Lets do it: The Ultimate Plugin Analysis Thread

[Steab]

So, the "digital eq myth&fact" and some other threads inspired me to run some tests on some of my plugins. I'm gonna share the results (screenshots, so you will able to see the plugin with its exact settings too) here and I think it's going to be interesting, as my plugin collection is by now quite big. I will be adding more plugin test results whenever I find the time.

I hope that with the contribution of you guys, we will be able to round up results of EQ, Compressor, emulation and various plugins in terms of frequency, harmonic distortion, harmonic response and maybe later dynamics.
I do not intent to test those plugins in every aspect.
Screenshots (or even png exports) of the plugin responses with different settings would be appreciated.

Let me clarify this, im not starting this to make the point that plugins are or are not good. The point is to see what those specific plugins do and what they don't. The last thing we want is comments like "I discovered the only truth, it's mixing otb, all plugins suck im so smart".Also,

Im will be using the free VST plugin analyser (running 1 kz sine wave) for the tests, and I advice you do the same, it's really handy.
(You can download it here:Christian’s Blog » Measurement)

Here we go:


Abbey roads tg12413 2005 and 1969:

[Steab]

Rocket, default, compressing, compressing + impetus.

[Steab]

Urs Nmix: Just inserted, then some eq carving. Is there a possibility those urs plugins emulate the sound of the hardware units they're supposed to?
I doubt it.

[Steab]

Warmifier: Mode 5842, then mode 12at7. Nothing else configured.

[mark.james]

Great idea! I wonder if that VST Analyser will work in the VST to RTAS wrapper... I don't see why not. I'll try it out when I get home. I've a good amount of plugins I can try this with.

Mark

[Teddy Ray]

Steab, how do you use the analyzer??

I have a .DLL file and an executable..

[SWAN808]

I wish there was an AU or Mac VST version of the plugin analyser...

[Steab]

Mark, the analyser is an executable file, unfortunately analyzing only vst plugin dlls. But thanks and hope you try your vsts!

Teddy, you have to open the exe file then load and browse for the plugin. In the menu you'll find the measument and other options..

Swan, I wish too, but for the time being there isn't, let's do our best for the accesible plugins.
You can always test your plugins with the traditional method, run a sinewave through the plugin and watch it in your frequency analyser plugin.

PSP Oldtimer coming!

[Amun Ra]

Cool idea!

Could someone please explain "harmonic response", though? I have no idéa what that means, and the graphs didn't help...

I vote for phase diagrams also.

[Teddy Ray]

Originally Posted by Steab

Mark, the analyser is an executable file, unfortunately analyzing only vst plugin dlls. But thanks and hope you try your vsts!

Teddy, you have to open the exe file then load and browse for the plugin. In the menu you'll find the measument and other options..

Swan, I wish too, but for the time being there isn't, let's do our best for the accesible plugins.
You can always test your plugins with the traditional method, run a sinewave through the plugin and watch it in your frequency analyser plugin.

PSP Oldtimer coming!

when I tried to load up several .dll I get an error..."wrong program version:"

[Steab]

I must admit I'm loving this compressor!

Oldtimer: Clear, then valve mode, frequency and harmonic distorsion.

[Steab]

Originally Posted by Teddy Ray

when I tried to load up several .dll I get an error..."wrong program version:"

That's odd.. I'm sorry I don't have a clue what could be causing it.

[Steab]

Stillwell VibeEq, one light, colored plugin.
Boosts, hp and lp (os) filters, in terms of frequency and harmonic distortion.

[Steab]

Event Horizon analyzed in some of its presets: default, limiter and analog warmth.


I would be doing urs strip pro but it's analyzed right here: URS Classic Console Strip Pro analyzed. - signaltonoize.com

[jdtrbn]

Note that in compressors it's hard to avoid distortion. Modulation distortion, which comes from the gain change and is signal dependent, is very different from constant distortion, which usually isn't implemented in compressors (it usually sounds bad).

EQs are nice for analyzing, because they're usually signal-independent, constant in time. Most processes are not.

I'd also like to note that generally distortion is not what gives gear it's sound, unless it's a very bad sounding unit. Especially it's not the secret "mojo" or "warmth" factor that magically un-digitalizes your sound. Distortion is often unwanted, and to me it's very understandable that many plug-ins don't implement it. Just like they don't produce the noise of the original units - they don't emulate breaking tubes or old crackling pots either. There are some exceptions, for example Vari-Mu designs and 33069 have some pretty funny distortion-related stuff going on that define a part of their sound.

Someone smarter than me has written about related subjects: tube_talk

PS. As a bit extreme example, tube guitar amps usually have a signal-dependent harmonic content that results from adding DC at transients. Transistor amplifiers don't generally behave like this. It's most likely an important part of their sound, but it's not simple to find this kind of time-variant effects in analysis unless you know what to look for.

[imaginaryday]

i would love to participate as i have many plugins, but i have no idea how to do this. can someone give some instructions?

thanks.

[Steab]

Originally Posted by imaginaryday

i would love to participate as i have many plugins, but i have no idea how to do this. can someone give some instructions?

thanks.

Start by reading the previous posts!
If vst is your plugin platform you can download the Vst plugin analyser, which is free, theres even a link in my first post.

[jdtrbn]

BTW if someone has UAD Harrison EQ, it would be interesting to analyze what makes people like the HP/LP filters so much.

[TonyBelmont]

Making measurements with graphs can tell you some things, like for instance if the plugin is imparting "harmonic distortion" to a 1K signal. But, I think that evaluating tools in the context of a mix is the best way to analyze if it's the right thing to you... it really is all about how the tool works for you in the context of a "working" situation.

