Perceived volume vs. mathematical volume.

[fdeck]

I agree. Nice way of showing it.

Thanks. By the way, were you ever tempted to plug your bass into one of those industrial amps?

 

[DavePlaysBass]

Thanks. By the way, were you ever tempted to plug your bass into one of those industrial amps?

I keep saying one of these days. It was the dream of making the world's biggest bass amp that led me to EE school back in the day. I have yet to realize that dream.

I am an analog guy and could modify an industrial class D power stage to suite speaker loads pretty easily, especially if we are talking fans and room temperature. It is the audio signal processing that I am not aware of many of the tips and tricks. Getting the music signal modulated into a pulse width modulated drive signal of course can introduce distortion. I have scanned a number of articles talking about increasing fidelity with switching power amps, but they usually are talking fancy DSP algorithms. In my work enviroment, I get to define what the DSP does, but someone else makes it tick.

I wish I could get some of my project teams at work focused on some of these bass amplification problems.

Dave

 

[agedhorse]

Let me expand on this discussion, it's very important stuff to us engineering types, it applies to all amplifiers, and relates directly to what you guys hear/feel.

Why does an amp appear to sound bigger that another similarly rated amplifier. There are 4 different common reasons and those in the (real) pro audio industry have known this for many years but it seems to escape a discouragingly large segment of the market.

1. Energy storage allows for higher dynamic power ride through under transient conditions, minimizing the effects of DYNAMIC sag regardless of continuous sag. Under most conditions, you HEAR dynamic power but the numbers tend to indicate continuous power. In previous diagrams, the power supply series resistence was placed after the "ideal" battery but in fact, it's actually more complicated than this. The power supply has a series impedance associated with it that is responsible for the long term (continuous) voltage sag as the current increases (load impedance decreases). In actuality, there are energy storage capacitors that serve (among other things) to hold this voltage up during transient loads by acting as an energy resivour. The charging (recovery) time is a function of the transformer source impedance and the amp load in conjunction with the capacitors determins the discharging time constant. The larger the capacitors the longer the discharge time constant but without considering the source impedance and the load dynamics the point of diminishing returns is met at some point. Knowing how this mechanism works is fundamental (IMO) to developing an amp that optimizes these parameters to give a bigger, more dynamic sound with good feel but still keep the package size/weight practical. This stuff all affects how the amp feels as it reaches maximum power.

2. Reactive loads... a speaker is not a resistor. It's a complex load consisting of incuctance, capacitance and resistance. If you were to analyze the load line of a real speaker, it shows a lot more potential for amplifier destruction compared to just driving resistors. Without adequate silicon in the output stage and the necessary drive electronics, the protection schemes (specifically VI limiters) must prematurely limit the outout under these conditions. It's common for manufacturers looking to save money to skimp on this part of the design, essentially crippling their own product and maybe not even knowing it. Also, VI limiting (not to be confused with signal limiting) is particularly nasty since the voltage clips referenced to the current and any phase angle between the two (which there is plenty) causes the clipping to be taken out of the sides ofthe waveform resulting in "yuck".

3. Output signal limiters and how an amp overloads has a tremendous effect in how an amplifier reacts approaching the clipping point. An amp the limits gracefully, overloads more smoothly and doesn't stick or chatter can be driven harder into clip without unpleasant artifacts. For some reason, designers seem stuck with fixating on how the amp performs prior to clipping but then all hell breaks loose as the amp hits the clipping point and becomes unuseable.

4. Preamp voicing and general small signal processing affects how loud an amp appears, and even more importantly how it feels in context with the instrument itself. If everything is complimentary and working together, it will feel unhindered rather than crippled.

 

[billfitzmaurice]

I think that this diagram approximately illustrates the situation.

Not really, as a loudspeaker is not a resistive load. It's primarily an inductive load, although there are both resisitive and capacitive components. The main problem with the diagram isn't the fact that the load shown is resistive though, it's that it's shown at a set value. Loudspeakers do not have a set load value, only a nominal average load value.

 

[fdeck]

Not really, as a loudspeaker is not a resistive load. It's primarily an inductive load, although there are both resisitive and capacitive components. The main problem with the diagram isn't the fact that the load shown is resistive though, it's that it's shown at a set value. Loudspeakers do not have a set load value, only a nominal average load value.

Yes that's definitely true. All the more reason to start talking about voltage instead of power.

This page gives a recipe for modeling a speaker load.

Link Removed

 

[fdeck]

I keep saying one of these days. It was the dream of making the world's biggest bass amp that led me to EE school back in the day. I have yet to realize that dream.

I am an analog guy and could modify an industrial class D power stage to suite speaker loads pretty easily, especially if we are talking fans and room temperature. It is the audio signal processing that I am not aware of many of the tips and tricks. Getting the music signal modulated into a pulse width modulated drive signal of course can introduce distortion. I have scanned a number of articles talking about increasing fidelity with switching power amps, but they usually are talking fancy DSP algorithms. In my work enviroment, I get to define what the DSP does, but someone else makes it tick.

I wish I could get some of my project teams at work focused on some of these bass amplification problems.

Dave

What a coincidence. My interest in bass gear fueled my interest in electronics as a kid. But I went to a college with no engineering school, so the closest thing was physics. Thus in an indirect way, bass amplification drove me into my present career too.

Now I work for a company that makes Fourier Transform spectrometers, and we have some guys who know DSP frontwards and backwards. In fact, we were one of the early makers of DSP's, when it still involved racks of boards stuffed with TTL chips.

 

[DavePlaysBass]

Not really, as a loudspeaker is not a resistive load. It's primarily an inductive load, although there are both resisitive and capacitive components. The main problem with the diagram isn't the fact that the load shown is resistive though, it's that it's shown at a set value. Loudspeakers do not have a set load value, only a nominal average load value.

The resisitive load represents the work done. Work in this translates to air movement. While reactive properties effect dynmics, they do not represent work, air movement, or translation of electric energy into air movement. At the end of the day, the air movement comes down to the effective voltage developed across the effective resistance. The effective resistance hear meaning the energy change from electron movement to SPL.

In terms of answering the OP question, this model is adequate. In terms of exploring every intricacy of power amp behavior, probably not.

BTW, I might guess that your expereience with speaker testing may be with amps that are a long way from clipping. In order to get good data on your speakers, you do not want non linear amp effects getting in your way. Hence, you may not see a voltage change of signicance at the operating points you are asking the amps to operate at. Effectively you want huge headroom when doing speaker tests so this does not mess with your results (common engineering practice, let's not change everything at once). But if you approach the problem from how hard can I push the amp (which is a very different question than how hard can I push the speaker), your test methods and observations may be different. Of course up until you clip the amp at which case you are testing both.

 

[DavePlaysBass]

What a coincidence. My interest in bass gear fueled my interest in electronics as a kid. But I went to a college with no engineering school, so the closest thing was physics. Thus in an indirect way, bass amplification drove me into my present career too.

Now I work for a company that makes Fourier Transform spectrometers, and we have some guys who know DSP frontwards and backwards. In fact, we were one of the early makers of DSP's, when it still involved racks of boards stuffed with TTL chips.

I would love to have a many beer conversation on preamp design if you are ever vacationing to the great state of CO. In fact that goes for anyone out there. Even if we disagree ;-).

Dave

 

[Eublet]

Ken, I've seen you post comments about your LMII being as loud and louder than some other heads much more powerful. I think Liam's comments are in the same category as yours. And depending on the head, I think twice as powerful is probably a fair estimate. Many heads have high specs that don't play out when your run the head with a certain voice in mind. VI limiting starts kicking in and affecting the overall volume when you're pushing hard. I know that there are a lot of players around my area using GK Backline, SWR Working Series, and Ampeg SVT3's just to name 3, and in a watt to watt comparison, Genz blows their stuff out of the water in terms of volume, transient attack, quick reaction, etc. It just doesn't compare. The same could be said by other quality heads compared to those, not just Genz.

The amps I just named are among the types of heads that I see people gigging with, especially college guys who have to stretch their dollars.

As for the whole 2 ohm thing, I don't understand why it's such a problem that a head that is designed to run optimally at 2 ohms would be branded as such in the name. Not to stir it up again, but perhaps the problem is everyone else whose heads aren't designed as such. Besides, on my GBE 1200, I lose 200 watts going down to 4 ohms. On the 750, you lose 125. Not a big deal, and GBE 750 has a much nicer ring than GBE 625.

 

[Eublet]

I'll add, the cabinet you use seems to make a difference in making volume judgements. I recently did a comparison between my Neo-pak and an SVT3 pro through an 8 ohm 810 fridge. The svtIII is 300 watts, the Neo-pak is 225. There was no comparison. The Neo-pak put it to shame, and it has a first-gen SMPS design. The GBE 600 would CRUSH the SVT III pro.

However, I've run both these heads through my NeoX 112. It's about impossible to notice any volume difference between the two heads as the cabinet size and single driver normalizes things a good bit.

 

[agedhorse]

There is no issue in designing a head to run all day long at 2 ohms if the designer knows what they are doing and the amp is manufactured accordingly. QSC, Crown and Crest all have excellent products that fit this catagory don't they?

 

[fdeck]

I know that there are a lot of players around my area using GK Backline, SWR Working Series, and Ampeg SVT3's just to name 3, and in a watt to watt comparison, Genz blows their stuff out of the water in terms of volume, transient attack, quick reaction, etc. It just doesn't compare. The same could be said by other quality heads compared to those, not just Genz.

I don't doubt that there are differences in dynamic reserve among amps. But differences in perceived volume can have other causes as well. I would want to know:

1. How much actual reserve is there, in dB? And also in terms of technical knowledge, how much of an audible effect does this have, compared to (for instance) letting the amp clip the first one or two cycles of each note?

2. Are the amps all equalized the same, and do they use the same speakers? I suspect that equalization and speaker response play a dominant role in perceived volume, even when we apply terms like "quick reaction" to what we hear.

 

[Eublet]

I don't doubt that there are differences in dynamic reserve among amps. But differences in perceived volume can have other causes as well. I would want to know:

1. How much actual reserve is there, in dB? And also in terms of technical knowledge, how much of an audible effect does this have, compared to (for instance) letting the amp clip the first one or two cycles of each note?

2. Are the amps all equalized the same, and do they use the same speakers? I suspect that equalization and speaker response play a dominant role in perceived volume, even when we apply terms like "quick reaction" to what we hear.

I can't answer your first question. As for the second one, EQ is tough to call, but I EQ an amp until I get as close as possible to the tone I"m after. I totally agree that alot of things can contribute to perceived volume, and EQ has a big part in that. But by and large, it's just easy to tell with the heads that I mentioned that they just don't hang when I compare them. It's an easy thing to hear. With more comparitively powered items that perform more identically, that is difficult I agree.

 

[agedhorse]

A cycle or two is not likely to be noticed as long as there are no artifacts like stickingm chattering or bursts of oscillation. The problem isn't so much a couple of cycles (20mSec at 100 Hz) but how about something along the lines of 200mSec of clipping? Add this to the other effects I mentioned and there are a lot of practical reasons why some amps sound bigger than others with similar numerical ratings.

 

[fdeck]

A cycle or two is not likely to be noticed as long as there are no artifacts like stickingm chattering or bursts of oscillation. The problem isn't so much a couple of cycles (20mSec at 100 Hz) but how about something along the lines of 200mSec of clipping? Add this to the other effects I mentioned and there are a lot of practical reasons why some amps sound bigger than others with similar numerical ratings.

Considering that I can play five notes in a second, I would agree that 200 ms of clipping would be objectionable. And like you say, how the amp actually behaves when it clips is important. When I work on circuits, I always pay close attention to clipping behavior.

 

[agedhorse]

When I work on circuits, I always pay close attention to clipping behavior.

Something simulators don't do a particularly good job of identifying in many cases. This is where experience makes a difference. It's easy to design the linear stuff but it's the non-linear stuff in a linear design (like an amp) that will get you every time.

 

[fdeck]

Something simulators don't do a particularly good job of identifying in many cases. This is where experience makes a difference. It's easy to design the linear stuff but it's the non-linear stuff in a linear design (like an amp) that will get you every time.

Discovering the difference between simulator and reality can be a humbling experience.

 

[agedhorse]

Discovering the difference between simulator and reality can be a humbling experience.

It's one reason why some amps appear to produce less real world power than they should. The simulator model must be accurate and that's pretty tough to do with a dynamic transducer that you may not know which particular model is being used with an amp.

For other less technical forumites, this is a bit of what goes on when designing any amplifier product. We make assumptions, fine tune the results and then based on experience adjust for real world conditions. It's not easy because the conditions themselves can change and the unknown changes must be factored in also.

 

[Fretlessboy]

Maybe that was his real world finding on the particular head he owns. JM 2 cents. Sometimes the prejudice is against they do not endorse.

We've had this discussion before I was mainly commenting on the 600's 425 watts somehow sounding like more than 425 watts. It doesn't to me.... that's not necessarily a bad thing (it's a very nice moderately powered head to my ear), but an endorsee stating that it sounds louder than 'most amps he's heard with 'double the wattage' was misleading and ridiculous IMO and IME.

Just trying to provide some 'real world' view based on not being aligned with any company... just what my ears tell me with no politics, friendships, or 'endorsements' impacting what I might be hearing.

 

[svtb15]

I used an ampeg svt3 pro for years...And i have played my friends Genz gbe600 with my same cabs, and the Genz 425 watts appears to be MUCH LOUDER than the ampegs 450watts each at 4 ohms... Much louder and fuller...