Getting your sound, science or art?

[JPJ]

6 hours ago, Phil Starr said:

 

 

.

We had hoped to measure a lot more amplifiers. There were plenty there at the bass bash but we simply ran out of time. The result for the Gnome was so unexpected that we spent precious time checking everything over. I had to set up and run a speaker shootout in the main hall and couldn't act as gopher for John.

 

Anyway it was an interesting experiment I think we'll repeat and try to grab more data on other amps. There is free software to do this so if you have a simple two in and two out interface you should be able to do this at home so maybe a few of us could work together and build a database

I’m happy to measure my collection if someone can provide a ‘how to’ guide and a link to the necessary software?

My collection includes:

1) SWR SM1500

2) SWR Marcus Miller preamp

3) SWR Grand Prix preamp (Green face)

4) Ashdown Head of Doom

5) Ashdown ABM NEO 400

6) TC Electronics BG250

 

P.S. don’t show my wife this list 😂

[Vin Venal]

My tone is 100% mistakes and doing everything the wrong way.

 

As is my entire life basically. 😅

 

I like the process of figuring everything out myself, and coming up with my own ways of doing stuff.

[Phil Starr]

7 hours ago, JPJ said:

I’m happy to measure my collection if someone can provide a ‘how to’ guide and a link to the necessary software?

My collection includes:

1) SWR SM1500

2) SWR Marcus Miller preamp

3) SWR Grand Prix preamp (Green face)

4) Ashdown Head of Doom

5) Ashdown ABM NEO 400

6) TC Electronics BG250

 

P.S. don’t show my wife this list 😂

OK it is starting to look like a plan, it might take a while to sort out the easiest way to do it and write a how to guide but worth doing. I'll have a chat with John when he gets over his covid and I'll have a good look myself. That might take a while as we have a couple of new band members to rehearse in but it could be an interesting project.

[3below]

53 minutes ago, Phil Starr said:

OK it is starting to look like a plan, it might take a while to sort out the easiest way to do it and write a how to guide but worth doing. I'll have a chat with John when he gets over his covid and I'll have a good look myself. That might take a while as we have a couple of new band members to rehearse in but it could be an interesting project.

 

A guide and a reasonably standardised way of doing this (physicist/engineer here) would be really useful.  In addition to amp heads I would like to contribute measurements of some pedals in my posession e.g Tech21 Leeds, VT, Joyo American etc. (if this is of interest).  In semi jest  we can then subject all the results to AI / machine learning and arrive at the universal eq emulator ranges.

Edited October 15, 2022 by 3below

[stevie]

23 hours ago, Chienmortbb said:

I will do that next time I visit @stevie. It won't be soon as I have Covid 😁

 

By way of explanation, it was my Ashdown that John tested at the Bash. I occasionally use the RM500 with the tone controls switched out. I can't say I've noticed any difference between that and having the controls at 12 o'clock.

[Phil Starr]

13 hours ago, 3below said:

 

A guide and a reasonably standardised way of doing this (physicist/engineer here) would be really useful.  In addition to amp heads I would like to contribute measurements of some pedals in my posession e.g Tech21 Leeds, VT, Joyo American etc. (if this is of interest).  In semi jest  we can then subject all the results to AI / machine learning and arrive at the universal eq emulator ranges.

I'm looking at using REW which is free and pretty powerful. It's also fairly well supported and there are plenty of instructional videos on You Tube. Any 2 in/2 out interface should work and I'm probably going to use my USB mixing desk. REW lets you do a calibration plot so if the interface does have any artefacts they can be allowed for. There might be a problem with connecting some BTL (bridged) amps to earth but most DI boxes should let you fix that. Currently John is doing his measurements with an oscilloscope and other software. I'm not sure if that is an issue, if they lead to different results then we can investigate.

 

The only other issue I can think of is all agreeing a protocol so data can be collected and compared. For measurement I think 20-20kHz measurements at low power makes sense and REW allows you to centre results at 0db. Displaying results as a graph is useful but a table summarising results would also be good as a quick reference (-3db points, frequency range, rate of roll off, centre and size of any frequency anomalies) Ideally it would be good to make the data files themselves available to anyone who wants them. I've no idea about how they are stored.

 

I'm no expert on this so any help/suggestions are very welcome. We might need to take this across to another thread if this takes off

 

[SamIAm]

21 minutes ago, Phil Starr said:

I'm looking at using REW which is free and pretty powerful

OMG that looks amazing!

 

S'manth x

[Chienmortbb]

1 hour ago, Phil Starr said:

I'm looking at using REW which is free and pretty powerful. It's also fairly well supported and there are plenty of instructional videos on You Tube. Any 2 in/2 out interface should work and I'm probably going to use my USB mixing desk. REW lets you do a calibration plot so if the interface does have any artefacts they can be allowed for. There might be a problem with connecting some BTL (bridged) amps to earth but most DI boxes should let you fix that. Currently John is doing his measurements with an oscilloscope and other software. I'm not sure if that is an issue, if they lead to different results then we can investigate.

 

The only other issue I can think of is all agreeing a protocol so data can be collected and compared. For measurement I think 20-20kHz measurements at low power makes sense and REW allows you to centre results at 0db. Displaying results as a graph is useful but a table summarising results would also be good as a quick reference (-3db points, frequency range, rate of roll off, centre and size of any frequency anomalies) Ideally it would be good to make the data files themselves available to anyone who wants them. I've no idea about how they are stored.

 

I'm no expert on this so any help/suggestions are very welcome. We might need to take this across to another thread if this takes off

 

REW may be the way to go. I use software allied to the Picotech range of PC oscilloscopes. This is versatile but expensive for just viewing the frequency responses of amplifiers. 

[Phil Starr]

Back on track, I did warn you I like a graph. This next one is probably familiar and I think the 'science' reason behind the 'smiley face' eq so many amps have when the controls are all set to noon. Each curve measures how much energy you need at each frequency to make sounds equally loud. Effectively it draws where you'd need to set the graphic to get a flat sound. You'll notice there isn't  watt in sight, sound is measured in decibels.  The easiest line to explain is the 'threshold' line. In this case 0db is set at the quietist sound you can hear with all the other decibels set relative to that. Originally the first measurement was taken at 1kHz so that is where zero was set, a bit like setting 0 as the lowest temperature they could get at the time at the bottom of the Fahrenheit scale.

 

So this is why we like a smiley face eq. If you look at all the lines they show a broad dip in the middle with lots of extra volume  to get you to hear the bass and treble as loud. And let's face it when you only had two tone controls on the radiogram you only ever boosted you never cut. Looking carefully to hear 40Hz (bottom E) at all you need 50db of bass boost. That's 100,000x the power! If you look at the other extreme at 100Phon (about the level of sound where the drummer sits in a rock band) you only need 20db of boost. The bass sounds 30db louder and only needs 1000x the power. Subjectively it will sound 8x louder relative to that 1kHz mid.

 

So turning up the volume of anything amplified makes it sound like you've boosted the bass and treble and here's the punchline: boosting the bass and treble at any frequency makes an amp sound like a bigger more impressive amp. I wonder why anyone would want to do that in a showroom situation

 

Edited October 19, 2022 by Phil Starr

[Merton]

Great post @Phil Starr - clear and succinct 👍🏻

[3below]

1 hour ago, Phil Starr said:

.... So turning up the volume of anything amplified makes it sound like you've boosted the bass and treble and here's the punchline: boosting the bass and treble at any frequency makes an amp sound like a bigger more impressive amp. I wonder why anyone would want to do that in a showroom situation

 

+1 ^ This.  

Next we will be adjusting bass / treble / eq in gig or recording situations to make it sound better  

[Phil Starr]

3 minutes ago, 3below said:

+1 ^ This.  

Next we will be adjusting bass / treble / eq in gig or recording situations to make it sound better  

Here's the thing though, and this has only just occurred to me re-reading my own post.  Turning up the bass and treble at means that you are effectively turning the mids down relative to the two extremes. The equal loudness curves show that just turning up the volume sounds like you are turning down the mid balance. Ever had that experience of a band being really loud but somehow not being able to pick yourself out of a mix? Well the mids are what you need to hear yourself and a nasal mid balanced sound always sounds awful at home but a better fit once you are mixed in with a band. Equal loudness could be most of the reason for that. Especially if everyone else is turning up at the same time.

 

That's why I love these discussions, I'll be mulling over that all day, and i now feel stupid for not having thought about it before

[Chienmortbb]

I am puzzled (could be covid or my innate stupidity). Equal loudness curves show the way we hear things. So you, or at least I, would expect that a truly flat response in a PA would be the best option, then set the mix to sound good. So why do people recommend boosting the low end by 6-20dB? Is it just that the preponderance of bass heavy, computerised music in recent years.

 

On a second point, the smiley face EQ or Shape with bass and treble boosted is much like the old Loudness controls on Hi-fi amps. These were there because the equal loudness curves change with Volume/SPL. Does this explain why smiley face EQ sounds great at home but rubbish on stage?

 

There is an excellent piece on this, it's quite short, Is Sound an Illusion? © 2009, Les Acres, Rod Elliott

Edited October 19, 2022 by Chienmortbb

[Downunderwonder]

7 minutes ago, Chienmortbb said:

Does this explain why smiley face EQ sounds great at home but rubbish on stage?

Not really. At home alone there are no competing sounds so your relatively low volume smiley sounds are complete and wonderously bassy. Throw in the band and the tonal information in the bass mids is smothered.

[Chienmortbb]

10 minutes ago, Downunderwonder said:

Not really. At home alone there are no competing sounds so your relatively low volume smiley sounds are complete and wonderously bassy. Throw in the band and the tonal information in the bass mids is smothered.

My point could have been made better, but what I meant was that at practice volumes, our ears do not hear either the low or high-end very well.

[Phil Starr]

26 minutes ago, Chienmortbb said:

I am puzzled (could be covid or my innate stupidity). Equal loudness curves show the way we hear things. So you, or at least I, would expect that a truly flat response in a PA would be the best option, then set the mix to sound good. So why do people recommend boosting the low end by 6-20dB? Is it just that the preponderance of bass heavy, computerised music in recent years.

 

Good question. I've also come across situations where the PA is set up with a smiley face on the graphic as a starting point. Please don't listen to this advice. I came across it only a few weeks ago when an inexperienced guy hired in a PA where we were playing. We had a BBC engineer on the monitor mix and the hirer on the FOH. Result was the on-stage sound was glorious and FOH a bassy mush throughout the whole day. 

 

So you want your PA to sound like what you are feeding in. The graphic is only there to adjust for room acoustics IMO. If you want the bass or kick louder turn them up in the mix. If the bass sounds thin then adjust it with the channel controls so it doesn't. Don't boost both if you only want to affect one of them.  The trouble is that a lot of these 'rules of thumb' date from the times when PA and backline were in their infancy. If your PA was composed of underpowered amps and WEM 4x10's and you were mixing with the bass coming through a guitarists 4x12 boosting the bass and treble made some sort of sense, though replacing drivers was routine as a result.

 

 

 

 

[3below]

1 hour ago, Phil Starr said:

..... Especially if everyone else is turning up at the same time......

The perennial problem, plus my drummer (not heavy handed) sometimes uses a Gretsch kit which is just "too loud".   

[JPJ]

2 minutes ago, 3below said:

The perennial problem, plus my drummer (not heavy handed) sometimes uses a Gretsch kit which is just "too loud".   

You know the old adage, if its too loud you’re too old 🤣

 

 

[Passinwind]

On 16/10/2022 at 04:01, Phil Starr said:

I'm looking at using REW which is free and pretty powerful. It's also fairly well supported and there are plenty of instructional videos on You Tube. Any 2 in/2 out interface should work and I'm probably going to use my USB mixing desk. REW lets you do a calibration plot so if the interface does have any artefacts they can be allowed for. There might be a problem with connecting some BTL (bridged) amps to earth but most DI boxes should let you fix that. Currently John is doing his measurements with an oscilloscope and other software. I'm not sure if that is an issue, if they lead to different results then we can investigate.

 

The only other issue I can think of is all agreeing a protocol so data can be collected and compared. For measurement I think 20-20kHz measurements at low power makes sense and REW allows you to centre results at 0db. Displaying results as a graph is useful but a table summarising results would also be good as a quick reference (-3db points, frequency range, rate of roll off, centre and size of any frequency anomalies) Ideally it would be good to make the data files themselves available to anyone who wants them. I've no idea about how they are stored.

 

I'm no expert on this so any help/suggestions are very welcome. We might need to take this across to another thread if this takes off

 

REW is pretty danged good IME and the price is right. I bought the full version of True RTA eons ago and so I mostly just use that because I like the memory storage format for A/B capabilties. One gotcha is that many commodity interfaces have non defeatable HPFs built into some or all of the input channels. Steinberg UR series, I'm looking at you! 😎

 

As far as your original post, I've been wrestling with this for several decades by now. I used to help out on installs in many different sorts of spaces and in churches I would rather hilariously often get drafted to read from the pulpit because I have a big voice and a decent feel for the behind the mic experience. Dead flat was pretty much never the goal or the end result, but it makes for a uniform reference at least!

 

I happen to be building some new preamp pedals this week and testing various HPF formats. By the time you factor in rolloffs in various amps and onboard preamp stages just spec'ing a net 4th order HPF at 30Hz is a lot less straightforward than one might expect. At some point taking the box to an open mic and letting half a dozen bassists play through it often teaches me more than a zillion RTAs and Spice models do. But of course once we can correlate all those things reliably we're way ahead of the game!

Edited October 19, 2022 by Passinwind

[Waddycall]

Thanks for this Phil - great thread 👍

[Phil Starr]

13 hours ago, Passinwind said:

As far as your original post, I've been wrestling with this for several decades by now.

This is my delight, the sweet spot of musical speaker design. Everything we hear goes through huge reprocessing in our brains and our response to sounds is so subjective. The language we use to describe sound is so imprecise yet our response to music so visceral. The engineering part is pretty simple but converting people's thoughts into a design spec and then seeing how that all works in a real life situation is what really fascinates me. I've had the pleasure of designing a few one off cabs for people here and it's been a real learning experience. I've recently had a chance to talk to @TheRevabout the problems of amplifying upright bass.  A fascinating new problem. This interface between music, science, woodwork and people is my happy place.

 

Great to hear from you

[Chienmortbb]

15 hours ago, Passinwind said:

Dead flat was pretty much never the goal or the end result, but it makes for a uniform reference at least!

As usual I have to wait for someone else to put my thoughts into words. You have to have a reference, a starting point. 
 

When I first worked as a young apprentice my first job out of training g school was testing RF amplifiers. The equipment and probes we had were far from flat but at the start of each session, each morning and after lunch we would calibrate the special scopes and probes.  
 

This told us what flat looked like and we had a reference. That was a big lesson for me. 
 

The next realisation in music is that it’s not All About the Bass, or at least not the fundamental. The next thing, the thing we learn at many gigs is that the room is the box that affects your sound more than the little Barefaced, GK, Ashdown or even LFSys (my current squeeze). However if your amp or cab has a baked in sound that is similar to the acoustic qualities of a room, you are in trouble. That is why I like flat. I then use the amp eq, bass eq and multi effects to get, “my sound” but in the end I go into the auditorium  and listen. 

Edited October 20, 2022 by Chienmortbb
Corrected the worst of the gobbledygook.

[agedhorse]

On 14/10/2022 at 23:28, Phil Starr said:

 

 

.

We had hoped to measure a lot more amplifiers. There were plenty there at the bass bash but we simply ran out of time. The result for the Gnome was so unexpected that we spent precious time checking everything over. I had to set up and run a speaker shootout in the main hall and couldn't act as gopher for John.

 

Anyway it was an interesting experiment I think we'll repeat and try to grab more data on other amps. There is free software to do this so if you have a simple two in and two out interface you should be able to do this at home so maybe a few of us could work together and build a database

Beware that it's MUCH easier said than done.

It's common for folks who do not do this all the time, with a setup that allows for anechoic measurements 1/2-space measurements (1/2-space is most commonly used for stage speakers because that's the closest approximation of the real world environment) to have errors of between 3 and 6dB SPL. Then they go proclaiming things that are simply not true because they have drawn conclusions based of incorrect and/or inaccurate data, which is unfair to everyone.

[agedhorse]

On 16/10/2022 at 04:01, Phil Starr said:

I'm looking at using REW which is free and pretty powerful. It's also fairly well supported and there are plenty of instructional videos on You Tube. Any 2 in/2 out interface should work and I'm probably going to use my USB mixing desk. REW lets you do a calibration plot so if the interface does have any artefacts they can be allowed for. There might be a problem with connecting some BTL (bridged) amps to earth but most DI boxes should let you fix that. Currently John is doing his measurements with an oscilloscope and other software. I'm not sure if that is an issue, if they lead to different results then we can investigate.

 

The only other issue I can think of is all agreeing a protocol so data can be collected and compared. For measurement I think 20-20kHz measurements at low power makes sense and REW allows you to centre results at 0db. Displaying results as a graph is useful but a table summarising results would also be good as a quick reference (-3db points, frequency range, rate of roll off, centre and size of any frequency anomalies) Ideally it would be good to make the data files themselves available to anyone who wants them. I've no idea about how they are stored.

 

I'm no expert on this so any help/suggestions are very welcome. We might need to take this across to another thread if this takes off

 

Yes, you want  to be VERY careful with bridged amps, and most class D amps are internally bridged. Most are not protected from a short between 1/2 of the output bridge and ground, this is not the normal way a fault occurs on a speaker out, the protection typically looks at either 1/2 of the bridge or (indirectly) the difference in current between the two halves of the bridge. Also, the level is (or can be) higher than many scopes can handle.

 

This is a common way folks who are new to testing, or are new to electronics, or do electronics as a hobby destroy their laptops or interface (depending on where the interface is located). The fault is large enough that it typically lets the magic smoke out pretty instantaneously and catastrophically.

 

I've been testing class D amps professionally for over 20 years, there are a lot of necessary tricks to get good results and data, especially power and THD data (depending on the modulation scheme and output reconstruction filter type). It's harder than it looks, especially if the numbers need to be accurate.

[Phil Starr]

5 hours ago, agedhorse said:

Beware that it's MUCH easier said than done.

It's common for folks who do not do this all the time, with a setup that allows for anechoic measurements 1/2-space measurements (1/2-space is most commonly used for stage speakers because that's the closest approximation of the real world environment) to have errors of between 3 and 6dB SPL. Then they go proclaiming things that are simply not true because they have drawn conclusions based of incorrect and/or inaccurate data, which is unfair to everyone.

We were measuring the output of the amplifiers only. No speakers were involved so none of these concerns would be applicable. I wonder if you are confusing these measurements with the speaker shootout I carried out at the same meet-up.