The No.1 Website for Pro Audio
 All  This Thread  Reviews  Gear Database  Gear for sale     Latest  Trending
Software for testing Mic Pres
Old 12th June 2019
  #1
Gear Maniac
 
pipelineaudio's Avatar
 

Software for testing Mic Pres

I'm gathering piles of mic pres in order to do some actual real world, quantified, not nonsense numbers to see just exactly how different these things are and especially to look at so many mythical or near mythical claims

I have been trying Rightmark Audio Analyzer, Plugin Doctor and Room EQ Wizard

Noise and frequency response measurements are easy enough in all of these programs, but distortion has been tricky

All three have THD analyzers, but what I really want to see is distortion vs level on these. It seems like all three just send a static level to check. Plugin Doctor seems to do it at 0dbfs as well which makes it tricky

Any suggestions on how I measure the distortion? The claim so often is that you can magically know at all times the level that is coming in and then you can "push" the mic preamp hard and reveal all this magic "character"

I really like the way this distortion chart goes, but I dont know which software it is

Old 12th June 2019
  #2
You can generate a ramp signal - e.g. a 1k sine wave ramping up from silence to full scale. You can then record the pre's response to it and look at it in 3d (waterfall graph). That might give you some insight into distortion characteristics.
Old 12th June 2019
  #3
Lives for gear
 
nucelar's Avatar
 

That particular plot in the picture is from a Rhode & Schwarz analyzer (hardware) you can tell from the logo. You also might have seen plots with Ap logo: Audio Precision.
The "Audio Measurement Handbook" can be downloaded from AP's website. It's A classic if you are into this stuff.
What you can do with software is limited by the lack of calibration and the quality of your converters.
Old 12th June 2019
  #4
Lives for gear
FWIW I’ve found Fuzzmeasure a lot faster and easier than Rightmark. But I don’t know any app that’ll draw a 3-dimensional distortion at frequency vs level. Multiple passes, easy. All in one burst? Is there anything that’ll do it?

Or is the OP interested in pre distortion at a single frequency, ignoring bypass caps and transformers?

—-
Trivia: guy who wrote Rightmark is now a senior dev at iZotope.
Old 12th June 2019
  #5
Gear Maniac
 
pipelineaudio's Avatar
 

I'll look into that Fuzzmeasure

Orson. Once I record that waveform, what 3d waterfall app or plugin could i look at it in?
Old 12th June 2019
  #6
Gear Maniac
 
pipelineaudio's Avatar
 

Quote:
Originally Posted by Jay Rose View Post
Or is the OP interested in pre distortion at a single frequency, ignoring bypass caps and transformers?
I'm not sure about bypass caps and transformers, but I would think a single frequency would be ok for what I'm doing.

Last night I ran one of the mic pres that people say gets magically "buttery" as you increase the gain, and found that no matter what I set the mic pre gain at, it always distorted at the same relative level of gain to input. I'm hoping to be able to make some graphs of that sort of thing
Old 12th June 2019
  #7
Quote:
Originally Posted by pipelineaudio View Post
I'll look into that Fuzzmeasure

Orson. Once I record that waveform, what 3d waterfall app or plugin could i look at it in?
I believe I've been using MATLAB at work. I think SpectraLAB/SpectraPLUS also has similar features. I also believe you can whip up a Python script to do that if you're savvy with a bit of programming.

If you can play the recording back or hook the preamp live, I believe you can check out free real-time analyzers like Visual Analyzer.

Cheers.
Old 13th June 2019
  #8
Gear Maniac
 
pipelineaudio's Avatar
 

This is exactly what I'm looking for, but its APwin so not really in my range...Any free or cheap software that does this?

