• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

(Informal) Chat with Dr. Earl Geddes of GedLee Audio

Jim Matthews

Major Contributor
Forum Donor
Joined
Mar 25, 2021
Messages
1,051
Likes
1,286
Location
Taxachusetts
So, that would make the SINAD rating not quite as important as the THD numbers, correct?
Color me curious about this.

I've owned lots of downmarket tube gear, mostly built on the Williamson topology. My cohabitation with an amplifier was inversely proportional to the amount of Global Feedback in the design.

My last amp was a direct heated triode with zero global feedback.

I had it for nearly three years.

Lord knows how it might measure.
 

puppet

Senior Member
Joined
Dec 23, 2020
Messages
446
Likes
284
@KSTR ... what you state about the VC is correct .. and a secondary effect of that is the crossover components going out of electrical spec resulting in crossover shift and the issues that brings. This is my understanding of the thermal effects imposed by power compression.
 

Tim Link

Addicted to Fun and Learning
Forum Donor
Joined
Apr 10, 2020
Messages
722
Likes
638
Location
Eugene, OR
Interesting that some masking actually increases at high volume. Also that Ford found out that people aren't happy if their sound system doesn't distort when they turn it up. I've also never heard of diffraction being thought of as a kind of group delay.
 

Newman

Major Contributor
Joined
Jan 6, 2017
Messages
3,449
Likes
4,210
Well, that’s what I thought until I saw this interview :)

More specifically, I don’t think they discussed why exactly the big, high efficiency speakers sound better at high SPL. What specific thing is responsible, if not the normal THD or IMD measurements?

Well, until we see a controlled listening test to confirm it, we don't even know if it's truly in the sound waves. Until then, the most likely explanation will be non-sonic: expectation effect, confirmation bias, placebo effect, etc etc etc.

cheers
 

Newman

Major Contributor
Joined
Jan 6, 2017
Messages
3,449
Likes
4,210
Wow, the JBL 305 actually does great there.

Or, looking at it from the opposite direction, if a small cheap speaker can pass this test so well, then this compression test as graphed is not really a useful test to distinguish between decent speakers.

cheers
 

sgent

Member
Joined
Mar 24, 2020
Messages
99
Likes
127
Or, looking at it from the opposite direction, if a small cheap speaker can pass this test so well, then this compression test as graphed is not really a useful test to distinguish between decent speakers.

cheers


Given the gross margin on speakers in general, it would not surprise me if someone taking an evidenced based approach to speaker design could build a cheap speaker that vastly outperforms one that costs an order of magnitude more -- especially if you only care about performance and not looks, etc. Also for a higher volume item like a JBL 305 you can amortize the design costs over a much larger number of units.
 

Newman

Major Contributor
Joined
Jan 6, 2017
Messages
3,449
Likes
4,210
I agree with your statement. There will be some bigger and more expensive speakers that don't do as well as the 305 at compression linearity. But I also bet that the vast majority of them will pass those tests as well as the 305, meaning the test isn't useful for choosing which speaker to use.

Most audiophiles buy speakers for filling a room with good sound when placed away from the walls and listening distances of 2.5-4m, and if they are listening to good advice, they will be buying decent sized speakers with decent output and power handling for the job. I bet, when you take those speakers, and graph the compression linearity as per the above samples, at 86 and 96 dB at 1 metre, they will all get a pass with non-meaningful differences.

What I would like to see, would be the same test to a higher standard, that refers to the HT industry convention of listening to movie playback with peak levels of 105 dB at the listening position, from each speaker. So let's measure it at 1m, 2m, and 4 m for the following use cases: 105 dB at 1m for desktop users, 105 dB at 2m for small room users, and 105 dB at 4m for large room listeners.

That will tell us the approx room size that a speaker will handle while delivering realistic peak levels that are not compressed. And we might find the 305 actually doesn't do so great, even at 1m for desktop use.

cheers
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,386
Location
Seattle Area
What I would like to see, would be the same test to a higher standard, that refers to the HT industry convention of listening to movie playback with peak levels of 105 dB at the listening position, from each speaker. So let's measure it at 1m, 2m, and 4 m for the following use cases: 105 dB at 1m for desktop users, 105 dB at 2m for small room users, and 105 dB at 4m for large room listeners.
With what test signal? A flat sweep is liable to overdrive the tweeter with unrealistic power. In my case I am dealing with borrowed speakers that I cannot risk damaging.
 

Newman

Major Contributor
Joined
Jan 6, 2017
Messages
3,449
Likes
4,210
I'm not expert enough to discuss test signals and damage to drivers. But, from the consumer point of view, it is most useful to see test data on linear compression that represents the use case where linear compression is likely to occur, and also correlates with the established playback convention for peak levels.

Perhaps the simple note that "I don't have enough confidence in this driver to sweep it at 105 dB at 1,2,4 metres" tells us something, compared to drivers where you do, and do show the result.
 

Sancus

Major Contributor
Forum Donor
Joined
Nov 30, 2018
Messages
2,923
Likes
7,616
Location
Canada
What I would like to see, would be the same test to a higher standard, that refers to the HT industry convention of listening to movie playback with peak levels of 105 dB at the listening position, from each speaker. So let's measure it at 1m, 2m, and 4 m for the following use cases: 105 dB at 1m for desktop users, 105 dB at 2m for small room users, and 105 dB at 4m for large room listeners.

That will tell us the approx room size that a speaker will handle while delivering realistic peak levels that are not compressed. And we might find the 305 actually doesn't do so great, even at 1m for desktop use.

Erin does 102dB on the instantaneous linearity test, and yes, the 305p(review) won't even play that loud, the limiter kicks in. As you can see in the graph.

Personally I think this is adequate. I don't think 105dB makes any sense in the long term compression tests, even 96dB RMS is unrealistically high.
 

Newman

Major Contributor
Joined
Jan 6, 2017
Messages
3,449
Likes
4,210
Maybe that's the answer: do instantaneous at 105 dB at 1, 2, 4 metres. After all, the convention I described is about peak levels.
 

Sancus

Major Contributor
Forum Donor
Joined
Nov 30, 2018
Messages
2,923
Likes
7,616
Location
Canada
Maybe that's the answer: do instantaneous at 105 dB at 1, 2, 4 metres. After all, the convention I described is about peak levels.

I think this would require a downsloping test signal tbh, even 102dB seems like it might be pushing it. No tweeter will ever play that loud because almost all the SPL in real content is from 20-1000hz. As @amirm states, there is probably a high risk of killing tweeters with 105dB sweeps even on some big floorstanders.

E: I could see an argument for a bandwidth limited sweep to 1khz.
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
I'm not expert enough to discuss test signals and damage to drivers.

<expert enters chat>
I've tested 100+ drivers over the past 10 years so I can tell you that I would absolutely kill <some> tweeters if I tested them at that kind of level even with a bandlimited sweep. I've done it at even lower output levels with "lesser" tweeters. That always sucks. And I used to spend time trying to work my way up to failure so I could identify that point. But over the years, I realized it is a lot of extra work for nothing. I try to make my tests streamlined not just for me but for my audience as well and really don't want to have a bunch of caveated tests where I have 105dB here, 108dB there, 96dB here, etc.

So, yes 102dB is pushing it. Not just for tweeters but for smaller midwoofers as well. Especially if they have bad suspension that doesn't limit the excursion well and then you wind up with a VC outside of the gap permanently. Some can play it. Others cannot. But that goes back to my streamlining of the test.

The 105dB you quote is peak. My 102dB peak is only 3dB off and that's a comfortable area for me. My updated compression tests (long-term) use 86dB and 96dB. The THX quote is 85dB with 20dB headroom (link). Factor in distance, stereo pair in a room... the numbers can change around. But, I'm good with what I have settled on. FWIW, the majority of speakers I tested with Max SPL topped out around 98-103dB @ 1m. Another bit of information I had in the back of my mind when I chose the values I chose.

