Finding peak of two measurements ??

Discussion in 'Official REW (Room EQ Wizard) Support Forum' started by mamba76, Jul 19, 2018.

  1. ddude003

    ddude003 Active Member

    Joined:
    Aug 13, 2017
    Messages:
    118
    Likes Received:
    14
    Location:
    Somewhere Northeast of Kansas City Missouri
    At one point in time the orthodox research community believed the earth was flat and there be dragons that lie beyond... The sun revolved around the earth and folks were put to death for imagining otherwise... And a car is not a room... Placement of drivers and the materials used in the construction of the car, metals, plastics, composites, especially its interior, and sound proofing and size make its domain unique and still open to discovery and invention... For instance, my friends daughter has a car that has the speakers in an acoustically manufactured roof... My own 1994 RX7 has a custom manufactured and very unorthodox, for its time, base driver... No cheesy Hall sounds here...

    "All our dreams can come true, if we have the courage to pursue them." -Walt Disney-
     
  2. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,211
    Likes Received:
    186
    I guess I am not sure of the point here, but if I have offended you or Mamba76, or even sound like I'm trying to stand in the way of your science, I do apologize. I'm really trying to be helpful, but without a better understanding of what is going on, I can't be of much help.

    I'll give you both the same advice I was given, which came off as harsh, but ultimately is true. Good science and novel innovation need to be peer-reviewed. If you have an idea that is truly novel and innovative, then publish it in a professional journal like JAES, and let the experts review it. This is true in every field of science.

    It is in fact not true that science ever believed the earth was flat, and that has long been disputed. However, it is true that our understanding of physics was fundamentally flawed at multiple times and it took innovative thinkers to document their beliefs, provide experimentation of their theory, and allow for the peer community to rally around these ideas. Science today is not like it was 200 years ago or more, so people don't tend to follow the non-sensical flat-earth like beliefs that once prevailed amongst the uneducated and non-science community. However, it wasn't that long ago that we believed in ether and it took Einstein to fundamentally change how we look at large body objects. He was a physicist with a Ph.D. and he fully documented his theory mathematically in a manner wholly consistent with reality and in a manner that was readily testable (even if it took time to test). Acoustical physics and human physiology are mature sciences, but that doesn't mean there isn't more to learn. However, if we are to innovate, the same holds true. New ideas must be documented, they must be sensible, anything that challenges the current orthodoxy must include experimental proof or the ready possibility of experimental proof to support such a view, and finally, be allowed to follow a traditional peer review process.

    I don't understand what is being done in this thread and have raised some skepticism. That is a casual form of peer review and should be embraced, not rejected. We need to thank people for asking questions or questioning what we are doing. It tells us we haven't done a good enough job explaining it such that our peers can understand.

    Mambo76 never discussed crossfeed, that was your comment, and to that I say, I'm skeptical why that would be of merit, or why EQing the sound of the arrival of the left and right channels differently such that they do not match each other would be a good thing.

    So to use this as an example, I can explain based on my understanding of human hearing and acoustics, why crossfeed makes sense with headphones. In any acoustical environment, we hear based on a mix of direct and reflected sound. Our left and right ears are fed cues that come from both the left and right side of the room and the amplitude and time delay differences help us discern the direct sound from the delayed sound, as does a precedence effect. Headphones don't have that and with headphones, we don't get any room information, our ears are fed a discrete left and right signal with no information being crossfed, as naturally happens in a room. Room reflections also rarely have the same response as the direct signal and that impacts our perception of the space. So with headphones, crossfeeding with eq has the potential to make a headphone image like speakers, at least to a point (and it's worth noting that this has been proven to be inadequate to truely provide that sensation, with better and more sophisticated approaches taking precedence today). A car is an acoustic environment with reflections, just like a room. The reflections are shorter and better damped because the car has more absorption per square inch of surface area than a room does, but a car also has other issues like very obstructed and unequal sound paths. I can't see why the theoretical benefit of crossfeed would come into play in a car.

