"House sound" Do the designers here have a "house sound"?

[hongrn]

Re: "House sound" Do the designers here have a "house sound"?

Let us know when this design is discovered.

If you set this device to the correct setting, your speakers will first glow, then disappear. No more house sound, just music.

[mattsk8]

Re: "House sound" Do the designers here have a "house sound"?

So do you all voice the speakers you design in highly treated rooms? Mildly? No treatment? This is easily my favorite thread!

My highly treated room for voicing is... (wait for it)... my garage :D. It's about 24x26 with 8' ceilings, and my wall treatments are a bunch of retired skateboards screwed to the wall, a large tool cabinet, and a beer fridge on the back wall; a freezer, lots of hardwood on a hardwood rack, and bunch of tools on one of the sidewalls, and a bunch of cabinets, a desk, and my ply on the other wall. Concrete floor. So it's pretty lively for voicing. If my speakers seem a tad dark, now you know why ;)

[carlspeak]

Re: "House sound" Do the designers here have a "house sound"?

My highly treated room for voicing is... (wait for it)... my garage :D. It's about 24x26 with 8' ceilings, and my wall treatments are a bunch of retired skateboards screwed to the wall, a large tool cabinet, and a beer fridge on the back wall; a freezer, lots of hardwood on a hardwood rack, and bunch of tools on one of the sidewalls, and a bunch of cabinets, a desk, and my ply on the other wall. Concrete floor. So it's pretty lively for voicing. If my speakers seem a tad dark, now you know why ;)

From your description, I would call that room anti-anechoic! :p

[mattsk8]

Re: "House sound" Do the designers here have a "house sound"?

From your description, I would call that room anti-anechoic! :p

Good description lol. I've gotten pretty good at knowing what to let go and what gets to stay when I'm looking at measurements . Then I'll bring them in the house and think, "I don't understand, these sounded amazing in the showroom"... :rolleyes:

Edit: I will say this... as harsh as you'd think that transfer would be, going from the garage to the house generally only consists of a maybe a tweak in a resistor somewhere in the xover. I do get a lot of reverb out there when I'm voicing, but if you do it enough you generally know what to tone out. You do have me curious though, I think I'll measure a pair out there, then bring them in my living room and measure there as well. Although, my living room is hardwood floors (with a rug) so I don't anticipate much change.

[arlis[email protected]]

Re: "House sound" Do the designers here have a "house sound"?

The answer is....maybe. I don't target a certain slope or crossover point. But let the drivers tell me where they want to meet. That might mean 2nd or 3rd order electrical, it might mean a notch. It really is driver dependent.

This is my method as well.

[DDF]

Re: "House sound" Do the designers here have a "house sound"?

There's another factor that causes us to have to "voice" our designs: measurement inaccuracy

It is extremely difficult to perform complete and accurate driver measurements that can then be used to achieve an accurate crossover simulation.

Measuring indoors requires impulse response FFT gating that leads to smoothing in the frequency domain that may easily hide small problems than need to be "voiced" around, such as a peak being larger than actually shown on the graph, or hole deeper, or worse, BDC that is actually a bit different than shown. It also limits the lowest frequency that can be measured to usually ~ 250 Hz, which is worked around using a near field model (or ground plane measurement) + diffraction model spliced together and then spliced to the measure. So you don't usually get a real measurement below 250 Hz and so may miss capturing any anomalies that may exist there, that would then need to be voiced around.

I spent 20 years refining my indoor measurement technique, trying Sonex and other absorbers, refining how the speaker is mounted ( I have results showing big differences caused by putting the speaker on a an open weave milk crate) to avoid even minor reflections or vibrations, changing how the mic is mounted and cable dressed etc; using very high quality mics. I think I have it as good as I can possibly get it. But you know what? Measuring outside with a tall ladder and a 10 to 15 ms window (allowing excellent frequency resolution) still provides results easily 1 dB different in some wide frequency bands, and as mentioned earlier, 1 dB difference over a wide band is pretty very audible.

Said another way, our graphs may be showing one thing (even when based on extremely well done indoor measures), but true anechoic measures would usually show something different enough that the frequency response you think you have in your simulation is actually not quite what is really happening. So you voice to fix it.

This error can be reduced by measuring on a really tall ladder with a really tall mic boom, on a quiet day with really understanding neighbours. Trying to get these all to line up at the same time can be a wee challenging. And how many of us send our mics back for calibration yearly? Can we really trust them to within 1 dB year in and out? If the mic frequency response is not perfectly calibrated, even broad 1 dB errors would need to be voiced around to fix.

