And yet all the answers I've got (though none seemingly to the actual question I asked), are implying I'm not allowed to make mistakes. I have to magically jump to all you guys level of expertise before you can even contemplate making a speaker, and I must aim to be perfect or its a waste of time.
Perfect I clearly dont want to make.
Member
Joined 2003
Make all the mistakes you want, don't let me stop you. You've asked for advice, the above post appears that you are surprised that the advice is to go down a path that results in minimal error in data which is where most "mistakes" originate from. No one is going to give you advice to design something poorly.
To be honest I had assumed you had already abandoned the DIY idea. It appeared that you only wanted to build a single set of speakers which was a rather ambitious omni-directional design for a first build, and were learning here that loudpseaker design is not as simple as plunking manufacturer data straight into the crossover design tool so the whole idea was a non-starter.
To be honest I had assumed you had already abandoned the DIY idea. It appeared that you only wanted to build a single set of speakers which was a rather ambitious omni-directional design for a first build, and were learning here that loudpseaker design is not as simple as plunking manufacturer data straight into the crossover design tool so the whole idea was a non-starter.
No such luck sorry. You cant get rid of me that easily.
I have never assumed loudspeaker building was simple, as you describe plunking in manufacturer data, and I'm well aware of the propper process, howerver you did link to "simulation without measurement" by Reid Towsley on how to do just that. I'm just tying to learn vituix
I'm assuming I was meant to follow this guide, and I have as close as I can see but I'm still unsure if I have it right.
My files are here All files for project.
Member
Joined 2003
I wrote that guide. It is meant to help you learn the tools, following a similar process to what would be done with your own measurements, and limitations of the process are spelled out at the start (and previous posts here). For real design work, you are meant to use a real measurement mic and 2-Channel measurement jig. ARTA manual and the measurement guides in the VituixCAD help is a good resource for that, and I have some documentation of my own as well on the jig if you’re interested.
I find the choice of words here a little dramatic. A "real" design work???
What does that even mean? Or other things not real, thus fake??
Are we "meant" to do things, is it a forced law?
Second of all, it implies a very discrete black-and-white situation.
Like there is nothing in between?
Which is just not even the case at all and it also depends on the context.
This all leads to confusion and frustration for people who try to understand it and learn from it.
A better approach is to explain what the shortcomings are and why.
"the data is just crap" doesn't explain anything at all.
So first of all, some issues with simulating responses from datasheets.
Fyi, the bafflestep simulation is in most cases fine, unless you're going for bizarre shaped cabinets.
When the shortcomings are known this automatically also means it's not entirely useless.
As long as we respect those issues and/or stay (far) away from it, those issues are non-existing for our purposes.
There are also techniques to kind of get a rough idea about these issues.
The first problem is something to really look out for.
It can sometimes be a challenge to figure out under what conditions manufactures measure their things.
Was it measured on an AES baffle, in a certain cabinet etc etc?
Also in many cases, the freq resp itself can be cherry picked, or contains to many reflections.
(See SB Acoustics for example).
A workaround here, is to find data that is more reliable.
Boxsim for example seems to have a decent (not perfect) correlation with actual data.
So one could for example also pick his components accordingly.
Phase information is only (very!) important with multiple speakers.
As soon as this phase information is not accurate anymore, the whole crossover filter is close to be meaningless.
(for the freq resp connection between the two, not for the speakers individually)
So when designing a fullrange speaker (so just one fullrange woofer), this is not important.
It's also not very important for subwoofers, or woofers that aren't being used up till the diffraction issues start.
For the last point, one can decide to make a baffle with as minimal diffraction issues as possible.
Another way is to just put the crossover before the point the diffraction issues start.
A couple of these things also count for 2-channel measurements.
This is only gonna be an issue* when there are more than one speakers involved.
Second to that, there are (many) workarounds.
Especially when active filters are being used, one can decide to measure on the fly.
Although this can be a very tedious process, it's doable.
In this case you measure with just a simple (USB) microphone, and change the filter parameters per speaker.
Next you measure both speakers at the same time**
Go back, rinse and repeat.
This also works with gated measurements.
Is this ideal? No!
Does it get you by? Yeah, kind of.
The point is to make people aware for what to look for.
Instead of just saying it can't be done and one can't make ANY good speakers with it.
Yes, without proper measurements and off-axis response with multiple axis, making a good professional sounding loudspeaker is a big challenge.
