• 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!

Measuring dynamics

Keith_W

Major Contributor
Joined
Jun 26, 2016
Messages
2,531
Likes
5,803
Location
Melbourne, Australia
I did a quick search and could not find an answer to my question, so if this has been asked and answered before, please direct me to the thread or merge this with an existing thread.

Like most audiophiles, I started off as a subjectivist. I listen to music, I know what I like, I then go on a journey to buy equipment that provides sound reproduction that I like. Along the way, I started to learn that measurements can predict aspects of performance that I like and dislike, so I started paying attention to them. These days I have a full active system with 8 DAC channels, 8 power amps, each directly connected to its own driver with all sorts of corrections done by DSP, so I can really tailor the sound to my liking and would make most purist audiophiles cringe. In the end it is my music system which I use to listen to music, so if I don't like the sound of the Harman curve I do away with it (which I did). My own target curve that I have tuned the system to create is quite different. I guess you could call me an objective subjectivist. I use objective methods, but in the end my music listening is subjective.

Sorry for the preamble, but I had to go on that to describe one aspect of objective audio that continues to elude me: dynamics.

In subjective terms, not only does this mean that there is enough headroom to go from soft to loud without the amplifier clipping or the speaker distorting, it also refers to the sense of ease that the system is able to reproduce soft-loud transitions. There are microdynamics and macrodynamics, but after years of listening I have decided that these two phenomena are one and the same. There is no such thing as a speaker that has one and not the other.

I was wondering if there is a measurement that predicts how "dynamic" a speaker would be. Impulse response?

Also, is there such a thing as DSP which would improve the dynamics of a system by exaggerating the contrast between voltage spikes in the signal?

The reason I ask is because my system is a hybrid one: horns for mids and treble, and conventional drivers for the sub (20 - 80Hz) and bass (80 - 600Hz). There is an obvious disconnect between the conventional drivers and the horns, and I can tell you it is NOT the frequency response, NOT amplifier headroom, and NOT distortion, and I have all the measurements to prove it. Well, not with me right now because I am in a hotel room. The conventional drivers just sound ... less dynamic, and I can not explain why using measurements. So please tell me where to look.
 

dasdoing

Major Contributor
Joined
May 20, 2020
Messages
4,209
Likes
2,675
Location
Salvador-Bahia-Brasil
after switching to horn mids/treble myself I searched for the answer, but couldn't find it.
It can have 2 reasons:
1) the rise time of the driver; if you get a membrane to vibrate at a certain aplitude, it wont achive this amplitude at the first wave. compression drivers might just be faster getting there?
2) it could be just dispersion related. horns will give you more direct sound and less reflected sound. So the reproduction is less smeared.
 

Hayabusa

Addicted to Fun and Learning
Joined
Oct 12, 2019
Messages
787
Likes
519
Location
Abu Dhabi
I did a quick search and could not find an answer to my question, so if this has been asked and answered before, please direct me to the thread or merge this with an existing thread.

Like most audiophiles, I started off as a subjectivist. I listen to music, I know what I like, I then go on a journey to buy equipment that provides sound reproduction that I like. Along the way, I started to learn that measurements can predict aspects of performance that I like and dislike, so I started paying attention to them. These days I have a full active system with 8 DAC channels, 8 power amps, each directly connected to its own driver with all sorts of corrections done by DSP, so I can really tailor the sound to my liking and would make most purist audiophiles cringe. In the end it is my music system which I use to listen to music, so if I don't like the sound of the Harman curve I do away with it (which I did). My own target curve that I have tuned the system to create is quite different. I guess you could call me an objective subjectivist. I use objective methods, but in the end my music listening is subjective.

Sorry for the preamble, but I had to go on that to describe one aspect of objective audio that continues to elude me: dynamics.

In subjective terms, not only does this mean that there is enough headroom to go from soft to loud without the amplifier clipping or the speaker distorting, it also refers to the sense of ease that the system is able to reproduce soft-loud transitions. There are microdynamics and macrodynamics, but after years of listening I have decided that these two phenomena are one and the same. There is no such thing as a speaker that has one and not the other.

I was wondering if there is a measurement that predicts how "dynamic" a speaker would be. Impulse response?

Also, is there such a thing as DSP which would improve the dynamics of a system by exaggerating the contrast between voltage spikes in the signal?

The reason I ask is because my system is a hybrid one: horns for mids and treble, and conventional drivers for the sub (20 - 80Hz) and bass (80 - 600Hz). There is an obvious disconnect between the conventional drivers and the horns, and I can tell you it is NOT the frequency response, NOT amplifier headroom, and NOT distortion, and I have all the measurements to prove it. Well, not with me right now because I am in a hotel room. The conventional drivers just sound ... less dynamic, and I can not explain why using measurements. So please tell me where to look.
I assume the difference in directivity between the low/mid/high is what you hear.
The mids and highs will have less room refections and will sound more 'direct' or 'dynamic' I guess.
 

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,219
Likes
1,360
Location
Budapest
@Keith_W
Not sure if this will help, but you can try: are all your drivers properly time-aligned?
You can see that in your overall step response curve when zoomed in
It shall have just one peak (=all drivers time aligned) instead of several ones (=drivers not time aligned)

I agree with the above comments re. horns having more direct sound (and therefore less reflections) but again, not sure if that is what you hear when you mention there is a disconnect between the drivers
 
OP
Keith_W

Keith_W

Major Contributor
Joined
Jun 26, 2016
Messages
2,531
Likes
5,803
Location
Melbourne, Australia
Hi all, thanks for your responses. I am aware that horns are more directional, but both my speakers are toed in so it shouldn't be a problem unless we are talking about room reflections. Even then, I have the room reflections cancelled out using DSP.

ppataki, my drivers are all time aligned, it was one of the first things I did. Although, I have read arguments elsewhere that time alignment is not audible unless there is a very wide disparity between the drivers. I can time align my drivers to a 1ms accuracy which I suspect is far more accuracy than I need because all I need to do is move my microphone and the time alignment is off again by more than 1ms. Regardless, it is as time aligned as I can make it.

And yes, I suspect that "dynamics" is about the rise time of the wave form. Is this measured in a standard sine sweep? If so, where?
 

Hayabusa

Addicted to Fun and Learning
Joined
Oct 12, 2019
Messages
787
Likes
519
Location
Abu Dhabi
Hi all, thanks for your responses. I am aware that horns are more directional, but both my speakers are toed in so it shouldn't be a problem unless we are talking about room reflections. Even then, I have the room reflections cancelled out using DSP.

Yes I think it is about room reflections.
DSP can cancel out the frequency response errors caused by reflections (on one listening position) but not the audible effect that the sound comes from different directions.
 

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,219
Likes
1,360
Location
Budapest
I suspect that "dynamics" is about the rise time of the wave form. Is this measured in a standard sine sweep? If so, where?
There might be members here who might not agree but I personally use this to measure the 'speed' (=the rise time and the decay time of the impulse response in the function of frequency)

In REW go to Spectrogram and set it like this:
1655128476146.png


The span before and span after times you need to set according to your needs - for low frequencies you need to set a wider range, for high frequencies you need a tighter range

It will look something like this:

1655128629886.png


You will also be able to see the Energy Delay (Peak Energy Time)
This graph can also show you if your filters are minimum phase or linear phase (the latter makes the IR curve more symmetrical while the first will have a faster rise time but a much slower decay time)
 

Dj7675

Major Contributor
Forum Donor
Joined
Jan 12, 2019
Messages
2,116
Likes
2,782
I did a quick search and could not find an answer to my question, so if this has been asked and answered before, please direct me to the thread or merge this with an existing thread.

Like most audiophiles, I started off as a subjectivist. I listen to music, I know what I like, I then go on a journey to buy equipment that provides sound reproduction that I like. Along the way, I started to learn that measurements can predict aspects of performance that I like and dislike, so I started paying attention to them. These days I have a full active system with 8 DAC channels, 8 power amps, each directly connected to its own driver with all sorts of corrections done by DSP, so I can really tailor the sound to my liking and would make most purist audiophiles cringe. In the end it is my music system which I use to listen to music, so if I don't like the sound of the Harman curve I do away with it (which I did). My own target curve that I have tuned the system to create is quite different. I guess you could call me an objective subjectivist. I use objective methods, but in the end my music listening is subjective.

Sorry for the preamble, but I had to go on that to describe one aspect of objective audio that continues to elude me: dynamics.

In subjective terms, not only does this mean that there is enough headroom to go from soft to loud without the amplifier clipping or the speaker distorting, it also refers to the sense of ease that the system is able to reproduce soft-loud transitions. There are microdynamics and macrodynamics, but after years of listening I have decided that these two phenomena are one and the same. There is no such thing as a speaker that has one and not the other.

I was wondering if there is a measurement that predicts how "dynamic" a speaker would be. Impulse response?

Also, is there such a thing as DSP which would improve the dynamics of a system by exaggerating the contrast between voltage spikes in the signal?

The reason I ask is because my system is a hybrid one: horns for mids and treble, and conventional drivers for the sub (20 - 80Hz) and bass (80 - 600Hz). There is an obvious disconnect between the conventional drivers and the horns, and I can tell you it is NOT the frequency response, NOT amplifier headroom, and NOT distortion, and I have all the measurements to prove it. Well, not with me right now because I am in a hotel room. The conventional drivers just sound ... less dynamic, and I can not explain why using measurements. So please tell me where to look.
I believe our perception of dynamics to be related to dispersion somehow. If we take 2 speakers that have very little/no compression and low distortion, provide them with ample power... the one with a narrower dispersion will seem more dynamic. What I find interesting is the in @amirm listening tests, he commented specifically on dynamics in 4 JBL models.... 708p, array 880, 4439, and srx835p. Al of these I believe have well controlled dispersion but under 50 degrees. In a completely subjective, unscientific observation.. I swapped out a Revel C208 with a JBL 708P and the perception of dynamics in movies is very noticeable. C208 is around 80 degrees and the 708p is around 50 degrees.
 

DVDdoug

Major Contributor
Joined
May 27, 2021
Messages
2,920
Likes
3,835
Generally, speakers are linear and they don't compress the dynamics (unless they are overdriven). When you start limiting or "pushing down" the peaks, that's regular-old harmonic distortion. So yes, you can drive a speaker in to distortion.

There are "rumors" of voice coils (or magnets?) overheating and the speaker loosing efficiency/sensitivity. if the speakers temporarily lose their efficiency and that would be a kind of slow-compression. Mark me down as "mostly skeptical". ;)

But... People do perceive some speakers as being more dynamic than others... It might just be that more efficient speakers give that impression (I think that's true with me) or maybe it's something else.

There are microdynamics and macrodynamics, but after years of listening I have decided that these two phenomena are one and the same.
These are MOSTLY characteristics of the MUSIC and RECORDING and the are different. I'd say a loud snare drum every measure is microdynamics. A song that's quiet through the 1st half and loud for the 2nd half if macrodynamics. And there's everything in-between... Dynamics are complicated and not easily measured or described.

When I want to be more precise, I use the term "dynamic range" to describe the system (hardware & software). i.e. CDs have more dynamic range than vinyl records where the dynamic range is limited by noise. For the music/program material I like to say "dynamic contrast". Classical music tends to have more dynamic contrast that rock/popular music.

I was wondering if there is a measurement that predicts how "dynamic" a speaker would be. Impulse response?
There are. You'd have to make a recording and then you can compare the Crest Factor or the UBU R128 Loudness range (LRA) is based on loudness perception so LRA should be better. dpMeter is a free plug-in that can measure LUFS (loudness) and LRA and it works in Audacity (free). Or, GoldWave ($50 after free trial) can do the measurements. I think dpMeter can also measure crest factor or with a couple of "tricks" and standard Audacity plug-ins, Audacity can give you the peak & RMS levels so you can calculate it.

