification, I guess what I’m waiting on are the final IR files Id need to load in my CabM+ And see how I like it for myself. Am i making any sense?

[Cuki79]

Here is an example of the new processes for a Gibson J-45 std (Baggs element pickup).



1) Baggs element (direct, no FX)
2) "Quackless" preset
3) "Dry" preset
4) "Mic-like" preset

I am really wondering about selling "Line 6 Helix/HX presets pack" on my website. What do you think?

Of course I won't shut down the free database but this is different. You can not get tone 2 & 3 with the database or open-source algorithms.

[dreamingOM]

Originally Posted by Cuki79

Here is an example of the new processes for a Gibson J-45 std (Baggs element pickup).



1) Baggs element (direct, no FX)
2) "Quackless" preset
3) "Dry" preset
4) "Mic-like" preset

I am really wondering about selling "Line 6 Helix/HX presets pack" on my website. What do you think?

Of course I won't shut down the free database but this is different. You can not get tone 2 & 3 with the database or open-source algorithms.

In my genre, which is church/religious music, the HX/Kemper/Fractal platform is VERY POPULAR.

which means, people who play electric and acoustic will have access to a really good presets for both. Most, if not all, Christian guitarists (Bethel - David Hislop, for example) sell their presets/patches for everything they use, Strymon stuff, Boss stuff, and Line6/HX. This includes Speaker IR’s, amp models, overdrives, delay and reverb patches. Its selling like hot cakes because church musicians who are a part of a church who sing their songs wants the same tone. So they buy it. However my rig is built different and I dont lioe copying tones and I make my own. I like my electric guitar tone. Hehehe

There’s a bit of gap with acoustic IR’s. Probably some would have stumbled upon your website like I have. Your website is one of the first links when I did a google search and i wont be surprised if all of them found and use the IR’s you made. However I dont know if they be as critical as you or everyone here in AGF for their pickup tone. Some may just accept their quack piezo tone, or some wont tweak anymore and accept the IR’s as it is and play. So this experiment and the technicality behind it might be ignored and just use it as it is.

My point is, it will sell and if you explain why in a nicer way why this is a bit harder to do in a plain jane english, they’d still be interested in buying and trying it out for themselves.

My search/journey is different. I tend to get obsessive about my gear and tone. Since Im mainly an electric guitarist my ears and knowledge is mainly concentrated to that, and the acoustic guitar side is lacking, so i want to learn and in doing so, i want to be able to come up with the best sounding rig for live/recording sessions with as less manipulation as possible because I want to sound like me, my gear. Unfortunately my finances cant buy me a nice Gibson J-45 and put the top of the line pickups in it, and a whole preamp/DI rig so i resort to IR’s for now.

If you have anything for the CabM, id buy it too. I bought a couple of IR’s anyway. Mainly dynIR’s which TwoNotes have developed to use for their products.

[Cuki79]

Originally Posted by dreamingOM

If you have anything for the CabM, id buy it too. I bought a couple of IR’s anyway. Mainly dynIR’s which TwoNotes have developed to use for their products.

You can download any IR from my website and try them in your CabM for free.

I'll make you some free CabM+ preset. I own this pedal (It's French so I can actually make it if you send me the recordings.

The EQ+IR will not sound as good but the "Full IR" and "Quackless" presets will be as good as coming from a Helix/HX (might be better).