[Steab]

Originally Posted by TonyBelmont

Making measurements with graphs can tell you some things, like for instance if the plugin is imparting "harmonic distortion" to a 1K signal. But, I think that evaluating tools in the context of a mix is the best way to analyze if it's the right thing to you... it really is all about how the tool works for you in the context of a "working" situation.

I get your point and it's ok.
But other than evaluating the plugins while mixing, I find it really interesting to really see what's going on there, and I dont mean that analysing will tell you everything about the plugin neither that you will instantly know if it's good for you. It's just good to know that one plugin does add harmonic content to the signal and another one doesnt (and how much), or "what does that hpf exactly do when i press it", or "what does that PHAT button exactly do?" etc.
Especially when people's views on the same plugins vary so much that you know for sure some of them talk nonsense.
It all adds up to our knowleldge data base, you know?

Anyway, come on people, analyse those plugins!

[pacorro]

ok, I play

I've passed a 1khz sine (generated in Wavelab 6, 24 bit/44.1) though a few plugins, mainly compressors.

this is NOT a scientific test, because most of the comps have fixed times for attack and release, so is really difficult to match the settings and compare them all.

I'm not pretendig to say wich is good or bad, I'm not talking about the sound.
I'm just testing the harmonic distortion added UNINTENTIONALLY by the plugin, I mean ALIASING.

From the wiki:

In signal processing and related disciplines, aliasing refers to an effect that causes different signals to become indistinguishable (or aliases of one another) when sampled. It also refers to the distortion or artifact that results when the signal reconstructed from samples is different from the original continuous signal.

Originally Posted by andy_cytomic

Too much aliasing is a big problem, and will mean your audio suffers hugely - even though a listener might not be able to say "oh that's aliasing" if you do A/B tests with the aliased and non aliased version, they will prefer the non-aliased one.

Ok. Let's start.

You'll see attached the original 1khz sine.
every plugin will show some peaks.
The big peaks are harmonics, the small ones are not harmonics. These small peaks means the plugin is not being accurate with the original file.

From the wiki...

A harmonic of a wave is a component frequency of the signal that is an integer multiple of the fundamental frequency

inharmonicity is the degree to which the frequencies of overtones (known as partials, partial tones, or harmonics) depart from whole multiples of the fundamental frequency. Acoustically, a note perceived to have a single distinct pitch in fact contains a variety of additional overtones.

So when a plugin has harmonic content, the sound is probably richer and can be pleasant to your ears. Most of the hardware adds this kind of harmonics and we love that sound.
In my test you can watch that the harmonics (big peaks) are 3k, 5k, 7k...being all of them multiples of 1khz (the original sinewave)

But when you see small peaks at 900, 800, 1100, 1200, etc, this is inharmonic content that doesn't sounds good. The sound becames less defined, more blurred, foggy...

Imagine you play the A key on a piano, but unfortunately you push at the same time a few keys next to the A that sounds at a lower volume than the A note. This A note will be the louder, but less defined, and less pure than if you just pushed the A key and nothing more.
So, when you add inharmonic content to the signal, you are adding a lot of notes that don't sound good in this context, but at a lower volume. You probably not hear them, but the sound looses definition, although you don't know the reason.
You and me probably can't hear nothing of it when this content is below -80db (as you can see in the graphics), but when the inharmonic content is above -60 and you stack a lot of tracks, the sound becomes less defined and dirtier to our ears.

This inharmonic content doesn't exist in the analog world, you'll never find it in the hardware, just the harmonic content.

Adding inharmonic content to the original signal means a reduction of quality. In a entire mix whith lot of tracks, this means less definition in the final sound, because you are adding a lot of notes that are not harmonics with the real ones in the music.

Aliasing is distortion added by the plugin. This distortion ever sounds bad, we're not talking about "warm" or "analogue vibe". The cause is a bad and poor processing and a sound quality loss.

Most of the plugins have very nice interfaces, some of them have a lot of features, but 90% have poor processing and they make your tracks TO SOUND WORSE, with less transparency because they're adding unpleasant distortion caused by aliasing.
Most of the time, we call this distortion "analog sound" or "color", but is just a matter of poor quality product.


In the next posts I'll upload some plots.
Because is very difficult to match the settings in all those plugins, I'll just do the test in different positions and different times. Gain reduction is always 6db.
All this measurements are made with native versions on Nuendo4 at 24 bit 44.1
Working at higher rates (88, 96, 192) the plugins work much more accurately and the looks cleaner than in 44,1.

Sorry for my english. I hope you'll understand this is my third language.

Look at the original 1khz sine

[pacorro]

Waves analog button.
There is NO SIGNAL passing though the plugins. Project is in stop mode.

[pacorro]

Waves SSL shows a lot of aliasing, even at slow times.

Waves Api2500 is a clean and high quality comp free of aliasing

[pacorro]

Density (free)

[pacorro]

Elysia Mpressor

Noisy, even with no signal passing though it.

[pacorro]

Fab Filter Pro C

[pacorro]

Softube FET

[pacorro]

Softube Tubetech Cl1b
Very very nice interface, but a lot of aliasing

[pacorro]

I really can't understand this comp, so I tested a few positions. Impossible to me to find a clean setting.

[pacorro]

Psp Oltimer, the most clean and accurate comp in the whole market

Look at these plots.
In the faster position, Psp Oltimer has 0'1 ms attack and 0'5 ms release. One of the fastest comps, and totally transparent with the original signal