I attached an example of one of my own systems, on axis, showing measurement indoor (below 250 Hz was near field+diffraction model) vs outdoor on a tall ladder (good down to 100Hz). I haven't had the time to investigate where the differences are coming from but both were carefully taken and illustrate how easy it is to get led down the wrong garden path in simulation, with reality being something a bit different

[mattsk8]

Re: "House sound" Do the designers here have a "house sound"?

There's another factor that causes us to have to "voice" our designs: measurement inaccuracy

It is extremely difficult to perform complete and accurate driver measurements that can then be used to achieve an accurate crossover.

So, you're basically saying that companies like TAD, Harman, B&W, Sonus Faber, etc... these guys all just plug measurements into a PC, then after they use that to design a xover they fire up the assembly lines? No tweaks in the middle?

I understand the relevance of good measurements, but I don't see poor measurements as a reason for voicing. Maybe a reason for a bad design, but not voicing.

[isaeagle4031]

Re: "House sound" Do the designers here have a "house sound"?

I would have to argue about not being able to get accurate driver measurements. With the programs available to us, along with the hardware, we can have extremely accurate data. Jeff B has shown how we can sim down into the teens.
Even in my "horrible" room, I have pretty good measurements down to about 250hz in room using just the omnimic. Then with jeffs programs and freq response blender, I have full confidence in the results.

[Sydney]

Re: "House sound" Do the designers here have a "house sound"?

So, you're basically saying that companies like TAD, Harman, B&W, Sonus Faber, etc... these guys all just plug measurements into a PC, then after they use that to design a xover they fire up the assembly lines? No tweaks in the middle?...

Matt: This provides examples of line testing methodology.
https://www.klippel.de/our-products/qc-system.html
https://www.klippel.de/fileadmin/kli...sion_Parts.pdf

[dlr]

Re: "House sound" Do the designers here have a "house sound"?

There's another factor that causes us to have to "voice" our designs: measurement inaccuracy

It is extremely difficult to perform complete and accurate driver measurements that can then be used to achieve an accurate crossover simulation.

I agree with regard to the W/M in a 3-way, but for a M/T for W/T, we can get good enough results with MLS/FFT.

Measuring indoors requires impulse response FFT gating that leads to smoothing in the frequency domain that may easily hide small problems than need to be "voiced" around, such as a peak being larger than actually shown on the graph, or hole deeper, or worse, BDC that is actually a bit different than shown.

It also limits the lowest frequency that can be measured to usually ~ 250 Hz, which is worked around using a near field model (or ground plane measurement) + diffraction model spliced together and then spliced to the measure. So you don't usually get a real measurement below 250 Hz and so may miss capturing any anomalies that may exist there, that would then need to be voiced around.

The are below about 1K is in many situations missing a lot the anomalies, agreed. Lots of smoothing effects below that, maybe even below 2k. But I don't think that affects the crossover Fc area of anything other than a W/M. It can be argued that a narrow trap might be appropriate and not implemented if there is a significant driver anomaly.

...changing how the mic is mounted and cable dressed etc;

Which is why I mounted my mic capsule on a long copper tube, supported by a PVC pipe mounted on a camera tripod with full felt coverage. Even the end of the PVC pipe whose end was 20" or more behind the mic capsule showed up in the impulse response. I put some felt around the front of the end the PVC to eliminate that. I wasn't even the one who found that. Feyz checked it, found a reflection in the impulse at a time point, then mentioned that. I checked the time, found it to perfectly match the distance to the end of the PVC. Added felt, reflection gone.

Measuring outside with a tall ladder and a 10 to 15 ms window (allowing excellent frequency resolution) still provides results easily 1 dB different in some wide frequency bands, and as mentioned earlier, 1 dB difference over a wide band is pretty very audible.

Given my experience with the camera trip (very small reflective area), I suspect that the ladder had more reflections that you may have considered. Wide bands would, though, likely be due to limited impulse length. I wish that I lived in an area where I could easily to outdoor measurements, primarily for the W/M, though. Some of the anomalies that are missing in measurements are due to inherent issues with drivers that many times can't be corrected, such as those below what we need for an Fc in a 2-way. Too often these are issues with surround or diaphragm resonances below Fc, one reason that I like to mod woofers and midwoofers.

So you voice to fix it.

I think this may be only partially true. IME, voicing has more to do with getting the tweeter balanced with the system and in adding my preferred "house sound", that being a more recessed sound that can sometimes be achieved by adding a slight droop in the 1-3K area. There are no series measurements issues with this other than incomplete measurements such as when one only measures and designs on one axis. Even with a full set of measurements, this sort of voicing is not reliant as much on the measurements.

And how many of us send our mics back for calibration yearly?

How much might a typical mic capsule change? I've taken drivers that I measured on my 2m x 2m baffle, remounted them years later and found that the measurement was almost identical, in absolute terms as well. This was with LAUD that normalizes the measurement. With current mics, I would lay odds that a driver will wander more over time than a mic, but that's just an assumption on my part.

I would have to argue about not being able to get accurate driver measurements. With the programs available to us, along with the hardware, we can have extremely accurate data. Jeff B has shown how we can sim down into the teens.

I do agree with DDF on this aspect. Measurements through these techniques are good, but the issues that Dave raised are valid, especially the part about missing details in FFT results at the low end of an impulse response measurement. Near field techniques also have specific limitations.

dlr

[fastbike1]

Re: "House sound" Do the designers here have a "house sound"?

Matt, I don't see how you get that. DDF is clearly talking about home measurements. You know very well (if you've ever read anything from Toole or Olive) that Harman takes multiple measurements in and anechoic chamber. The other names do as well. In essence, he has just said that measurements may not be as accurate as you think. There are always uncertainties to anything we do, and a host of possible error factors in home measurement. Still and all, these measurements are generally "good enough" but still may require some small changes to XOs for all the reasons discussed in this thread.

So, you're basically saying that companies like TAD, Harman, B&W, Sonus Faber, etc... these guys all just plug measurements into a PC, then after they use that to design a xover they fire up the assembly lines? No tweaks in the middle?

I understand the relevance of good measurements, but I don't see poor measurements as a reason for voicing. Maybe a reason for a bad design, but not voicing.

[maynardg]

Re: "House sound" Do the designers here have a "house sound"?

At home, in the living room, it is sometimes even difficult to replicate a measurement result between session. Each time, you move the furniture, place the speaker, set up the mic, quiet the ambient noise to the extent possible. Results are rarely quite the same. So, for me, measurements help me model the XO, provide info in tweaking sessions, and set me up for scores of hours of voicing.

You can chase every little blip and wiggle until you grow old, flogging the XO into submission with ever increasing XO parts count, but it is the voicing the yields the right sound. After all the "ear work", I follow up with a measurement and it nearly always looks pretty good.

After you get the basic (and hopefully fairly simple) XO spec'd using measurements, trust your ears to get the voicing right.

[carlspeak]

Re: "House sound" Do the designers here have a "house sound"?

At home, in the living room, it is sometimes even difficult to replicate a measurement result between session. Each time, you move the furniture, place the speaker, set up the mic, quiet the ambient noise to the extent possible. Results are rarely quite the same. So, for me, measurements help me model the XO, provide info in tweaking sessions, and set me up for scores of hours of voicing.

You can chase every little blip and wiggle until you grow old, flogging the XO into submission with ever increasing XO parts count, but it is the voicing the yields the right sound. After all the "ear work", I follow up with a measurement and it nearly always looks pretty good.

After you get the basic (and hopefully fairly simple) XO spec'd using measurements, trust your ears to get the voicing right.

Another builder who's found 'his voicing'!! :applause:

[craigk]

Re: "House sound" Do the designers here have a "house sound"?

why after reading all this I am amazed that we even listen to music through speakers. we all must be so deaf and incompetent in our measuring techniques that we do not know that we can not design speakers that are worth listening to. you can live in the theoretical world of nothing is good enough, or in reality and we do the best we can with what we have. so to the measurement points being made. guess what, the world functions on measurements. measurements that are not very accurate and will never be exact, and we still do just fine. so the first rule, there is no such thing as the perfect measuring device and there never will, so you will just have to live with that. you can live in theoretical world and pretend speaker design is a land od of perfection or try reality. yes reaching perfection or attempting to is a nice idea and worthy goal, but will never happen. voicing is not about fixing mistakes in measuring, it is about fixing mistakes that machines make because machines can not " hear and feel music." voicing has as much to do with bring that emotional part of the music out as it does with correcting any imperfection. if you like dry and sterile, fatiguing dull sound, by all means don't voice. if you are happy with that good for you, I prefer something a little different, so please let me voice and be happy with imperfection in the speaker and myself.

[dlr]

Re: "House sound" Do the designers here have a "house sound"?

if you are happy with that good for you, I prefer something a little different, so please let me voice and be happy with imperfection.

Nobody told you to do anything one way or the other. Please give us the courtesy to discuss this as we please without making unwarranted charges. I'm sure no one here is looking for perfection. Many of us do want to improve with whatever tools we may have at hand since, as you state, there is no "perfection". I for one, am not content with any status quo.

dlr