Alignment, design and operation of professional sound systems. Training seminars on dual-channel FFT analyzers and system design. Author of the Subwoofer Array Designer calculator.
At 2:15 how does an "All Pass" filter choose a corner frequency ? It's an ALL pass filter meaning it passes everything, right ? And by the way, what's the point OF an ALL pass filter anyway ?
I'm curious why two 120V legs of 208V 3 phase system, separated by 120º, sum to 208V instead of 120V as this calculator indicates? Anyone know why that is?
Excellent question! Sound pressure is about summation. Whereas three‑phase electric power considers the - difference - between phases, i.e., subtraction. I posted an illustration here: facebook.com/merlijn.vanveen/posts/1028349245335764
Wouldn't delay in amplitude envelope mean that audio gets delayed (if we use real audio for testing)? No one listens to pure sine waves rt? And that's why audiophiles concerned with GD.
I have never seen in my career, in the real world, among the guys who work in the world of concert installations and not those who live among books, someone who aligns Main and sub with the delay finder. perhaps the author has seen them in the world of Car Tuning? if this were to happen, it would mean that decades of smaart and mwyer schools have been of little use, do you think? As Jacop said in his comment, the author confuses Delay Fidner with the possibilities offered by IR if visualized and studied with the right tools. And yes, because here it is simply a matter of software not suitable for doing tuning work with IR, which is possible in other software
This appears to be a faulty delay finder issue, not a "using impulse issue". The impulse clearly showed a 10ms start time from original t=0. The delay finder got this wrong but the user could clearly see (as you did and pointed out) this is wrong. The impulse could clearly be used to align the mains to subs, just don't use the delay finder to do so. Might want to update your title to not be misleading. I'd suggest Reasons to avoid using automatic delay finders to align mains to subs.
This video is awesome! Starting with a delay line, going for a regular filter and ending with a constant phase (= frequency dependent delay) filter is an excellent explanatory journey. For a week now I have been breaking my head over the question if phase actually introduces a delay in the output. e.g. let's take the constant 90° phase shift filter and place 1000 of it in series. Does the 1000x90° correspond to a measurable (or audible) delay at the output? or does the phase wrap around at the 180° mark and there is no delay at all? According to your video it is the latter. I am sure if I keep thinking about this video enough, i can finally wrap my head around it.
A constant phase has a group delay of zero. But group delay is not necessarily the latency that most people think of when they hear the word delay. A time delay gives a linear phase.
Merlijn, how might someone be sure the have the right delay value if perhaps the Subwoofer naturally/unnaturally lags in its performance? If I just use a tape measure or laser to calulate the proper delay offset for Smaart to read the trace how can I be sure that I've accounted for all possible delay inducing issues, without blindly trusting the physical distance? (Are there even such anomalies in a signal path or are you saying that using the phyical distance is a fullproof method?) ... I'm just a beginner so be gracious with me :)
Very interesting that the listening experience of a squarewave is mostly independend of the phase of its frequency components. Is an allpass somehow audible with real music?
Thank you for you comment. APF audibility is a point of contention. Carefully executed listening examples-typically involving headphones-have resulted in certain audibility thresholds (search AES library), but I sincerely question whether or not typical audience members will be able to tell at a conscious level.
It makes sense then that IR is not a good measurement tool for adjusting timing difference between subwoofer and mains for what we're ultimately trying to do, which is to align the phase slopes of mains and sub throughout the crossover region for good summation. Thus, acoustic slopes, assuming they match a good and proper crossover slope, should sum to a nice flat response in the region. I've found that phase "slopes" of each in the region will show if the sub is lagging behind the mains or mains behind the sub by the steepness of the phase wraps. As you delay a driver, the phase response will exhibit more wraps. So it's a matter of using either delay or phase adjustments in the DSP to make sub wraps overlap the mains throughout, and it's done.
Are these numbers only true when the sine waves combined are 1 decibel? If the sine waves are 10 decibels do you get 50, 30, -60 decibels or are two 10 decibel sinewaves combined in pefect phase to become 16 decibels?
I have a question about the height of main: The height that we must indicate must be that of the highest speaker, the one that is lower, or the one in the center of the line array? Thank you very much for this fantastic tool.
Thank you for your inquiry and your kind words. For height I typically, take the middle of the entire array. At low frequencies, it acts effectively the same as a single point source.