Also, is there such a thing as DSP which would improve the dynamics of a system by exaggerating the contrast between voltage spikes in the signal?
There is! But, it's kind-of rare and I don't know how easy it is to program into something like the miniDSP. Dynamic expansion, and there is a variation called "peak unlimiting".

In audio production about the only time expansion is used with a noise gate. A noise gate is downward expansion where the sound is reduced or killed completely during quiet parts (when presumably there is only background noise).

I assume you know about dynamic* compression. Dynamic compression reduces the dynamic range (or "dynamic contrast") by making the loud parts quieter or the quiet parts louder. Limiting is a kind of fast-compression, rounding-down the peaks. Compression is used (or over-used) on almost every (pro) recording/production. In practice compression and limiting are used to reduce the loud parts and push-down the peaks. Then make-up gain is used to make everything louder. These are the main weapons in the Loudness War.

The easiest way to play around with expansion is to download the trial version of GoldWave (an audio editor). It has a built-in compressor/expander effect. But there is one screwy thing about it... They don't use the definitions of compression & expansion correctly... The presets ARE correct so you can start with Boost Loud Parts or Reduce Quiet Parts presets and then you can play-around with the settings from there. What it really does is: When you choose "compressor" it's working above the threshold and when you choose "expander" it's working below the threshold. You'll notice that with the presets.... Boost Loud Parts will incorrectly show as compression but Reduce Quiet Parts will be correctly shown as expansion.

Of course, if you boost the peaks you'll usually have t reduce the overall level to prevent clipping.

You can also find 3rd-party plug-ins for Audacity, but I don't know of one to recommend.

Both GoldWave and Audacity are audio file editors. They don't work in real-time.

You can't "properly" reverse Loudness War compression and limiting. :( With limiting it's impossible to know the original peak. With compression there are several unknown settings and there is multi-band compression which compresses different frequency bands separately/differently, and often the individual tracks are compressed/limited before mixing, and then the mix is compressed/limited again. Too many unknowns and irreversible processes.


* Don't confuse this with file compression like MP3. Some people think MP3 compresses the dynamics but it doesn't. In fact, it changes the wave shape making some peaks higher and some lower. It doesn't change the sound of the dynamics but some people measure "dynamics" by using the Crest Factor and the MP3 will measure better than the uncompressed original! (And MP3 has actually higher dynamic range capability than CD.) Something similar happens with vinyl so if you use crest factor, the vinyl will measure more dynamic than the CD even if they were made from the same master.
 

tinnitus

Active Member
Joined
Jan 26, 2021
Messages
144
Likes
199
Location
Germany
Hi, Keith what DSP Software are you using for Filtergeneration ?
Have you measurements from your LP?
 

Kvalsvoll

Addicted to Fun and Learning
Audio Company
Joined
Apr 25, 2019
Messages
878
Likes
1,643
Location
Norway
The reason I ask is because my system is a hybrid one: horns for mids and treble, and conventional drivers for the sub (20 - 80Hz) and bass (80 - 600Hz). There is an obvious disconnect between the conventional drivers and the horns, and I can tell you it is NOT the frequency response, NOT amplifier headroom, and NOT distortion, and I have all the measurements to prove it. Well, not with me right now because I am in a hotel room. The conventional drivers just sound ... less dynamic, and I can not explain why using measurements. So please tell me where to look.
Correct. There is a difference, and in your observation you also have the answer - install horns for bass and low mid, and enjoy full-range dynamic sound.