Old 13th June 2019
  #9
Lives for gear
 
nucelar's Avatar
 

You can always produce a graph manually if you have the patience. In fact, the interesting part of this measurement is around the saturation breakpoint, if you can manually increase the input level in small steps you can obtain a nice graph for that transition . The above graphs seem quite coarse in that respect, although it probably is adjustable.
In many preamps the gain cell is really a fixed gain element, e.g.60 dB. With the "gain" knob you're not really adjusting the gain of the amplifier, but attenuating the signal that gets to the amplifying circuit. Therefore, moving the "gain" knob or adjusting the input level is equivalent in most cases.
Old 13th June 2019
  #10
Also I'm not sure what you are testing them for, but you might not want to see the THD figure because it is mostly useless. For example, last time I used the waterfall spectrum plot was to find out the amplitude of the input signal at which a certain guitar pre was stepping over the 50/50 duty cycle square wave (all even harmonics), which was clearly visible. That is, if you need some qualitative and quantitative analysis, you might still consider using full spectrums instead of a condensed arbitrary THD figure.
Old 13th June 2019
  #11
Gear Maniac
 
pipelineaudio's Avatar
 

AhhhI I think I can load wave files into REW for a waterfall plot, but I didnt think that did distortion, I thought it was level vs frequency vs time
Old 13th June 2019
  #12
Oh yeah, I forgot REW did waterfalls too. Your three axes would need to be frequency for X, level for Y (like an ordinary spectrum plot), and then time for Z. Exactly what REW does I think.
Old 13th June 2019
  #13
Gear Maniac
 
pipelineaudio's Avatar
 

I dont quite understand why that would show distortion well, I would think it needs a level ramp from really quiet to really loud, but I'll go try it and see!
Old 13th June 2019
  #14
Yes, you need to generate a "ramp" input signal - that is a sine wave of a single frequency (I used 1kHz) that is getting louder from total silence to full scale. I did it in Audacity (free) and made it 10 seconds long to get better temporal resolution.

So when you interpret the result, the 1kHz "ridge" (when you look at the waterfall) is what you expect to come out. Everything else is distortion by definition - the harmonics that are being added by the circuitry. And as opposed to the single figure THD, you can actually see what is happening. You can also see if there are any 60/120Hz harmonics (and their respective intermodulations) being added by the power supply, you can see the order of the harmonics being added (which is VERY significant in terms of listening experience) etc.
That is why in my opinion, this 3D plot is way better than the 1D THD estimation.

Feel free to ask away any questions.
Old 13th June 2019
  #15
Gear Maniac
 
pipelineaudio's Avatar
 

Were you able to import that wave into REW or does Audacity have its own waterfall? I couldnt figure how to get a sine sweep into REW, but I'm happy to use audacity if it does that!
Old 13th June 2019
  #16
Lives for gear
 
nucelar's Avatar
 

I might add that to analyse the differneces between devices it would be very useful to see the evolution of individual harmonics vs input compared to each other. See if your software supports it (individual harmonic distortion) , otherwise it is just a bandpass filter that focuses the output on a particular harmonic of interest and measures its amplitude.
PD. I guess you can see this in a well defined 3d waterfall, but the visualisation is more messy imo.
Old 13th June 2019
  #17
Audacity has a spectrogram view, but its not that great. You can see here. I use it only for test signal generation.
As for REW, I haven't tried it yet. I can fiddle with it at home in a few hours and see if I can produce the same output I saw in MATLAB on my test files.
Old 13th June 2019
  #18
Quote:
Originally Posted by nucelar View Post
PD. I guess you can see this in a well defined 3d waterfall, but the visualisation is more messy imo.
Yeah, what you're talking about is a "slice" along the time axis of the waterfall at a given frequency. I don't think I've seen this anywhere. That calls for some software development
Old 13th June 2019
  #19
Lives for gear
 
nucelar's Avatar
 

Problem with using a waterfall (like REW), is that it is defined as a "spectum evolution over time", so in the z axis (towards you) you have time. These acoustic measurement software are designed around a certain methodology of sending a noise or freq sweep, which does not apply in this case.

The OP instead wants to see what happens depending on input level, so unless you have a mathematically defined relationship between time and input amplitude, you only will see qualitative results, difficult to interpret.

As i said if I were the OP I would carefully increase the input level stepwise and take the output spectrum at each step. with this array of spectra you can plot very interesting data, for example evolution of the 2nd harmonic vs 3d harmonic.

Before finding an automated all-in-one method one must undestand what basic measurements are behind
Cheers
Old 13th June 2019
  #20
Quote:
Originally Posted by nucelar View Post
so unless you have a mathematically defined relationship between time and input amplitude, you only will see qualitative results, difficult to interpret.
Yes, and this is exactly what happens when you generate the input ramp signal - of course you have control over the relationship between time and amplitude.
Old 13th June 2019
  #21
Lives for gear
 
JohnRoberts's Avatar
 



Have fun..

