I won’t repeat myself, but to me you are running in circles. A lot of cheap measuring microphones show a rising top end, +3 to +6dB at 10kHz isn’t unusual, while response often drops again above that. When you tilt the microphone (45 to 60 degrees is enough) the rise often disappears.
On the NTI, that should have a factory correction file, or NTI should still be able to reproduce it. If kept properly, it should be within reasonable specs. I’d send that one off to mentioned parties above.
On the NTI, that should have a factory correction file, or NTI should still be able to reproduce it. If kept properly, it should be within reasonable specs. I’d send that one off to mentioned parties above.
Of course I am running circles!!! That is the only way to get my measurements AND simulations right - by repeating. The purpose is to get better at them each time.
Out of the all mics, only one (dbx) shows very very tilting response - so if I simulate according to it the sound will be very very dark - no highs. Are we talking about the same? Or maybe it is "Harman" curve, which is dark? Maybe Vituixcad default scoring, which is biased to pretty dark sounds, or everything is because of measurements? I have no idea which combination I have. Maybe measurements are bad, or maybe Harman/Vituix curve is too dark for me. My mind says to trust calibrated mic(s), but they together differ a lot between them: 1-2dB, which is a lot. To my mind if you produce measurement device, which can be called "calibrated"... It has to be accurate. Calibration procedure is dead simple if you are microphone producing company - put mic on stand, run sweep, put in a box, press couple buttons on computer, where it generates comparison file to the most expensive and trusted mic of Earthworks, NTI or whatever hi-fi company which itself is calibrated by more complex means. It can be automated enough to make it in 10 minutes including mechanical work. Even 5 min. Even you could do it in batches by placing lots of mics at once. My head does not understand how "calibrated" can be so uncalibrated. Yes, I would like to trust Sonarworks, as it is Latvian company, North Europe has high quality standards. My current idea is to look for ASR reviews for some relatively popular speaker, go into hifi store, listen to it, or ask to let me measure it, then compare to my sims. Or simply listen some well known to me record.
I haven't heard about 45 or 60 degree tilts. I will try in a day or two, but isn't it the similar effect as tilting speaker?
To make a very long story short : the rising response of the microphone is a diffraction caused by the edge of the the rod of the microphone. it starts above 2kHz. This can be calculated and corrected. This rising response disappears if the wave is perpendicular the microphone.
See this documentation : https://www.grasacoustics.com/filea...ment_microphone_affecting_the_sound_field.pdf
It is a laboratory microphone with extremely stable characteristics : low noise, low distortion, low temperature deviation etc. They are very expensive. Sold by B&K, GRAS etc.
An other information The measurement microphones have precision ranges : Class 1, Class 2.
If you want a high precision you should go to a high level precision like Earthworks, NTI etc.
But I think if you want just measure a sound level of speaker, a Sonarworks is enough. But for distortion measurements it is an other story.
Understood. So if Sonarworks mic has calibration file for 0 degrees and for 90 degrees, I should use it on 90 degrees and trust 90 degree calibration file or should use it leaned 45-60 degree and use some kind of approximation or mix of those 0 and 90 degree cal files?
Also... If mic is calibrated - shouldn't already this effect included in mic cal file???
Nope. I want to measure just levels. As accurate as possible. Up to 15000Hz
There are a lot of open data on the internet available about this driver distortion and it is fairly easy to predict after crossover and build crossover according to it.
I don't understand this, sorry.
MEASURED response is for example 85dB @ 10kHz, when the REAL is 82dB @ 10kHz, so I have rising response by 3dB, right???
I have 3 cal files for Sonarworks - 0, 30 AND 90 deg. 0 and 30 seem to be the same.
Real example:
On 0 deg measurement I have 73.5 dB on 10kHz, after calibration correction - 72.2 dB. And when I simulate FLAT with corrected after calibration measurement - I have too dark sound. If I simulate without correction - I will get even darker sound. PKAudio with the same crossover is flat, when I get raised highs in the simulation.
On 0 deg measurement with 90 degree calibration file I have 73.5 dB on 10kHz, after calibration correction - 75.74 dB. I haven't measured with 90 deg mic AND 90 deg file. But in this case if I try to simulate flat - I will get no highs at all.
Last edited:
Yes
I don't understand why you have three calibration file. There are no diffraction when the wave has 90 deg. of incidence, you should not have a 90 deg. correction. Perhaps my assertion is false...
With my microphone I have one calibration file for 0 deg. incidence and my measurement and my simulation are the same. And one case I don't a good balance is on an open baffle, I have to add treble, otherwise no problem with other speakers.
Note we pollute the subject with microphones instead with the SICA coax. But again measurements never say how a speaker sounds, just you have something with a response curve you can see.
I think your taste is for fresh sound. No one stops you from tuning your system that way. The majority might prefer something else, but that is of no concern.
I also think your loudspeaker system and room interface might mess up your listening experience. We design loudspeakers to be flat to have a starting point. Room influence is quite important for perception.
I think you will find a better explanation and some comparison examples at this website
https://www.jochenschulz.me/en/blog...7150-vs-behringer-emc8000-vs-beyerdynamic-mm1
https://www.jochenschulz.me/en/blog/which-angle-is-correct-for-the-measuring-microphone
The Sonarworks is the more reliable of the calibration files, Dayton had some problems with their calibration files.
For measuring drivers for designing speakers the SW mic should be used at 0 degrees.
Why not compare with the driver you already own to PKAudio or Vineeth's measurement of the tweeter driver?
The differences in common measurement mics are almost all above 2K, below that there is often very little change or need for calibration.
That should tell you if there is something very different going on with your measurements. You want the mic to tell you what you have, you can then decide if you like it or not and adjust from there.
Yes, it may be, but it is not the case this time. I compare human voice. It is very very easy to understand when it sounds natural. All the eq or non-neutral system gives artefacts to the voice and you start to recognize. You don't need golden ears for that. Low and highest freq are hardest (for me) as I dont know exactly how sythesizer or some instruments have to sound. Most of the best singers have really the best mastered records: I listen to Daft Punk, Adele. And yes, now my primary listening device is Sica 5,5 coax, as this is my first proper coax system, and I love it so much (another was some old Philips system, MX3 I think?)
Would like to agree to this one. But Dayton has high freq curve which is dropping from ~2kHz to 15kHz around 1,5-2dB what is more realistic sound. Yes I use 0 degrees.
I did. #133 post.
It sounds WAY too hot on high end. So my taste is not for "fresh sound". Ok, to be correct it is not "hot" but lots of sibilance effects and the drum snares are not correct.
Last edited:
I don't see any comparison, certainly not in any way I can understand what and where the difference may be. I meant load one of Petr's measurements of the raw driver and overlay your own of the same angle and distance and scale to see where any differences are. Then you can see if your measurements are comparable.
Make sense if all Sica drivers and différents batchs are measured identical ... Consistency is a real challenge
There is no requirement for the drivers to measure identically, even if the drivers measured within 0.5dB themselves, a different baffle size and shape will make that irrelevant. The point is to see if there is broad agreement or an obvious tilt to the frequency response from the microphone response and calibration.
Cannot write it better! Those 2 dips and 2 spikes in response from 5 to ~15kHz are not helping. For me - PKAudio suggested crossover sounded unnatural in ~10kHz region, I think there all the vocal sibilances, lip movements are heard (too much), Maybe this is the task of "hi-fi" gear, but I seriously doubt that.
I just got a dedicated 4 channel amp with very flexible DSP settings. After blocking one of the ports and with some careful heavy eq'ing in the DSP I really enjoy these little cabs. Very deep bass for a 5.5". Again this is with heavy DSP eq'ing. In the original simulation I was clearly not getting what I was hoping for with both vents, glad I had the possibility to block one.
One question: does this electrical impedance chart indicate a slightly under-sized cab? I was under the impression that if both peaks reach about the same amplitude the cabinet size is balanced to the driver. Might explain why I need rather heavy eq to get the bass I was hoping for out of the speakers.
The cab is:
8.5 internal liters total (incl port)
0.4 liter round port, 26 cm deep 4.5 cm diameter.
Amp is https://www.thomann.de/se/the_t.racks_dsp_4x4_mini_amp.htm
Attachments
Last edited:
I don't think so, it just indicates a low tuning. that makes sense for a downward sloped bass rolloff.
Cannot answer other questions, only can add to them, as want to see full your config, but... For me ~3,6cm diam port is adequate. Maybe 4cm is the upper limit, as there is not much turbulence when used at the nearfield levels. I use smth like 55-65Hz tuning, still haven't firmly decided which sounds best, yours is... ~43Hz, I think this is overkill for this driver. But have no idea about your DSP settings
EDIT: oops, with port volume included probably ~45Hz?