Impossible? Maybe, in my opinion likely (that it's not possible).
On the other hand, I know people who do almost everything by ear and on experience.
These people make fine sounding loudspeakers, yes there is definitely room for improvement.
Which leads to the last point. Some people are just simply not looking for the "best possible speaker" (whatever THAT even means).
But these people are looking for something between "randomly wild guessing" and having to put all the effort of getting and buying all this measurement equipment.
Again, as long as people are very well aware of the shortcomings, and most importantly, understand these issues and also clearly change their expectations to it.
I don't have an issue with it and still don't understand at all why it almost seems a taboo?
In some cases it sounds more like just a personal OCD thing than anything else to be very honest.
Makes me wonder if some also weight the ratio between tea and water to the exact micro gram and monitor the water temp closely to 0.1 degrees C ?
Other people just get the cheapest tea they can find and shove it in a mug and add some milk to it.
* Way back I did most of my measurements with just single channel, and the issues were to small to have any significance.
We are talking about WinXP era with a simple 16 bit USB soundcard and just a regular normal amp.
These days I don't know exactly how that works anymore with Win8/10 driver issues and delays.
But even here there are workarounds available. Again, it's not ideal.
** Important is to take at least two off-axis measurements as well btw.
So like 15 degrees and 30 degrees or something. Absolute bare minimum.
What does that even mean? Or other things not real, thus fake??
Are we "meant" to do things, is it a forced law?
Second of all, it implies a very discrete black-and-white situation.
Like there is nothing in between?
Which is just not even the case at all and it also depends on the context.
This all leads to confusion and frustration for people who try to understand it and learn from it.
A better approach is to explain what the shortcomings are and why.
"the data is just crap" doesn't explain anything at all.
So first of all, some issues with simulating responses from datasheets.
- measured freq resp isn't always reliable itself
- no phase information, plus phase information can differ as well, and/or needs to be derived
- the simulation model for diffraction and interference effect isn't accurate enough. (read: in reality becomes quickly to complex to model)
Fyi, the bafflestep simulation is in most cases fine, unless you're going for bizarre shaped cabinets.
When the shortcomings are known this automatically also means it's not entirely useless.
As long as we respect those issues and/or stay (far) away from it, those issues are non-existing for our purposes.
There are also techniques to kind of get a rough idea about these issues.
The first problem is something to really look out for.
It can sometimes be a challenge to figure out under what conditions manufactures measure their things.
Was it measured on an AES baffle, in a certain cabinet etc etc?
Also in many cases, the freq resp itself can be cherry picked, or contains to many reflections.
(See SB Acoustics for example).
A workaround here, is to find data that is more reliable.
Boxsim for example seems to have a decent (not perfect) correlation with actual data.
So one could for example also pick his components accordingly.
Phase information is only (very!) important with multiple speakers.
As soon as this phase information is not accurate anymore, the whole crossover filter is close to be meaningless.
(for the freq resp connection between the two, not for the speakers individually)
So when designing a fullrange speaker (so just one fullrange woofer), this is not important.
It's also not very important for subwoofers, or woofers that aren't being used up till the diffraction issues start.
For the last point, one can decide to make a baffle with as minimal diffraction issues as possible.
Another way is to just put the crossover before the point the diffraction issues start.
A couple of these things also count for 2-channel measurements.
This is only gonna be an issue* when there are more than one speakers involved.
Second to that, there are (many) workarounds.
Especially when active filters are being used, one can decide to measure on the fly.
Although this can be a very tedious process, it's doable.
In this case you measure with just a simple (USB) microphone, and change the filter parameters per speaker.
Next you measure both speakers at the same time**
Go back, rinse and repeat.
This also works with gated measurements.
Is this ideal? No!
Does it get you by? Yeah, kind of.
The point is to make people aware for what to look for.
Instead of just saying it can't be done and one can't make ANY good speakers with it.
Yes, without proper measurements and off-axis response with multiple axis, making a good professional sounding loudspeaker is a big challenge.
Impossible? Maybe, in my opinion likely (that it's not possible).
On the other hand, I know people who do almost everything by ear and on experience.
These people make fine sounding loudspeakers, yes there is definitely room for improvement.
Which leads to the last point. Some people are just simply not looking for the "best possible speaker" (whatever THAT even means).
But these people are looking for something between "randomly wild guessing" and having to put all the effort of getting and buying all this measurement equipment.
