High Frequency Compression Driver Evaluation

[weltersys]

Also, while the tests are interesting, I am wondering about the transition between the compression driver's short exit/throat and the throat of the horn. If there is a discontinuity there - the angle changes - then you may have a place where there is diffraction. Might be effecting the distortion and the HF response. Wondering about that.

_-_-bear

As stated in post # 1, I spent a great deal of time an effort insuring all horns have a smooth transition at the throat.

Having tested a few of the drivers on some other horns, I would be quite surprised if you could find another horn with less distortion at similar output levels.

At any rate, all drivers used the same horn type, which has the same or slightly greater throat angle than the drivers.
Any horn traits are equally reflected in all the responses.

 

[Pano]

Thanks Art - Positive and Negative dB, got it! 0dB is referenced to 4 volts, right?

 

[Pano]

Art, I'm not sure everyone understands what you've done here, there is a lot of info. It's a very clever test, so let me see if I can rephrase it:

Six different compression drivers were tested on the same model horn. Two crossover points were used, 630Hz and 1250Hz. There was no woofer! The woofer is "virtual" having been edited into the sound files from the original source (low passed).

Art made the recordings of the drivers on the horn, only. Then in software he added the bass from the original file, after it had been low pass filtered to match with the horn. Therefore, the woofer is "perfect" - only the compression drivers change.

The compression drivers were run at different drive levels to test their distortion. That recording was normalized and combined with the bass portion. So everything you hear in the test results is at the same volume, just the drivers and their distortion changes. You would not be able to do this with direct listening! A very clever test.

Does that about sum it up?

 

[hunter audio]

Great - n - Thanks

Suranjan Dasgupta

 

[weltersys]

Art, I'm not sure everyone understands what you've done here, there is a lot of info. It's a very clever test, so let me see if I can rephrase it:

Six different compression drivers were tested on the same model horn. Two crossover points were used, 630Hz and 1250Hz. There was no woofer! The woofer is "virtual" having been edited in the sound files from the original source (low passed).

Art made the recordings of the drivers on the horn, only. Then in software he added the bass from the original file, after it had been low pass filtered to match with the horn. Therefore, the woofer is "perfect" - only the compression drivers change.

The compression drivers were run at different drive levels to test their distortion. That recording was normalized and combined with the bass portion. So everything you hear in the test results is at the same volume, just the drivers and their distortion changes. You would not be able to do this with direct listening! A very clever test.

Does that about sum it up?

Yes, Pano, that mostly sums it up.

In addition to your summation, test recordings were also made with dual sine wave tones at different drive levels and RTA screens of each have been posted, making it possible to see the distortion you can hear.
Also, an inverse distance test was performed so one can estimate throat SPL at any given test SPL level.

Dual VHF tones (above most human's hearing) were also tested to see if they resulted in lower frequency audible distortion.

All the information is either in the first post, or footnoted at the bottom of post #1 as to where it may be found.

All in all I’m glad I performed the 50.25 hours of testing, but the 84.75 hours it took to mix, analyze, write up and post the findings got more than a bit old. A total of 135 hours of unpaid work in the project, while I had answered my own main questions in the first two days of testing a month ago.

As the hours of mixing and writing slowly unfolded, I often regretted mentioning on line that my “High Frequency Compression Driver Evaluation” was forthcoming .

But as a man of my word, I slogged along until done, with no short cuts, the “Full Monty” has been posted .

I appreciate the thanks and interest shown by the respondents thus far.

Art Welter

 

[weltersys]

An excerpt from Ed Kabotie’s song, “7 Cities of Gold” with voice, acoustic guitar and wood flute was used to compare the output of various HF drivers all fitted to the same type horn.
Ed Kabotie Freedom Songs - DreamCatcher.com

The specifics and SPL level of each music recording are in post #1 "Tests" section.

The file type suffix < .zip > must be changed to < .mP3 >, the music files will then open with your mP3 app.

Art

 

[weltersys]

An excerpt from Ed Kabotie’s song, “7 Cities of Gold” with voice, acoustic guitar and wood flute was used to compare the output of various HF drivers all fitted to the same type horn.
Ed Kabotie Freedom Songs - DreamCatcher.com

The specifics and SPL level of each music recording are in Post #1 "Tests" section.
This post contains the last of the sound files listed in Post #1.