As pointed out by several others here already, this has to do with the radiation pattern. And if you enjoy live volume levels, never underestimate the need for capacity. At low frequencies (<100hz) it is mostly about capacity, because bass couples to the surfaces close to the bass-system and sort of creates a large radiator.
 

hex168

Senior Member
Joined
May 29, 2020
Messages
396
Likes
338
While browsing Musicanddesign (recommended!), I found this fascinating analysis of the potential for thermal compression in hifi drivers:
 

Ciobi69

Senior Member
Joined
Jan 27, 2022
Messages
401
Likes
260
Does an 86db driver vs an 96db driver for the example able to deliver the same sensation of Dynamics? I feel like the 86 db driver needs a ton of watts and even then they feel like they don't compare, i know they less sensitive one needs a lot of watts, but at the same spl it seems different, could i be dreaming?
 
Joined
Feb 14, 2023
Messages
52
Likes
128
I believe our perception of dynamics to be related to dispersion somehow. If we take 2 speakers that have very little/no compression and low distortion, provide them with ample power... the one with a narrower dispersion will seem more dynamic. What I find interesting is the in @amirm listening tests, he commented specifically on dynamics in 4 JBL models.... 708p, array 880, 4439, and srx835p. Al of these I believe have well controlled dispersion but under 50 degrees. In a completely subjective, unscientific observation.. I swapped out a Revel C208 with a JBL 708P and the perception of dynamics in movies is very noticeable. C208 is around 80 degrees and the 708p is around 50 degrees.
It's interesting that @amirm is noting high subjective dynamics on the JBL horn speakers reveiwed but there is no measure of dynamics in speaker evaluations? I was thinking rise time and decay that @ppataki notes may be a measure, directivity of speakers like horns might help us understand dynamic measures, or speaker efficiency, or maximum SPL that can be reached by a speaker to understand loudness headroom. These could be measures that help us understand dynamics.

Some speakers seem boring to me, meaning not dynamic, they distort when turned up and seem to strain, while others seem effortless and dynamic. The measures here at ASR perhaps indicate that dynamics is something that is hard to measure, or cannot be measured, or that speaker dynamics does not matter you just buy lots of cheap clean power.

Still, this thread on measuring dynamics seems loose and still open. I wonder what experts like @ctrl and @Bjorn think when considering how to evaluate a "dynamic" speaker.
 

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,219
Likes
1,360
Location
Budapest
It's interesting that @amirm is noting high subjective dynamics on the JBL horn speakers reveiwed but there is no measure of dynamics in speaker evaluations? I was thinking rise time and decay that @ppataki notes may be a measure, directivity of speakers like horns might help us understand dynamic measures, or speaker efficiency, or maximum SPL that can be reached by a speaker to understand loudness headroom. These could be measures that help us understand dynamics.

Some speakers seem boring to me, meaning not dynamic, they distort when turned up and seem to strain, while others seem effortless and dynamic. The measures here at ASR perhaps indicate that dynamics is something that is hard to measure, or cannot be measured, or that speaker dynamics does not matter you just buy lots of cheap clean power.

Still, this thread on measuring dynamics seems loose and still open. I wonder what experts like @ctrl and @Bjorn think when considering how to evaluate a "dynamic" speaker.
To be honest I no longer believe that the Wavelet diagram will tell much about the dynamics (or maybe we shall call it transient response?) of a speaker
It is definitely a diagram to look at since it will give you hints about the energy delay and the nature of your filters (minimum phase vs linear phase or post-ringing vs pre-ringing)
But when it comes to dynamics, here is my totally subjective 2 cents:
- I find higher sensitivity speakers to sound more dynamic (I know there is zero science behind that but still this is my experience from the last 15 years)
- I find bigger cone surface area to sound more dynamic (does not matter if there is one bigger driver or multiple smaller drivers; I am just referring to the sum of the surface area) - this point actually might have something to do with large surface area --> lower excursion needed to achieve the same SPL --> lower excursion = lower distortion usually
- I find flat frequency response to sound more dynamic (I mean that a loudspeaker with a jugged FR will sound way less dynamic vs the very same speaker with full DSP-d ruler flat frequency response combined with the below point)
- If I apply an adequate low-shelf boosting filter to that ruler flat FR I will get the feeling of having a more dynamic system
- Last but not least: having a good measuring DAC and an amplifier with enough headroom also helps a lot
 
Joined
Feb 14, 2023
Messages
52
Likes
128
To be honest I no longer believe that the Wavelet diagram will tell much about the dynamics (or maybe we shall call it transient response?) of a speaker
It is definitely a diagram to look at since it will give you hints about the energy delay and the nature of your filters (minimum phase vs linear phase or post-ringing vs pre-ringing)
But when it comes to dynamics, here is my totally subjective 2 cents:
- I find higher sensitivity speakers to sound more dynamic (I know there is zero science behind that but still this is my experience from the last 15 years)
- I find bigger cone surface area to sound more dynamic (does not matter if there is one bigger driver or multiple smaller drivers; I am just referring to the sum of the surface area) - this point actually might have something to do with large surface area --> lower excursion needed to achieve the same SPL --> lower excursion = lower distortion usually
- I find flat frequency response to sound more dynamic (I mean that a loudspeaker with a jugged FR will sound way less dynamic vs the very same speaker with full DSP-d ruler flat frequency response combined with the below point)
- If I apply an adequate low-shelf boosting filter to that ruler flat FR I will get the feeling of having a more dynamic system
- Last but not least: having a good measuring DAC and an amplifier with enough headroom also helps a lot
This is a key issue and I would expect the experts at ASR would have some way measuring. Perhaps something as simple as max SPL? I high max SPL coupled with low distortion may give you an idea of dynamics. I agree with your experience, 15 years that point toward high sensitivity being important. Zero science maybe as you say for high efficiency, but maybe the science needs to try and understand this common observation.

I would also think just having more surface area means less cone movement, less distortion, faster rise time and faster decay, which makes for better dynamics. There is some discussion of horns here at ASR, and @amirm there is a subjective report that horn loaded JBL's have a certain excitement to the sound. I would think that Harman corporation could develop some measure of what makes a speaker dynamic or dynamic headroom.

Purifi may be addressing this be reducing long through low distortion bass, meaning the move enough air to make stereo based music more exciting. The drivers have subjective high dynamics because of low distortion at higher SPL. So it seems like it should be measurable since a lot of effort went into a Purifi driver for example to create deep bass in a smaller format.

Or. We define a rock concert as 110 db at 3 meters from your speakers or we call this a Saturday night party. That means sound of 110 db in the mid to far field. You could define a speaker scoring how well it can "rock" by this definition. If you have loud parties, or you like to rock out you have a better idea of what you are shopping for.
 
Last edited:

ozzy9832001

Senior Member
Joined
Feb 19, 2023
Messages
404
Likes
257
To be honest I no longer believe that the Wavelet diagram will tell much about the dynamics (or maybe we shall call it transient response?) of a speaker
It is definitely a diagram to look at since it will give you hints about the energy delay and the nature of your filters (minimum phase vs linear phase or post-ringing vs pre-ringing)
But when it comes to dynamics, here is my totally subjective 2 cents:
- I find higher sensitivity speakers to sound more dynamic (I know there is zero science behind that but still this is my experience from the last 15 years)
- I find bigger cone surface area to sound more dynamic (does not matter if there is one bigger driver or multiple smaller drivers; I am just referring to the sum of the surface area) - this point actually might have something to do with large surface area --> lower excursion needed to achieve the same SPL --> lower excursion = lower distortion usually
- I find flat frequency response to sound more dynamic (I mean that a loudspeaker with a jugged FR will sound way less dynamic vs the very same speaker with full DSP-d ruler flat frequency response combined with the below point)
- If I apply an adequate low-shelf boosting filter to that ruler flat FR I will get the feeling of having a more dynamic system
- Last but not least: having a good measuring DAC and an amplifier with enough headroom also helps a lot

I agree with what you say about cone size, but I get mocked when I say that. And it probably does have to do with distortion. I also agree that the flattest bass seems the most dynamic. I usually try my hardest to get up to 500hz as flat as possible. Rest doesn't seem as important.
 
Top Bottom