JR
Old 13th June 2019
  #22
Quote:
Originally Posted by pipelineaudio View Post
Were you able to import that wave into REW or does Audacity have its own waterfall? I couldnt figure how to get a sine sweep into REW, but I'm happy to use audacity if it does that!
Well, I did that. and while you can see some interesting stuff, the controls for the visuals are not extremely ergonomic. Nevertheless, I manage to get something like this (see attached).
Attached Thumbnails
Software for testing Mic Pres-harm.jpg   Software for testing Mic Pres-psu.jpg  
Old 14th June 2019
  #23
Gear Head
 
mskala's Avatar
 

I don't think it is realistic to expect to do this with "software." At best, you will be measuring the performance of your audio interface as much as that of the preamp. Even if you are willing to buy a good audio interface for the project, you'd get better value and more accurate results putting that money into test equipment specifically designed for this kind of task rather than a PC audio interface (which will be optimized for other purposes) and PC software.
Old 14th June 2019
  #24
Quote:
Originally Posted by mskala View Post
Even if you are willing to buy a good audio interface for the project, you'd get better value and more accurate results putting that money into test equipment specifically designed for this kind of task rather than a PC audio interface (which will be optimized for other purposes) and PC software.
Is there an affordable bit of lab kit that can do these kind of measurements?
Old 14th June 2019
  #25
Gear Head
 
mskala's Avatar
 

Quote:
Originally Posted by Orson Maxwell View Post
Is there an affordable bit of lab kit that can do these kind of measurements?
In general it's the kind of task handled by an "audio analyzer," but I don't have a specific brand or model to recommend. I'd look at flea markets and other used sources. Depending on features, new/used, and fancy interface, they should range from a couple hundred dollars into the thousands. At the low end of that range someone might have to make individual measurements and do the plotting on a graph themselves (or use a PC just for that part); but the advantage would be that the measurement numbers would actually reflect the device being tested instead of whatever unknowable distortion is built into the PC's audio interface.
Old 14th June 2019
  #26
You can test the interface closed loop though.

But I agree - I mostly used this approach to test various thermionic guitar devices, you can run into the practical noise floor when testing higher end mic pres.
Old 15th June 2019
  #27
Gear Maniac
 
pipelineaudio's Avatar
 

Are a set of apogee converters so amazingly deficient that you wouldn't be able to measure differences big enough to be audible? Especially as claimed by so many "TOTALLY AUDIBLE OMGZORZ YOU GOTTA BUY THE FIVE THOUSAND DOLLAR PREAMP"

I would hope something like a Rosetta was at least flat enough for that! If not, why do we buy these damn things?
Old 15th June 2019
  #28
Gear Head
 
mskala's Avatar
 

People think they can hear differences that don't exist in objective measurements, and objective measurements reliably pick up differences that people don't hear. If you want to pick a preamp that sounds good to you, you can do that by using your ears; but I thought this thread was about making accurate, reproducible, objective, numeric measurements... and my point is that PC software is not the important part of solving that problem. It requires proper test equipment, which is not a PC audio interface.

With preamps in particular, there are also important issues of how it responds with different input and output impedances. The best PC audio interfaces will not be able to present realistic simulations of a microphone or other signal source's impedance to the preamp, because that's not what the interface is designed to do. So at best you can measure how the preamp behaves when attached to the impedance of the PC audio interface, which may or may not (probably not) realistically simulate a microphone.

Last edited by mskala; 15th June 2019 at 04:00 PM.. Reason: note impedance etc.
Old 16th June 2019
  #29
Lives for gear
 
Radardoug's Avatar
 

I agree , you need dedicated test equipment. And then you need to spend hours doing frequency response, distortion, and noise testing. This is not a trivial process. I have built mike preamps, and spent months doing measurements on just one preamp.
Old 16th June 2019
  #30
Gear Addict
When it comes to distortion testing, I use REW but do it manually:

When you run a frequency sweep, you get a distortion plot which shows distortion level by component as well as THD.

I use that, and ramp up levels manually. That way, you've got distortion (by component) vs frequency vs level. A really useful set of data.

Chris
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump