IR Properties

Discussion in 'Digital & Modeling Gear' started by Jay Mitchell, May 17, 2019.

  1. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    This is a thread about IRs taken at different distances from the cab and the implications of close-mic'ing vs. alternatives. Below are two charts. Both graphs display IR data taken from the same speaker, using the same mic. One trace in each graph was taken on axis with the mic (a Bruel and Kjaer 4007) at 2 meters, the other with the same mic directly in front of the center of the speaker cone right at the grille. The cab is an open back with a single 12" speaker. The microphone is a pressure-sensing (omni) condenser mic. As such, it has no proximity effect. All differences are due to the mic placement.

    Amplitude vs. Frequency​
    [​IMG]


    Log Amplitude (dB) vs. Time​
    [​IMG]

    Any guesses as to which is which? The color designations are the same for both charts; one color is close-mic'ed, the other was taken at the 2 meter mic distance.
    (Hint: there is a floor reflection clearly visible in one of the traces.)
     
    Last edited: Sep 8, 2019
    Cenzo, pinfloi, rhummellh and 7 others like this.
  2. camstudio

    camstudio Member

    Messages:
    664
    Joined:
    Jun 11, 2012
    Location:
    Scotland
    Green is the close mic.
     
  3. ML Sound Lab

    ML Sound Lab Vendor

    Messages:
    393
    Joined:
    Sep 4, 2018
    Good idea about creating another thread Jay. I'll also mention that even though I'll actively post in this thread and my company logo is my avatar, I'm not here to promote my products and I wish to be able to partake in constructive conversation.
     
  4. ML Sound Lab

    ML Sound Lab Vendor

    Messages:
    393
    Joined:
    Sep 4, 2018
    I'm not entirely sure what we're trying to get at here. I thought the conversation in the other thread was whether or not reflections give you undesired results. It was not whether they exist or not.

    Where to begin? Well first of all I would never recommend placing mics at the center of the speaker cone right at the grille but also not many people will have experience with that mic so it's hard to say. I'll just say what I think I'm reading. Looking at those graphs makes me think I wouldn't like either of those sounds since neither seem to be balanced - ideally a balanced sound would have about the same amount of lows and highs. Purple would probably be my preference since it actually has a low end while green low end essentially starts at 200hz which some people consider "low mids". I don't know this mic but not having low end and having that huge scoop at 1500hz makes me think that's got to be the far mic. The purple line seems to be getting that 7ms reflection that you were talking about right?

    How do they sound?
     
  5. John Mark Painter

    John Mark Painter Silver Supporting Member

    Messages:
    6,684
    Joined:
    Jun 8, 2016
    THat's what I would say too.

    I used to have a pair of those B&K and used them on String sections.
    @ML Sound Lab
    That mic is so FLAT that it is almost boring
     
  6. AlbertA

    AlbertA Member

    Messages:
    49
    Joined:
    Mar 21, 2006
    Green seems like the close mic.
     
  7. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    First things first. This was definitely not an attempt to establish a preference poll. It is the beginning of a lesson. If you don't recognize which IR has early reflections that are not part of the cab's sound, then that's the first thing to learn to look for. (FYI, the cab is a very popular combo amp, so there are quite a few folks out there who like how it sounds.)

    I chose the above two IRs because they serve to dispel some of the myths surrounding mic'ing. There were two clues in my post: the mic has no proximity effect, and I revealed that there are obvious floor reflections in one of the IRs.

    Reveal: the green trace was taken with the mic at two meters. It has the exact amount of bass that you get from an acoustic dipole (aka "open back"). In the far field, radiation from the rear of the cab - which is reverse polarity from what comes out of the front - diffracts around the cab, propagates forward, and does a pretty good job of cancelling low frequencies. It also may be a contributor to the notch ca. 1.5kHz and the elevated high-mid level. Note that very few players would aim this cab directly at their ears, for reasons that should be obvious.

    The purple trace was taken with the mic centered just off the grille. Contrary to contemporary mythology, it does not have more high frequency content than the IR that was taken at a realistic distance but the identical angle (0 degrees). The increased level of LF energy is due to the cab's proximity effect. As I pointed out in the other thread - and the famous Dr. B insisted on misunderstanding - this is a direct result of the inverse square law. This IR has obvious floor reflections. See if you can identify them and, from that information, calculate the approximate elevation of the cab off the floor in the measurement.

    I'll post some other IRs of the same cab later, but, from the responses thus far, these IRs warrant quite a bit of further study.
     
    lowyaw, Leonc, Slapshot1977 and 8 others like this.
  8. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    I don't use either one. I don't aim guitar cabs - at least, cabs that don't have my directivity modifier installed - at my head, and I never use close-mic'ed IRs. I'll post another 2-meter IR of the same cab that sounds the same as the cab at a typical playing angle when played through a neutral monitor.
     
    ML Sound Lab likes this.
  9. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    That's it. It's right at 8ms. I wanted some near field IRs of the cab with as little effect from a floor bounce as possible, so I set the cab about 4 feet off the floor. Although that wasn't enough to get the bounce beyond 20ms, it reduced the magnitude by more than 12dB compared to what it would be with the cab on the floor. I could have hung the cab from the ceiling at ~12 feet off the floor and kicked the floor reflection out of the first 20ms altogether, but that was more of a project than I was prepared to take on at the time.
     
    Last edited: May 21, 2019
    yeky83, ML Sound Lab and Mark Al like this.
  10. AlbertA

    AlbertA Member

    Messages:
    49
    Joined:
    Mar 21, 2006
    Interesting, it was that reflection at 7-8ms that made me think it was the far mic instead of the close mic. Well this just shows how close mic'ing doesn't really solve the room reflection problem :)
     
  11. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    Does the length of the reflection-free zone in your IRs exceed 20ms? If not, you're wasting your time.