Thank you for your inquiry. I hope you are doing well! I kindly ask you to be more specific. You do not hear the group delay, or you would not describe the phenomenon heard, as group delay but something else in stead? If so, what would you call or describe that what you are hearing? Maybe I have done a poor job at underscoring what exactly it is that is being demonstrated here. The all-pass filter - as shown to the right - is applied to track 1. Track 2 is a duplicate of track 1 - without processing - and is subtracted, by means of a polarity reversal, from track 1. What is left over, is the residual difference between tracks 1 and 2. At 100 Hz - where the APF's corner frequency lives - I can hear a sustain which is not audible when the filter is bypassed. Is it possible that you may have watched this video on a laptop computer or computer with small speakers in which case 100 Hz most likely will not be reproduced. If so, please try headphones or larger loudspeaker in stead. Does that answer your question?
Yeah, and I loved loved loved how the SPL section view coincides with the floodlight in the RL-delta section view. The penny on sub alignment dropped in my head right there and then, bazinga and kablammo!
Hi Merlin! In about 13 minuts you talk about L1, L2 and L3 phases of an AC generator. In this case the circuit is a summing circuit or an push pull circuit? Thank you!
Hey Merlin, first of all good explanations (also in many of the other videos an essays!). But one thing that confused me first (as I have a diploma in electrical engineering and pretty used to the concept of dB): seeing a "formula" telling me x= -20dB + 2dB will be in my world -18dB (as we are summing values in logarithmic scale or in other words multiplying factors in the linear scale). So watching the video I understood "what will we get if we sum a -20dB Signal with a +2dB Signal; but then - sorry for insisting in correct details here as an engineer - is totally different question (which you did explain correct and very well to understand btw.). So the "formula" questioned by the original post could be: what is it if I add 2dB (gain f.e.) on to -20dB -> answer would be -18dB (still attenuation so to speak, as we normally use dB without any index for gain/attenuation values) OR the question could be "what do we get if we add -20dB(u) to an +2dB(u) signal, then the right answer will be 2,66dB(u) - then it is crucial t tell, that we are talking on dB not as fraction but as given value (as fraction of a defined L1 in case of "u" f.e. 0.775V). I often meet people who are not safe with the concept of dB (as you did mention in your video), but mess with given values (like dBu, dBV, dBFS) or just factors (fractions as you say) or gain (which is just a factor). So they get into trouble by f.e. setting limiter values because they have to separate gain (typical 30-34dB on todays amps) from Levels given in dBu/dBV (like 0dBu oder +4dBu). I hope I could somehow make clear what my point is.Your explanation of the concept for dB is pretty good, but I still see danger in misunderstanding "adding levels given in dB" or "adding gain given in dB".
Hey Pete. Thanks for this message. Do you have a document, website or anything witch can clarify and explain what you say and be safe with the concept of dB? Thanks!
Hi Merlijn, Do you know where we can find those Meyer Sound webinars in 2020?, They're not in Meyer sound website anymore, do you know why they were all removed?
To have the measurement of the main and the sub simultaneously running and displaying their separate responses (both phase and magnitude), the reference for each must be the actual crossover's band limited output, correct? Unless the two loudspeakers are completely acoustically isolated from each other and separate microphones are in use for measurements made from the same full spectrum reference signal (non crossover). Otherwise, if the main and sub are side by side, or in a typical PA setup, and one measurement mic is used with the same reference signal for both measurements, then there will not be separate, usable traces found by the TF. The only way to have these measurements running simultaneously in a 'venue' shall we say, would be to have the reference signal for main be the main loudpseaker's crossover output, and the sub reference be the sub's crossover output. Though realistically, being able to run the two measurements at the same time is really only more useful for demonstration’s sake and perhaps as a time saver if the measurement setup is already capable of that, it would seem to me.
Hi Merlijn, and thank you for all the work you have done. I have noticed that, in rare occasions, the 60 degrees corridor is not enough. Meaning that, even if you achieve to match the phase slopes at the Xover +10dB to -10dB range, you might have to check the slope characteristics of the TOP, after the 60 degree corridor ends. If, by any case, the TOP phase slope starts to lead progressively, comparing with the phase slope of the sub, then it is possible that at some frequency point, after this range, the 2 elements are going to be out of phase by 180 degrees(worst scenario). If at this point, the main dominates by more than 18 dB then the loss is going to be less than 3 dB. But if it dominates by 12-16 dB, you may have a little 3-5 dB dip(hi Q) like a mini notch, somewhere after the Xover. I repeat, this is something that happens to me rarely, especially when processing has altered the phase response characteristics of the loudspeaker or the elements come from different manufacturers and the phase traces behave differently. Of course, all this depends on the time you have and maybe as you said "There are bigger fish to catch" than this but, nevertheless, I am curious if this had happened to you in the past and if you 'd like to share ways to deal with this. Thank you again.
