No speaker type / technology will please everyone. I would not blame a super sharp microscope for showing every minute, ugly, detail. Perhaps you should view better looking, prepared, samples !!! Proper horns and great quality cd's are a tool. This should make you look at the entire audio supply chain, before the drivers. Some people think there are no differences in the sound of capacitors, amplifiers, preamps, digital sources, etc--after all, digital is just ones and zeros ??? Right !! We know that statement can be VERY complicated. This is a great hobby, and you can jump off this bus when YOU are happy with the sound. Some will stay on board and search for something better.
Joe
Joe
I agree. I find any peaks in the 2000 Hz to 3000 Hz range are quite irritating. Even a few dB bump is noticeable to me, especially in female vocals. I always measure the response and add an appropriate notch filter. Then, I find compression drivers and horns are a treat to listen to.
Earl Geddes have conducted a test here
GedLee LLC
Subjective Testing of Compression Drivers: He concludes that FR is the most important factor for subjectively judging sound quality. He also concludes that THD is not a significant factor in compression driver sound quality. What he has found is that reflection / diffraction (HOM) in the horn has a negative impact on perceived sound quality.
Here is an interesting test (in french, but google translate will help) on 1" CD. The popular B&C DE250 does not fare well in the measurement, but is still regarded as a fine CD. Actually Geddes endorse them 🙂
Grand Comparatif de Compressions 1 pouce - JustDIYIt !
GedLee LLC
Subjective Testing of Compression Drivers: He concludes that FR is the most important factor for subjectively judging sound quality. He also concludes that THD is not a significant factor in compression driver sound quality. What he has found is that reflection / diffraction (HOM) in the horn has a negative impact on perceived sound quality.
Here is an interesting test (in french, but google translate will help) on 1" CD. The popular B&C DE250 does not fare well in the measurement, but is still regarded as a fine CD. Actually Geddes endorse them 🙂
Grand Comparatif de Compressions 1 pouce - JustDIYIt !
FR is certainly a huge factor, but I think what EliGuy and some others are saying is that even if the FR is basically right,* there is still something fatiguing in the sound. That could be narrow spikes or reflections causing trouble. Figuring out what is causing that irritation is important. It's likely more than just overall FR.
*IME horn FR often isn't right.
*IME horn FR often isn't right.
In my experience stored energy, especially in the treble-range can cause listening fatigue even if the frequency response is flat. A CD/horn combo is more likely to have this than a dome tweeter, simply because there is more factors in the design one is likely to not get right.
I haven't heard many of the cd/horn combinations like others here but what I find bothersome is a directivity mismatch due to poor crossover implementation. Some systems lack a "seemless" quality as a whole, the individual drivers stand out and tonal balance suffers. Perhaps some ask too much of a particular cd/horn combination ie. how low can it be pushed.
Yeah, it's an all too common goal to try to get more out of a horn+driver than it can comfortably do. I'm guilty of that. Took me a long time stop doing it. 🙂
Agree puppet and Pano,
just because a CD can go low on a horn, doesn't mean it should.
Going through the trouble of taking polars is the only way i know to get it right,
....well, other than a lot of trial and error via listening....
just because a CD can go low on a horn, doesn't mean it should.
Going through the trouble of taking polars is the only way i know to get it right,
....well, other than a lot of trial and error via listening....
Gotcha, those marketing videos are quite often a bit out of sync with the manual's specs huh ? 😛
I made a mistake too.
I've always been impressed by the DEQX and its comprehensive manual, so i took a little time to read thru it last night.
What i see is a maximum tap count which gives a delay of 24ms, or about 4600 taps @ 96kHz output, in either single-amp mode, or active multi-amp mode.
In single amp mode, 3ms is reserved for the main speaker correction, and 21ms is available for xover to sub. (That's where i mistakenly got 21ms before.)
In single amp mode, it means the entire speaker correction (excluding the sub), is being done with only 576 taps.
Ime, this is simply too low a tap count, to make decent corrections anywhere below about 1kHz.
(especially at 96kHz which halves taps effectiveness compared to 48kHz))
I've come to call what DEQX is doing in single-amp mode, correcting a speaker with xovers already in place, as "global FIR correction".
I believe it is a suboptimal technique at best, compared to DEQX's active mode.
Although in the case of the DEQX, given it only has about 4600 taps to allocate across all its multi-amp channels, the degree of suboptimal might be hard to hear.
Multi-amping, correcting driver by driver with sufficient taps for each channel, is audibly more optimal ime/imo.
Anyway, point of all this, is i'd suggest don't give up on FIR based on the DEQX. It's a beautiful piece of gear, but it's a long way from being a first class FIR platform, imho.
One other aside I picked up on from the manual. In the section about timing the subwoofer to the main speaker, it uses impulse peaks to establish time alignment.
This was practice for quite a while until a bunch of smart folks with good dsp tools started realizing it's a mistiming.
Most everyone today acknowledges its the beginning of impulse waveforms that need to be timed together, not peaks.
Just meant as another example of why not to give up on FIR (or your choice of whatever type processing)
Yep, what a nice elegant program. I wish i could get it to work on my computer...tried several times with no luck.
I think because of my experience in the music studio I immediately focus in on FR and quality of Eq. The accuracy in which filtering is implemented, is critical to final outcome...somehow the expensive stuff is still low quality in resolution regarding manipulation of the signal curve (in the loudspeaker application world). In my DAW my favorite linear eq has many points of control
Attachments
Last edited:
It does need the .net framework to function. You should already have that on Windows. I went back to the 32 bit version, the 64 bit just wasn't stable and many functions were dodgy.
The REW interface and capabilities with the HOLM's impulse math and audio driver awareness would be the killer app for sure.
Please don't blame bad horn sound from misuse. From my experience implementation is paramount and there are many pitfalls in horn speakers. My former 4 way horn system with tractrix horns played nicely, but the c2c could never be solved. Synergy horn takes everything to the next level. Much better polar response, better phase and excess group delay, much easier to time align and not just in one spot, better step response, etc. Sensitivity is not as high as big FLH systems, but I find the dynamics better in the Synergys, probably because of better integration between drivers.
For music production, I can probably tolerate "a lot"...like when I get several of these cpu intensive plugins going I have to increase the buffer size on my focusrite which in turn further increases latency....I just set my focusrite2 to 1024 buffer size and in Ableton it says 22ms output latency....for every 3rd party plugin I run, add 1-2ms of latency/per... if I recall? Linear plugins add more latency than that, per instance.....there are ways to keep plugin instances low and still use as many as needed, but I'll have over 20 plugins running on a project...
I won't tolerate high system delay because I play fps games competitively...I have since 2000 actually lol...not straight but you get what I mean...so I am well aware of latency vs input relationship...as well as your output (sound and visual) and the top competition want to see low numbers.... everywhere. So for the sound system you could see why I personally would sacrifice linear filtering for latency.
I don't understand how taps relate to resolution with minimum phase eq. I have around 16 points of parametric eq per channel with the crown input cards....whats the equivalent of that in the linear filtering world.
Last edited:
I do not get it, how is this done? You designed speakers that sounded a bit bad and kept on doing it?