    What you apparently aren't aware of is that no material you can apply to a structure like that will absorb well at low frequencies.

    Far more likely is that placing a speaker in a resonant cavity lined with frequency-selective absorption makes it sound like crap. If you'd asked, I would have told you that more than 30 years ago. What were you doing back then, BTW? When you try something and then draw general conclusions from your experiment, it behooves you to actually understand what it is you're trying. Otherwise, your premise can turn out to be false (as it is in this case), thereby invalidating any conclusions you might draw. Did you perform any tests to establish just how effective your "absorption" was?
     
    Last edited: Sep 8, 2019
    FlyingsCool and ML Sound Lab like this.
  12. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    My far field IRs never have reflections that early.
     
  13. ML Sound Lab

    ML Sound Lab Vendor

    Messages:
    393
    Joined:
    Sep 4, 2018
    To some extent I think if you've found something that you really like you should stick to it and not use something else because someone says it's "more correct". That being said there's a lot of "crucial information" all the way up to about 200ms and if you pull the mics back a lot you're going to need even longer IR's. So "what the heck is happening" is that you're not getting a very authentic representation especially of the low end. Also if you slice raw phase information like that weird things will happen. :) Again if it sounds good to you it works.
     
    Tmidiman likes this.
  14. ML Sound Lab

    ML Sound Lab Vendor

    Messages:
    393
    Joined:
    Sep 4, 2018
    Come on Jay this is the third time that you've tried to take a personal stab at me. :D This is age-ism - don't play the age card. I've already said many times that you know this stuff better than anyone else and I think I've shown respect. The topic we're talking about is IR's and I think I have a pretty good card hand if this turns into a card game. I'd just much rather talk about the topic than measure d... reflections. I'd love to keep this talk on topic in hopes of everyone (including myself) learning something useful.

    On topic: Having no low end below 200hz is a much bigger problem than any reflection. I'm still not understanding the problem with "regular IR's". All guitar tones people have ever heard have had reflections. Why should we now get rid of them?
     
  15. Ed DeGenaro

    Ed DeGenaro Supporting Member

    Messages:
    19,977
    Joined:
    Jan 3, 2002
    Location:
    Malibu
    If I had to guess it's acting like a high pass filter.
     
  16. antcarrier

    antcarrier Member

    Messages:
    60
    Joined:
    Aug 28, 2011
    Location:
    Sunshine Coast, Australia
    I would have said that green was the FF IR just looking at the first chart, from my experience capturing FF IRs (I don't use NF IRs myself either). Obviously the second one proves it. What program are you using to generate these charts?
     
  17. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    That would indeed be a problem if it were to happen when you play through an IR. It does not.

    First, all the room boundary effects that are not present in the IR - or in the direct sound from the speaker it came from - are added by the environment you're playing in. Those effects will pretty reliably add to the low-frequency content of your sound.

    Second, the voicing of the amp - combined with its interaction with the speaker impedance - adds low frequency content. If playing through cabs with the low frequency response characteristics in the above IR sounded bad, Leo Fender would have failed at amplifier manufacturing.

    Edit: Have you ever noticed that most vintage Marshall amps are voiced much "thinner" than Fender amps of that era? Ever wonder why? What are the typical cab types those amps are used with, and how do those types differ in low-end response? That's a useful subject to contemplate for awhile. Could even lead to an epiphany.

    It is my position that the IRs I acquire are "regular IRs."

    If you can't play an IR through a transparent monitor and have it sound like the cab it was taken from, IMO that's a problem.

    We aren't, we shouldn't, and we couldn't if we wanted to. As I have pointed out already, there will always be reflections from the room you're playing in. The reflections you don't need are an incomplete set from some other room all coming at you from a single source.
     
    Last edited: May 18, 2019
  18. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    You aren't reducing latency when you shorten an IR, assuming the data you remove is from the end and not the beginning. Beyond that, it's impossible to say what might be happening. It would depend on exactly what is in the IR to begin with. If you like 5-ms IRs you might like the cab sims in some of the earlier modelers, which were very short IRs.
     
    FlyingsCool likes this.
  19. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    I was hoping you'd take it as the good-natured jab it was.

    With all subjects of any depth, novices may think they're discovered or proven something when they haven't. It is often the case - particularly in acoustics - that their initial assumptions/premises are wrong, in which case everything that follows is also wrong. Sometimes, they haven't studied the field sufficiently to recognize something that's either been definitively disproven or is common knowledge.

    When you say you built an absorbent "cave," it is a valid question as to how you quantified the absorptive properties of the material you used. If you simply assumed that the material was ideally and uniformly absorptive, that assumption would definitely be false. If you don't know the absorption vs. frequency characteristics of the exact material you used in the exact thickness you used, then you can draw no valid conclusions from your experiment. By far the most likely result was that you built a resonant cavity with some frequency-dependent absorption (certainly better at absorbing high frequencies than low), so you didn't eliminate reflections, you instead added some with altered spectral content.

    My question regarding your experiment was sincere, valid, and crucially important. What testing did you do to determine the absorptive properties of the material you used? I set up and perform such tests on occasion, so my question is not an unreasonable one.
     
    Last edited: May 18, 2019
    LordByron, adamquek and djd100 like this.
  20. Jay Mitchell

    Jay Mitchell Member

    Messages:
    5,448
    Joined:
    Oct 13, 2007
    Room EQ Wizard.
     
    antcarrier and Lele like this.

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice