RTA comparison on vintage speakers

EDIT: See post 8 for a more scientific, data-based comparison.

Here's a Behringer DEQ2496 RTA comparison on five different speakers.

1. Dynaco A25 XL (has been recapped)
2. Indignia MC-1000 (old cabinet, new innards)
3. KLH 17 (has been recapped)
4. ZLH 20 (eWave variant, tweeter pot set at 12 o'clock, looks like the pot may need to be turned down a little ;-)
5. Taylo Reference Monitors (the control in experimental terms, cost way more than the rest of the pack)

ECM8000 mic was set up at 3 feet away on axis with the tweeter. Pink noise volume was not adjusted for each speaker, so overall level differences reflect the differences in speaker efficiency.

Comments??

Appears to me that more money may truly buy a flatter response, the Taylo's appear to have the flattest curve.

Wish I could overlay these charts for easier comparison, guess I could make an Excel chart out of the data (will do that when I get the time and post it here). On that note, I understand the ECM8000 needs phantom power to work with a PC soundcard and REQW software. What the hell is phantom power and where do you buy it?

To my eye, it looks like the A 25s actually might be smoother over most of the range than the Taylos -- not as deep, but the Taylos have a couple of fairly high amplitude peaks, and a humped overall response.
But I take my own comments with a grain of salt, since its hard to know what part of the response is actually the room contribution.
What a neat toy! Wish I had one. I can see how it might make training your ear rather easier.

Measurement microphones are typically condenser type, and require phantom power, which is a flea-power current-limited DC voltage superimposed on the signal lines by the mic preamp, i.e., there's still only two wires to the mic. It biases the microphone element for proper function, and is removed via simple capacitor coupling at the preamp input to extract the signal.

'Course, I'm always most interested in how speakers I was involved in designing perform in comparison to others. Your Indignias are running +/- 2.5 dB over the entire spectrum, and some of that variability may be room modes and reflections. With practice and attention to measurement conditions, you'll learn how to clean some of that up. Overall, I'd say they are the best performers in the group, with the flattest response, as measured.

ZLH-17 is by far the flattest above 2.5 kHz, with the most extended VHF. The highs are indeed running too "hot" there, and it may be just a simple matter of adjusting the L-pad down a bit more; the flat HF response will remain intact, but be better balanced with the woofer. You can watch the response change in the RTA active mode as you do that, then switch back to "Average" to view the result. Also, draw up a simple protractor on cardboard to measure the response at different off-axis angles, 7.5° or 10° increments, perhaps, to check dispersion, and try some nearfield bass response measurements as well to see how they're all performing at the low end.

DEQ2496 is WAY fun; there's about a half-dozen AKers doing speaker measurements with it now, and y'all should settle upon a particular thread for the primary purpose of exchanging tips and techniques on how to use it most effectively, and sharing your experiences, a "DEQ2496 Clinic," as it were, maintained strictly on topic. Much of the information you develop there will also be applicable by others working with different RTA measurement systems.

Read up on how to use its AutoEQ function, and you'll get to hear how any of your speakers would sound if they were actually playing flat, or with ~1 dB/octave sloped response in the "Room Corrected" mode. Tons of cool stuff in that device.... :thmbsp:

BmWr75 said:

Wish I could overlay these charts for easier comparison, guess I could make an Excel chart out of the data (will do that when I get the time and post it here).

Click to expand...

You can read the dB measurement in each of the bands using the cursor and transcribe them into a spreadsheet. That's labor intensive, but works well enough to generate curves on individual drivers, for example, suitable for use in crossover modeling software, even.

There should also be means to digitize it directly from the images in a graphics program, but I have ZERO knowledge or skill with using those, alas....

I gotta get me one of these :yes:

markus said:

I gotta get me one of these :yes:

Click to expand...

Do it.

BmWr75's post demonstrates how easy it is to measure and compare the actual performance of various speakers, and ascertain and analyze any "issues" they may exhibit.... :thmbsp:

Scott, most microphone preamps supply phantom power nowadays. I don't think your DEQ2496 does, however. It's not listed in the specs. You may have to find a minimal preamp that will do that for you. Generally, if the unit will provide the power, it says so somewhere in the lit.

I'm going a different measurement route, with a TASCAM 144, a Behringer mic and a laptop with some of the software we all know and sometimes love. Skywave-Rider is watching to see how it works for me as he also wants an extremely portable solution. The TASCAM unit is a combo preamp and signal processor, w/MIDI interface and some other neat functions and it supplies the phantom power. Felt the need for more precision than the SPL meter gave, although it's still a handy tool for quick studies.

If anyone is interested in the TASCAM, Guitar Central has them on sale combined with a 2-microphone package, including cables. Good deal! If I had seen that one first, I'd have sprung for it. $200, IIRC. I have no relation to Guitar Central.

Anyhow, your RTA's are pretty neat. Looks like you might still have some room affects there, as Zilch said. All in all, pretty neat, though! Can't wait to be doing some of that myself. :thmbsp:

EDIT: I take that back. Reread the 2496 lit. It says you've got phantom power on the mic input. There may be a button you have to push to make it happen.

OK, after much eye strain, here's the 5 speaker comparison data put into Excel and then copied/pasted to Word files for uploading here. I normalized the frequency response data for each speaker to center it at -75 dB. This just corrects for the fact I didn't adjust the volume level when taking the measurements (which allowed me to do the efficiency comparison below). The KLH 17 was the most efficient speaker, so used it as the reference and arbitrarily designated it 0 dB. So, the Indignia MC-1000s are 6.6 dB less efficient than the refurbished KLH 17s.

Efficiency (measured at 3 feet, same voltage input)

Indignia MC-1000 = -6.6 dB (click link to see build thread)
Taylor Ref Monitors - -4.6 dB
ZLH 20 = -2.6 dB (click link to see build thread)
Dynaco A25 XL = -1.4 dB
KLH 17 = 0 db

After close examination of the charts, here's what I see.

The Indignias have the flattest response curve (Zilch was right...as usual).

The KLH 17s have the worst bass and are very peaky above 1250 Hz.

Taylo Ref Monitors.....cost 9 times more than the Indignias, but don't play as flat.

The ZLH 20s don't have quite as good bass extension as the Indignias and I need to turn the Lpad on the horns down. Other than that, they look a lot like the Indignias, no surprise since both were designed by the Zilch and other AKers.

Dynaco A25 XLs, compared to the Indignias, the bass starts rolling off at 80 Hz and highs do the same thing a little bit above 10000 Hz.

Bottom-line: I got about $200 cash invested in the Indignias (everything bought at Parts Express), some of my time and a couple of free Realistic MC-1000 cabinets a buddy gave me. I'd say that is good bang for the buck.

Edit: Two more vintage speakers added and one remeasured/adjusted, same measuring technique as descirbed above:

1. ZLH 20 - remeasured and adjusted the tweeter pot to as flat as possible
2. Original Large Advent - tweeter switch set to Normal
3. KLH 6 - tweeter switch set to Normal

Using the Indignia MC-1000 as the flat response benchmark, here's what I see.

The ZLH 20 does not have the >10,000 Hz peak and is not quite as flat in the 630 - 10,000 Hz range. Otherwise, it matchs the Indignia overall pretty well below 630 Hz and rolls off the bass slightly faster.

The OLA has better bass extension below about 50 Hz and rolls off the highs very quickly above ~12,000 Hz. Between 1000 Hz and 5000 Hz, it is a roller coaster ride of peaks and valleys. But, from 30-400 Hz it plays very flat, except for a dip at 80 Hz.

The KLH 6 matches the Indignia frequency response well up to about 1800 Hz. But between 1800 and 5000 Hz, there is a large peak in comparison. The highs start rolling off quickly above about 14,000 Hz.

The last Word file shows the response curves for all 7 speakers on one chart. It is busy, but shows the OLA has the best bass extension of the group. But it is also graced with the most rolled off highs. To my eyes, the worst bass award goes to the stock, recapped KLH 17 and it also wins the award for having the worst frequency response curve with respect to flatness. It is all over the place.

No opinions, just the facts. Hope you have enjoyed this audio science experiment. Plan to add a Dynaco A25, JBL L112 and maybe some old Heath DIYs to this analysis one of these days.

For now is it time to go listen to some music.

Regards, BmWr75

well, the real question is which ones you prefer to listen to.

mhardy6647 said:

well, the real question is which ones you prefer to listen to.

Click to expand...

Am currently listening mostly to a set of speakers not on the above list, ADS L1230s.

AR2Axs are the other pair I spending some time with on the tube system.

The ZLH 20s are in the upstairs great room and are getting some listening time also.

The Dynacos and Taylos share a downward-sloping HF response, generally considered to be pleasant listening. The Dynacos are lacking VLF; they need sub(s), but the Taylos seem to be losing bass response below 200 Hz. They could benefit from boundary reinforcement or Baffle Step Compensation. You really need to do nearfield measurements to get an accurate look at the LF responses.

Your KLH-17s measure very much as they do in ZilchLab - rolled-off at both ends of the spectrum, and a roller-coaster ride in between with forward highs.

I do flat and accurate, for the most part, though I'm careful not to allow rising response in the highs, and sometimes tilt them downward if they don't sound clean up there, or the speakers are too bright overall. AutoEQ with Room Correction will show you how Indignias and ZLH-17s would sound mimicking the Dynaco/Taylo response characteristic.

You'll soon develop a recognition of the correllation between what you see and what you hear, if that hasn't "clicked" already.

[Yah, Indignias are "special;" I believe we were all very pleased with the outcome of that AK Collaborative project.... :thmbsp: ]

Paratima said:

Scott, most microphone preamps supply phantom power nowadays. I don't think your DEQ2496 does, however.
EDIT: I take that back. Reread the 2496 lit. It says you've got phantom power on the mic input. There may be a button you have to push to make it happen.

Click to expand...

Les, the DEQ2496 does supply phantom power, don't have to do anything special, just plug the mic into the back of the unit on the input marked MIC. :yes:

Thanks for the info on the TASCAM system. I'd like to have a way to use the ECM8000 mic with the REQW software. That will make comparing different speakers a lot easier than the exercise I went thru today transcribing the data into Excel.

By a casual look at the DEQ displays, I noticed all units show a notch at ~315Hz. That can't be a fault of all units at the same time so I am suspecting room resonances. Had a similar experience when "testing" my satelite+sub combo in the living room. It showed a larger (more than 10dB) notch at 125 and 200Hz which is audible as a fuzzy and boomy bass response.

The morale of my short experience so far is that before I can use tools to evaluate loudspeakers I need to evaluate the measurement technique and prove the method (at least to me) before any of the readings make any sense. For example, when the mike is in the listening position, several feet away from the speakers, it's not the speakers that are measured but mostly how they interact with the room - an altogether interesting subject by itself.

Out of curiosity, does the DEQ use a calibrated ECM8000 mic? I've seen several correction curves for that mic around showing that it's readings must be corrected before put to use, especially in the <100Hz and >15kHz range. Does the DEQ take any such calibration into consideration?

SaSi said:

Does the DEQ take any such calibration into consideration?

Click to expand...

We don't know for sure, but they are certainly aware of this issue.

Best guess is that there's a correction curve in there based upon statistical ECM8000 mic samples; if there's an ability to install a custom one, that'd be an undocumented "feature."

I wholeheartedly agree the readings I've posted should not be taken as absolutes. Am fully aware my man cave is not the perfect infinite baffle. However, the readings on different speakers should be comparable because they are taken under the same conditions.

For reference, the five readings (to date) in this post were taken directly in front of my home theater center channel speaker. There is a 102" fixed HT screen on the wall behind the speaker being measure, which should do a good job of absorbing a lot of rear wall reflections (that is my guess at least). This puts the front baffle of the speaker being measured about 3' 4" from the rear wall, a very long way from the side walls and about 14' from the rear wall. The ceiling in the room is acoustic panels and about 7' 6" tall.

BmWr75 said:

Here's a Behringer DEQ2496 RTA comparison on five different speakers.

1. Dynaco A25 XL (has been recapped)
2. Indignia MC-1000 (old cabinet, new innards)
3. KLH 17 (has been recapped)
4. ZLH 20 (eWave variant, tweeter pot set at 12 o'clock, looks like the pot may be to be turned down a little ;-)
5. Taylo Reference Monitors (the control in experimental terms, cost way more than the rest of the pack)

ECM8000 mic was set up at 3 feet away on axis with the tweeter. Pink noise volume was not adjusted for each speaker, so overall level differences reflect the differences in speaker efficiency.

Comments??

Appears to me that more money may truly buy a flatter response, the Taylo's appear to have the flattest curve.

Wish I could overlay these charts for easier comparison, guess I could make an Excel chart out of the data (will do that when I get the time and post it here). On that note, I understand the ECM8000 needs phantom power to work with a PC soundcard and REQW software. What the hell is phantom power and where do you buy it?

Click to expand...

Is one of the dimensions of your room approximately 14'? You have room modes at 40 and all multiples of that frequency, which corresponds to 14' (I just read the rest of the posts and you mentioned that it was about this distance from a wall). This would indicate a live room. Anything you can do to increase sound absorption will minimize these cancellations and smooth the response. Bass traps will smooth the bass response, too.

Move the speakers to different positions to find the best location. It will vary from one speaker to another. Also, the position of the mic makes a big difference.

Mics can be dynamic- which has a diaphragm attached to a coil in a magnetic field, electret- which has a battery (usually a AA) to provide the energy for the magnetic field that needs to vary in order for the signal to be created, or condenser- which has a capacitor that's specially designed to be "microphonic", like a bad tube or capacitor. The Phantom power is for charging the capacitor and is usually 48V. Sound cards don't generally have the capability to provide 48V, so an input module, like M-Audio, Tascam, Line 6 or some other brand is needed. You wouldn't need to worry about the sound card when you use this kind of device because it goes in through a USB port.

I usually run my measurements in the 15 dB range. Shows all the warts. :tears:

I usually use my measurements in the 15 dB range. Shows all the warts. But then, you do hear the warts. :tears:

SaSi said:

By a casual look at the DEQ displays, I noticed all units show a notch at ~315Hz. That can't be a fault of all units at the same time so I am suspecting room resonances. Had a similar experience when "testing" my satelite+sub combo in the living room. It showed a larger (more than 10dB) notch at 125 and 200Hz which is audible as a fuzzy and boomy bass response.

Click to expand...

The room is in control below 200 Hz, not the speaker, and even time-windowed measurement technology cannot resolve these lower frequencies. We have two straightforward solutions: nearfield LF measurements according to Small, Keele, et. al, or measure in a VERY large room, i.e., out-of-doors.

Anyone reading response measurements must do so in the context of the limitations of the particular methodology used to generate them; there are no anechoic chambers large enough to capture the lowest octaves accurately, which is why rooftops and parking lots are common measurement venues.

For the most part, with RTA indoors, anything measured below 1 kHz reflects influences of the room under the best of conditions, and the room modes are well resolved. Consider measurements in the range below that frequency as progressively more indicative rather than definitive of the true low frequency response.

Nearfield response, BTW, is considered of similarly limited utility above 200 Hz. Read up in Dickason and D'Appolito, particularly. The Keele paper is #6 here:

http://www.xlrtechs.com/dbkeele.com/papers.htm

toddalin said:

I usually use my measurements in the 15 dB range. Shows all the warts. But then, you do hear the warts. :tears:

Click to expand...

Response deviations less than +/- 2 dB are generally considered acceptable, even if audible, and anything less than +/- 1 dB, "academic."

I can tweak any speaker to +/- 0.25 dB with high-res parametric EQ, so long as nobody farts in the room, and yes, I can also post measurement data with sufficient resolution to demonstrate the effect, should anyone care to "anal"yse it....

Two more vintage speakers and a remeasured AK design have been added to the 8th post in this thread.