    With the original post, my skepticism came more from a lack of understanding of what was being done. How would comparison of the EQ being applied to the left and right channels as measured at the left and right ears help? A few concerns come to mind that, in trying to be sure this is effective, are important:
    1. an Omni-mic doesn't hear like we do, if trying to compensate for the differences in what our left and right ear hear, I think a dummy head is needed to ensure that the mic has the same directional pattern as human ears, an HRTF, and an ability to simultaneously measure the impulse response at the left and right ear. This is not possible with a single microphone.
    2. Differences at the left and right ear are expected and necessary for our perception of sound. The inter-aural time difference and amplitude difference is precisely how we perceive the acoustic space, soundstage, and image. Manipulating that to create a more accurate space requires a foundational assessment of what the acoustic space should measure like. I believe that would require dummy head measurements of a good acoustic space with manipulation to recreate that through DSP in the car. This is what Harman, Bose, etc. are doing in their own system development. I don't believe that can be achieved through simple EQ, rather it requires the use of discrete speakers and up-converting to surround to do that.
    3. Measuring this is far harder than it seems and I say that as a DIYer who has done so recently. I have an EARS measurement device and I used it to take binaural measurements. I compared it with the measurements I took using an omni-mic at the same left and right ear position. There is little match and no ability to convert one to the other. Further, because both are USB mics, there is no way to correctly lock the timing between the left and right measurements and as such, no easy way to assess the time delay between the left and right ear (something that is necessary for this analysis to be possible). I did eventually develop an approach that I think is working, but is harder to verify. I use the acoustic timing beep in the original recording to manipulate the channels such that the timing tone is aligned and that initial time difference is fed back after the timing tone but before the impulse such that the delayed ear has a more delayed impulse. This seems to work, but without validation, is questionable. While John suggested what I'm doing is working, an acoustical physicist I talked with is skeptical and feels I need to use a loopback to be sure (which is impossible to do with REW and USB microphones). Even doing binaural impulse measurements is impossible with REW. I'm doing it using Audacity, editing the files, and then using REW just to convert the measurement into an impulse response and analyze the data.
    4. Eqing the left and right channel differently needs to be done carefully if using minimum phase EQ. Different EQ can create a different phase response. In a room this is bad and should never be done. In a room there is symmetry. In a car, this is still a problem, but it isn't symmetric. This makes me feel that you have to apply some different EQ, that makes sense, but that you also need to keep an eye on the phase response of the system at the listening position. The problem is that the phase response of a system like this is usually garbage, it is just a lot of noise caused by the non-minimum phase behavior. It's really hard to ensure the speakers have good phase tracking like this (and that they track across drivers and channels). Before all of the above, I would want to crack this nut first.
      1. If I can find some time, maybe I can take some measurements in one of my cars of the left and right channel using both an omni-mic and my EARS to see what I get. To see if it is possible to get a clean phase measurement that could be used in the EQ process.
     
  3. EvenChu

    EvenChu New Member

    Joined:
    Jul 25, 2018
    Messages:
    1
    Likes Received:
    0
    Hi...i am a new user here. As per my knowledge you can never make a speaker in a car door perform like a speaker in an enclosure in a room 12 feet in front of you. That is impossible. What you can do is adjust a set of speakers placed in disparate locations so that their phase tracking and time alignment is correct, potentially even ensure they are linear phase.
     
  4. ddude003

    ddude003 Active Member

    Joined:
    Aug 13, 2017
    Messages:
    118
    Likes Received:
    14
    Location:
    Somewhere Northeast of Kansas City Missouri
    No worries Matt... I can't and don't speak for Mamba and I am not offended... As for your advice, I will take it with a grain of salt being in the R&D community for some 40 years now... My focus has been on solving customer problems not publishing papers... And for some special customers you are never given more information than you need to know to solve a particular step in their problem chain... We all stand on the shoulders of giants... Choose your giants wisely.

    Cordially,
     
  5. Matthew J Poes

    Matthew J Poes Staff Writer
    Staff Member

    Joined:
    Oct 18, 2017
    Messages:
    1,211
    Likes Received:
    186
    This is fine. The advice is just to point out the need for scientific advancements to follow a certain process. Another reason beyond just peer review is that if it is a fundamental evolution or revolution in science, this is the only good way to spread the message.

    Most of us do lots of science this ‘tis never published. I am a researcher by trade yet most of my work is in government reports and white papers. I have very few journal publications. That also means my work has had very little impact on the field.
     
  6. Jean Ibarz

    Jean Ibarz New Member

    Joined:
    Jun 23, 2017
    Messages:
    2
    Likes Received:
    0
    Hi,

    I think it is clear that some people don't believe in your experiments for some, possibly legitimate, reasons. I have no opinion about that, but I can help you to realize what you want to do in another way.

    If I had to do what you want to do (calculate the minimal amplitude of your frequency response, I will ignore the phase response for now...) I would :

    • export the measurements to frequency response files (text files)
    • open them using a software like Matlab or GNU Octave
    • calculate the min of your magnitude responses like min(signal1,signal2)
    • save your results to a file again (use the ascii argument...)
    • import them again in REW
    You can do more complex things (export impulse as wav, open the wav file, fft then calculates the min of your two signals, then inverse fft and export again as wav files, or any variants we can imagine...). It will be a little bit time consuming, but you can make scripts to simplify the process : at least it should do what you want. If you have troubles doing that, I can help you if you try the explained process using GNU Octave.

    Cheers :T
     

Share This Page