TL;DR
I chose the levels I did based on my experience of testing drive units and loudspeakers and am happy with them. My goal isn't to test until failure. My goal is to have an easy comparison of trends. If someone else wants to stress test drivers past that then they are welcomed to it. :)
 
Last edited:

Newman

Major Contributor
Joined
Jan 6, 2017
Messages
3,449
Likes
4,210
IMO then, the TLDR should be there is no test level for this that is both meaningful and safe for drivers, if the 305 passes it with flying colours.

My suggestion: why not do the 1m test at increasing level until the DUT shows (say) 1 dB of non linearity, then report that level?

cheers
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
IMO then, the TLDR should be there is no test level for this that is both meaningful and safe for drivers, if the 305 passes it with flying colours.

My suggestion: why not do the 1m test at increasing level until the DUT shows (say) 1 dB of non linearity, then report that level?

cheers

Well, like you said:
“IMO”

And that’s your opinion. You’re entitled to it. And I am entitled to the method I have chosen.

Of course, you are always welcome to do such a series of tests yourself and publish them. That’s not sarcasm. At some point, if you believe that’s the only way to get useful measurements then what is stopping you from taking the next step? I mean, that’s why I got involved in testing. I didn’t like the way things were done so I took it upon myself to make it happen. It’s an easy enough series of tests to do if that’s what you want to do. Easier than SPIN data. You can use (free) REW and a $99 USB mic and follow @napilopez ’s tutorial on how to obtain quasi-anechoic measurements. Then figure out how to set up the power steps yourself. Post-process the data and share here. I’d be happy to give some advice if you need help, though it sounds like you already have the process you want.
 
Last edited:

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,403
Likes
5,296
Location
Somerville, MA
Well, like you said:
“IMO”

And that’s your opinion. You’re entitled to it. And I am entitled to the method I have chosen.

Of course, you are always welcome to do such a series of tests yourself and publish them. That’s not sarcasm. At some point, if you believe that’s the only way to get useful measurements then what is stopping you from taking the next step? I mean, that’s why I got involved in testing. I didn’t like the way things were done so I took it upon myself to make it happen. It’s an easy enough series of tests to do if that’s what you want to do. Easier than SPIN data. You can use (free) REW and a $99 USB mic and follow @napilopez ’s tutorial on how to obtain quasi-anechoic measurements. Then figure out how to set up the power steps yourself. Post-process the data and share here. I’d be happy to give some advice if you need help, though it sounds like you already have the process you want.

The ARTA manual has some useful notes about mic selection; finding a mic which doesn't distort under these high SPLs is something which needs to be considered. One of the many reasons I don't do many distortion measurements..
 

Newman

Major Contributor
Joined
Jan 6, 2017
Messages
3,449
Likes
4,210
@hardisj a bit too defensive there mate. Chill.

When you start punching out flat lines for 99% of non-shoebox-sized speakers, you might be more interested in genuine feedback that could have saved you a lot of wasted effort.

And my suggestion, that you blanked, of measuring and reporting the SPL at 1m that yields 1 dB of compression, and makes speakers meaningfully distinguishable, might have deserved a better response.

cheers
 
Last edited:
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
@hardisj a bit too defensive there mate. Chill.

No need for me to chill. I'm snowman status over here, mate. I stated my reply in the most civil way I could manage via internet text. Even stated "no sarcasm". I don't know how else to put a matter-of-fact reply in more friendly terms.

As I said, you are welcome to perform your own tests.

Cheers.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,386
Location
Seattle Area
Perhaps the simple note that "I don't have enough confidence in this driver to sweep it at 105 dB at 1,2,4 metres" tells us something, compared to drivers where you do, and do show the result.
What do you get out of a "sweep" with respect to peak playback capability?
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
The tube amp thing actually came up later in the video. I sounded like Dr. Geddes view is that while tube amps do have higher 2nd harmonic distortion, it's usually not high enough to overcome the very high masking our ears have for such distortion. So, people probably don't like tube amps for that higher distortion, but rather for some other reason,

Tube amps typically have higher output impedance than solid-state amps, so that with typical speaker impedance curves, they will sound warmer.

My Rogue M-180s have about 0.5 dB more amplitude in the bass, which is enough for a subjectively warmer sound.
 
Top Bottom