Again, as long as people are very well aware of the shortcomings, and most importantly, understand these issues and also clearly change their expectations to it.
I don't have an issue with it and still don't understand at all why it almost seems a taboo?
In some cases it sounds more like just a personal OCD thing than anything else to be very honest.
Makes me wonder if some also weight the ratio between tea and water to the exact micro gram and monitor the water temp closely to 0.1 degrees C ?
Other people just get the cheapest tea they can find and shove it in a mug and add some milk to it.
* Way back I did most of my measurements with just single channel, and the issues were to small to have any significance.
We are talking about WinXP era with a simple 16 bit USB soundcard and just a regular normal amp.
These days I don't know exactly how that works anymore with Win8/10 driver issues and delays.
But even here there are workarounds available. Again, it's not ideal.
** Important is to take at least two off-axis measurements as well btw.
So like 15 degrees and 30 degrees or something. Absolute bare minimum.
To summarize the above.
There are always multiple ways to skin a cat (poor kitty )
Yes there are ways that will give more reliable and more consistent results.
Which sometimes is nice if you do this on a professional level.
But there are still ways that maybe won't give you a a perfectly skinned cat ( ), but is good enough for the context its being used in.
Knowing it's not perfect, and it never will be perfect.
And you know, sometimes imperfection even has its charm.
There are always multiple ways to skin a cat (poor kitty
)Yes there are ways that will give more reliable and more consistent results.
Which sometimes is nice if you do this on a professional level.
But there are still ways that maybe won't give you a a perfectly skinned cat (
), but is good enough for the context its being used in.Knowing it's not perfect, and it never will be perfect.
And you know, sometimes imperfection even has its charm.
Member
Joined 2003
My definition of "real design work" would be something that is meant to be built and enjoyed for years, with design data that closely matches reality. The "unreal" design work would be something using bad or missing data, unknown factors making the data at risk of not matching reality. This sort of task may still be useful to learn the tools and experiment in "what-if" scenarios, but that's it.
If the goal is to use VituixCAD to its fullest, just follow the tried and true method to skin the cat. It's spelled out in the VituixCAD documentation clearly.
I understand the "data that matches reality" part.
Although one could rightfully so make an argument how important that really is?
(again, for professional work that is a totally different story).
As long one is happy with the result what does it really matter?
I don't follow the "enjoying for years" part at all?
Next to me at this very best moment, I have an old tube radio from 1948.
The thing still purrs and choo choos great after all those years with some repairs.
From an idealistic stand point, there is a lot "wrong" with this radio.
Nevertheless, from a pure subjective and nostalgia feeling I really enjoy listening to it.
Sometimes even forget "how bad" it is.
Missing data or unknown factors are definitely NOT just only for learning.
Than you're missing an extremely important point, even in physics and science if we put this in a more general perspective.
In the end it's all about questioning how predictable certain techniques are.
Or in other words; "how likely will the speaker perform as intended".
Some techniques will give you a much bigger correlation, other techniques will be less.
They are mostly just tools that can be used in development.
Although they don't necessarily have anything to do with the outcome perse.
How important it is that this correlation ratio is very high, totally depends on context.
In some cases predictability isn't important or less important.
One can still build an extremely good speaker with limited predictability.
Although, it's probably more LIKELY that you won't get the results you were aiming for.
Or again in other words, when good techniques are being used it's a lot more likely you will get the results you were aiming for.
How important that is, is completely subjective.
Some people are fine having very loose expectations.
In the end, there is absolutely no "true" way to skin a cat. That is just a very silly meaningless statement on itself.
The act of doing something doesn't have anything to do with the end result. (maybe it was a bad cat to begin with?)
Following certain guidelines only says something about the predictability of having a certain end result.
Which doesn't automatically mean you can't get the result on a different way.
Fundamental part of physics, by absolute 100% definition we have to deal with some kind of error as well as statistics.
Which also means by definition there will always be another way.
The changes might be very small sometimes, but never zero.
Obviously an extreme example, but we are talking about succeeding making a loudspeaker one is satisfied with.
Which is a lot less extreme, meaning that the possibilities are also higher from either side.
Summary in simple words;
Some people are totally fine with the idea that the methods they used to design a loudspeaker weren't that accurate.
They had very loose expectations to begin with and know and are (very) aware that this probably will show in the end result (or the lack thereof)
Seeing the problem from that point of view, they just skinned the perfect cat, since their expectations met the end result.
Plus like I mentioned above, there are techniques or design choices that will get you damn close.
Although one could rightfully so make an argument how important that really is?
(again, for professional work that is a totally different story).
As long one is happy with the result what does it really matter?
I don't follow the "enjoying for years" part at all?
Next to me at this very best moment, I have an old tube radio from 1948.
The thing still purrs and choo choos great after all those years with some repairs.
From an idealistic stand point, there is a lot "wrong" with this radio.
Nevertheless, from a pure subjective and nostalgia feeling I really enjoy listening to it.
Sometimes even forget "how bad" it is.
Missing data or unknown factors are definitely NOT just only for learning.
Than you're missing an extremely important point, even in physics and science if we put this in a more general perspective.
In the end it's all about questioning how predictable certain techniques are.
Or in other words; "how likely will the speaker perform as intended".
Some techniques will give you a much bigger correlation, other techniques will be less.
They are mostly just tools that can be used in development.
Although they don't necessarily have anything to do with the outcome perse.
How important it is that this correlation ratio is very high, totally depends on context.
In some cases predictability isn't important or less important.
One can still build an extremely good speaker with limited predictability.
Although, it's probably more LIKELY that you won't get the results you were aiming for.
Or again in other words, when good techniques are being used it's a lot more likely you will get the results you were aiming for.
How important that is, is completely subjective.
Some people are fine having very loose expectations.
In the end, there is absolutely no "true" way to skin a cat. That is just a very silly meaningless statement on itself.
The act of doing something doesn't have anything to do with the end result. (maybe it was a bad cat to begin with?)
Following certain guidelines only says something about the predictability of having a certain end result.
Which doesn't automatically mean you can't get the result on a different way.
Fundamental part of physics, by absolute 100% definition we have to deal with some kind of error as well as statistics.
Which also means by definition there will always be another way.
The changes might be very small sometimes, but never zero.
Obviously an extreme example, but we are talking about succeeding making a loudspeaker one is satisfied with.
Which is a lot less extreme, meaning that the possibilities are also higher from either side.
Summary in simple words;
Some people are totally fine with the idea that the methods they used to design a loudspeaker weren't that accurate.
They had very loose expectations to begin with and know and are (very) aware that this probably will show in the end result (or the lack thereof)
Seeing the problem from that point of view, they just skinned the perfect cat, since their expectations met the end result.
Plus like I mentioned above, there are techniques or design choices that will get you damn close.
Member
Joined 2003
Bad data, missing data, and poor interpretation of data are the cause of many audio misconceptions, because people have "come to conclusions" based on it. If the goal isn't to design using data that represents reality then I don't know what we're doing here. There's lots of people enjoying their designs using a USB mic and Xsim, maybe you will enjoy a design using an online calculator, but neither are going to be my recommendation. Maybe we have different goals, but your tube radio from 1948 is not the performance bar that I am trying to reach. VituixCAD is designed to follow CTA-2034-A standard measurement data set, you might call it the current "standard way to skin the cat". What are you saving from taking short cuts of "at least 2 off-axis angles" and whatever else nonsense? It's just bad advice.
I don't think this discussion is really going anywhere, so let's just move on from it .
Last edited:
No they are absolutely NOT the cause of many misconceptions.
The approach and ignorance (nofi) of people (or the lack there off) are causing misconceptions.
In science, experts constantly have to deal with crappy data.
Sometimes to even get pretty reliable results.
What are you describing now is also personal preference.
aka; What I already asked before, define "best speaker".
The the example I just gave, was just one tiny little example.
Not my absolute personal preference. (if that wasn't obvious clear already)
My personal preference differs from project to project, from customer to customer.
Sometimes that goal is to represent something without any concessions, which automatically also has its price.
Sometimes budget or esthetics are more important.
Point is, totally different discussion, but please don't force your personal approach for audio into other people and what they MUST make or follow.
(also best way to loose clients or customers btw)
Fyi, standards are again tools to be used, but not necessarily the truth and absolutely not the only way of doing so.
Those are two very different things.
It's a standard way to skin a cat, but not just the best way and certainly not the only way.
It's very likely the way with the most predictable result. (often in a relative sense, not an absolute sense)
We can't move on the discussion if people here are very clearly not able to understand the difference between standards and predictability.
Yet seem to mix up even their personal opinion in it as well.
Which makes me sad, because people with a scientific, physics or electronics background ought to be familiar with those concepts.
And fyi, please read again, because I haven't said any nonsense or advice in absolute sense at all.
Very obviously 2 off-axis angles isn't enough to get the whole picture, everyone knows that.
Really don't appreciate when things are being taken out of context.
The approach and ignorance (nofi) of people (or the lack there off) are causing misconceptions.
In science, experts constantly have to deal with crappy data.
Sometimes to even get pretty reliable results.
What are you describing now is also personal preference.
aka; What I already asked before, define "best speaker".
The the example I just gave, was just one tiny little example.
Not my absolute personal preference. (if that wasn't obvious clear already)
My personal preference differs from project to project, from customer to customer.
Sometimes that goal is to represent something without any concessions, which automatically also has its price.
Sometimes budget or esthetics are more important.
Point is, totally different discussion, but please don't force your personal approach for audio into other people and what they MUST make or follow.
(also best way to loose clients or customers btw)
Fyi, standards are again tools to be used, but not necessarily the truth and absolutely not the only way of doing so.
Those are two very different things.
It's a standard way to skin a cat, but not just the best way and certainly not the only way.
It's very likely the way with the most predictable result. (often in a relative sense, not an absolute sense)
We can't move on the discussion if people here are very clearly not able to understand the difference between standards and predictability.
Yet seem to mix up even their personal opinion in it as well.
Which makes me sad, because people with a scientific, physics or electronics background ought to be familiar with those concepts.
And fyi, please read again, because I haven't said any nonsense or advice in absolute sense at all.
Very obviously 2 off-axis angles isn't enough to get the whole picture, everyone knows that.
Really don't appreciate when things are being taken out of context.
Last edited:
Well the way I see it vituix does not know if I entered correct measurements or not, nor does it issue a warning if my inserted data only had 0, 15, 30, and 45 degree measurements. Also I'd wager that if Bill took actual measurements and John took actual measurements, they would differ to some degree ..none of this imho should prevent me learning the software. Why should I go ahead and invest in equipment and drivers if having them I still can't use the software successfully.????? I totally understand how real measurements in a cabinet make the process easier and better.....except for the buying part.
Thank you. It was much easier for me to understand than the vituix guide.
Yes I'd be interested in this also. Please post link
Member
Joined 2003
ignorance = poor interpretation, I think you may actually be agreeing with me. Bad data and poor interpretation go hand in hand, I competely disagree that one is absolutely not the cause of the other.
Predictability and repeatability are important, hence the standard. I would suggest that if you have a better way, then please document it thoroughly so it can become the new standard.
VituixCAD follows a standard and provides instruction on how to create data quickly and easily that follows this standard (2 channel, all angles, etc), so in the context of using this tool, I'm not sure what argument you're trying to make.
In the context of instructing a new designer/builder, providing an instruction that follows an established standard should be a good starting path, as in a known good reliable and repeatable method of skinning the cat. Trying to explain every different method and pros/cons, you can use a USB mic is some cases, but not others, etc. is really over-complicating (probably over-whelming) what can be a simple path to "rather good" results. (everyone knows best = opinion). Following this standard method is not even that complicated or expensive, so I'm sorry but I'm not grasping the point of this debate.
Member
Joined 2003
The jig that I use is just a simplified version of the ARTA jig here:
https://artalabs.hr/AppNotes/AN1-MeasuringBox-Ver2-Rev3Eng.pdf
I designed it to be used in conjunction with a USB audio interface such as the Steinberg UR22 or Motu M2 or similar. If you want the cheapest option out there that is "good enough", look at the Behringer UMC202 or UMC204.
It uses TRS connection, so it's just a few resistors and some connectors and some short patch cables to patch it into the USB interface. On a USB interface the TRS pins are not energized with phantom power, only the XLR pins are. The 10 ohm resistor should be at least 5W, the other resistors can be small basic 1/4W.
It is detailed in the ARTA jig document, but I have included a spreadsheet that calculates the attenuation provided by the probe resistors, so you can understand why the values are what they are and adjust them if needed. 20kOhm input impedance of the audio device is assumed, but you will notice that Zin from 10k - 100k doesn't make that big a difference in the attenuation result. Consult your audio device specifications for the input impedance. The voltage and power calculations really only come into play for high SPL testing such as distortion. For normal frequency response, generally <1W is needed.
It is important to use the jig with single ended amps only. Low cost class-D chip amps are generally BTL connected internally and cannot be used.
My recommended process moving forward to start into DIY speaker design is as follows:
1. Purchase a USB audio interface. I use the Steinberg UR22mkii, a low cost alternative would be the Behringer UMC series of products.
2. Purchase measurement mic. I use Line Audio OM1, look at Sonarworks SoundID for a low cost solution to get started.
3. Build the jig. Just a few resistors, connectors and some scrap wire can get you started.
4. Play around with ARTA/LIMP. All features are available to play with for free as shareware, you must purchase a license to save the measured impulse response which will be very convenient for VituixCAD.
5. Purchase ARTA License and have fun making speakers. From here you can start by following Kimmo's guide for ARTA measurements for VituixCAD.
REW is a free alternative to ARTA, but I'm not experienced with it and dislike the interface the few times I've put it to use. YMMV.
Attachments
Last edited:
So what's the difference IN THIS CONTEXT between USB and balanced mics. I have a decent audio interface already (not USB) and 32 ADC'c and a bunch of clean mic pre's with phantom power. All of which colour the sound in some way....but it would seem to me that a USB mic has less (external) components and their internal components are a known quantity to software like Rew. What does the more complex setup with more physical parts achieve?
The point is making people less ignorant and teach them more.
Teach to think about the choices they have to make, what the consequences are and why.
Blindly following manuals, tutorials and standards is not only not teaching people anything.
In some cases/fields/areas it can actually be very dangerous or even lethal.
The main reason why I getting (obviously) a little irritated by it, is because you explain it on a way that's the only right way of doing things period.
Which is fundamentally just not true.
I am also NOT saying it's untrue, but I am saying that it contains shades of gray.
Extremely important difference.
Other methods might not be as efficient, or smooth, doesn't mean you can't get similar (or better) results.
If you don't explain those shades of gray, people still keep being ignorant.
Result is that another myth story rises. (which on some other forums seems to be the standard already
)
Last edited:
Member
Joined 2003
USB mic is just a standard condenser mic with built in USB audio codec and phantom power.
When measuring multiple drivers for loudspeaker design, it is important to keep a locked timing reference for the measurement to capture time of flight so that the delay, or the difference in acoustic centre, can be captured accurately and effortlessly. This is where a 2channel measurement comes into play. The “reference” channel is a closed loop feedback of the electrical signal as if leaves the amp, so time of flight from driver to mic is captured. There is no “time of flight” accurately captured with a USB mic.
Additional benefit of the reference channel feedback is that all nonlinearity of the amplitude response from up to the reference probe can be compensated for by simply normalizing the reference channel input. You can observe this with the gear you have and ARTA. Simply loop back output to input, both channels. Run single channel and dual channel measurement (“1FR” and “2FR” buttons) with high sample rate so you have response to 40kHz+, and see the difference between single channel and dual. One will show you the nonlinearity of your equipment, and the other will be dead flat apart from noise .
It sounds like all you need is the jig and a measurement mic, not much of an investment to get started in the exciting world of speaker design
Last edited:
Member
Joined 2003
I don’t propose to follow blindly, question the instruction and understand it’s purpose and limitations is of course very important or learn.The point is making people less ignorant and teach them more.
Teach to think about the choices they have to make, what the consequences are and why.
Blindly following manuals, tutorials and standards is not only not teaching people anything.
In some cases/fields/areas it can actually be very dangerous or even lethal.
The main reason why I getting (obviously) a little irritated by it, is because you explain it on a way that's the only right way of doing things period.
Which is fundamentally just not true.
I am also NOT saying it's untrue, but I am saying that it contains shades of gray.
Extremely important difference.
Other methods might not be as efficient, or smooth, doesn't mean you can't get similar (or better) results.
If you don't explain those shades of gray, people still keep being ignorant.
Result is that another myth story rises. (which on some other forums seems to be the standard already
My way (err, the Vituix way
) is not the only way, maybe not the right way in someone else’s eyes, but it is the way that I am recommending. I have tried to explain, a newcomer can be easily overwhelmed with all the different ways to do the same thing, conflicting information depending on who you listen to, get pulled in multiple directions can be quite frustrating and confusing. There’s a lot of bad advice out there. So if you ask me which one way I recommend to learn as a design path, this is it. If someone asks me why I don’t do things some other way , I do my best to explain why, as I have just done above with regards to the USB mic question.