The file type suffix < .zip > must be changed to < .mP3 >, the music files will then open with your mP3 app.

Art Welter

 

[ErnieM]

Thanks for all the work!

I've learned quite a bit over the years from your posts. It's a treat to be able to hear direct comparisons in a well thought out controlled test.

 

[weltersys]

Thanks for all the work!

I've learned quite a bit over the years from your posts. It's a treat to be able to hear direct comparisons in a well thought out controlled test.

Thank you, Ernie.

Even though I recorded the tests a month ago, the editing allowing the direct comparisons at the same levels was only finished a few days ago.

I look forward to hearing others opinions of the various driver's sound.

Art

 

[ErnieM]

I listened to all the music samples including horn only....and I have to agree with your conclusions. The bms drivers would do well crossed over high. I liked their top end detail. It was a toss up for overall sound between the b&c and ev. They really do sound quite similar when eq'd. Though as mentioned the ev goes louder cleaner.

It would be nice to see how the big BMS and B&C models do. Also 18 Sound of course. Too bad they're all quite expensive...

I did the horn only test awhile ago with my DE250's. It didn't sound so smooth after that.

 

[blakkshepeaudio]

Wow, it's going to take a while to assimilate all this info.

Big up for all the work and the serious attempt to provide a valid and meaningful set of tests. DIY to-the-max.

Rgds,
blakkisblakk

 

[keto]

Thats an amazing report! It looks like the BMS 4550 is a little bit more powerful than the BMS 4552 in the lowest part of the pass-band. For a Unity/Synergy horn, with a 4th-order effective high-pass (and 10-15dB of padding due to EQ) at 1200hz, and judging from all your measurements and listening, would there be much of a performance difference between the two drivers?

 

[weltersys]

There would not be much difference at 1200 Hz, but the 3 dB difference below 1200 Hz is enough to make me favor the 4550 if the cones used below it were not flat to 1200 Hz.

Art

 

[keto]

All the graphs I've seen make the 52ND look a little "smoother" than the 50. Does it sound a little smoother?

 

[weltersys]

Keto,
The difference in “smoothness” is quite little as you can see in post #33, the 4552 having a slight peak around 5000 Hz, the 4550 two small dips around 6500 and 8500.
It could be argued that the 4550, with lower and upper response more equal in level, is “smoother “ than the 4552.

The recordings were posted for you to listen to and make your own decisions, please share which sounds “smoother” to you.

Art

 

[Hentai]

Art,

Thanks for this impressive effort. The uploads are sure of great value for diy community.
The B&C performed pretty well.

I have a question though. I'm trying to figure a DIY compression driver and what I've noticed it behaves quite differently when modeling is done with different types of horns.
So my question/concern is that the test are only conclusive on that type of horn. Could the results be different with another horn?
If this matter has already been addressed please direct me to it.

I have seen many agitated spirits here so don't take my question the wrong way, its just a concern i stated in another topic i opened on compression drivers.

 

[weltersys]

I have a question though. I'm trying to figure a DIY compression driver and what I've noticed it behaves quite differently when modeling is done with different types of horns.
So my question/concern is that the test are only conclusive on that type of horn. Could the results be different with another horn?

Different horns using the same driver will result in different frequency response, pattern control, sensitivity, diffraction effects etc.

The relative differences in the drivers in my test can be interpolated to how they would respond on other horns.

Good luck with your DIY compression driver project.

 

[keyser]

Woow... I'm impressed! I'll do some listening tests when I find the time. Many thanks for sharing!

 

[ra7]

Art, I had looked at this several times, but never posted. I read Pano's summary above, and the first post could really use this high level summary. Could you add an abstract summarizing the test?

This is tremendous work. I haven't yet listened to the files, but I will soon.

Thanks!

 

[weltersys]

Art, I had looked at this several times, but never posted. I read Pano's summary above, and the first post could really use this high level summary. Could you add an abstract summarizing the test?

This is tremendous work. I haven't yet listened to the files, but I will soon.

Thanks!

Good point, summary added to the top of the (rambling) first post.

Had to edit the post just to fit it in, was already at the 30,000 word limit .