Before & After Analysis = Awesome

We could all use less “hmm, i wonder what this is doing exactly?” and more “aha, i understand exactly what that is doing!” moments in our lives, am i right?

Well, perhaps you have everything sorted out with absolutely perfect mental clarity at all times… but nobody that i know of does. The built-in perceptual capabilities of human bodies have limitations. Consequently, methods we can use to uncover and keep track of finer layers of detail to our perceptual input than normally possible can be quite useful for stepping outside our our usual—highly subjective and somewhat amorphous—human frame of reference.

When i look back on the last few years of my audio production journey, the one software tool which has proven the most valuable (well other than Ableton Live, obviously) has been a particular analyzer plugin. This blog post is definitely going to be promoting that specific product, so i’d like to point out that i am not affiliated with the developer in any way and this post is not endorsed. i just… really love it.

The plugin in question is the IQ Analyser by HOFA.

Continue on for an brief overview of a couple of innovative qualities that make this tool stand out, interspersed with some philosophy about the general usage of spectral analysis.

The IQ Analyser has lots of awesome attributes, but what makes it particularly extra nifty to me are two particular traits: The “Energy” Spectrum and “Difference” Mode.

“Energy”

This is an analysis mode that provides the average spectrum curve of the audio content (either played through it in real time or dropped in for auto-scanning). The time length which is averaged can be any, from a brief one-shot sample to an hours-long field recording. I find energy mode to be more valuable than a typical peak spectrum histogram, since it somewhat “absorbs” individual occasional nuances that stray from the average contour, rather than keeping them blatantly poking out at full strength even if they only happened for a brief moment. I think that momentary transient peaks are less important than the areas where energy builds up over time through sustained tones, when it comes to listening fatigue

Respecting Tonal Variety
Music is more interesting when all songs don’t sound the same, so tonal variety is a good thing; therefore i’d rather my decisions be more based on the overall picture than on occasional moments that “stray from the ideal”, as those very moments often enhance the individual character of a musical piece. Also, if you’re making static EQ decisions which are applied to a lengthy time section but are based on reactions to brief momentary issues, you may be detrimentally affecting otherwise solid portions of time. Dynamic EQ or other types of multi-band dynamics can help address this by for example only curtailing a harsh frequency when it hits a certain dB, however, this also serves to reduce tonal dynamics by homogenizing the contour over time, which may not be ideal for retaining the heart of a song’s emotional dynamics.

Scanning the spectral energy of different songs and matching them to the same integrated lufs value is a great way to see how they compare to each other in a “big picture” sense. For example, if one song has a lot more upper mids at 4k compared to another, it will be obvious by looking at the loudness-matched curves.

Average Spectrum Energy feels like the spectrum analyzer display equivalent of what integrated LUFS is for level metering. Instead of just showing us the maximum peaks attained and a temporary fluctuating RMS value (neither of which provides a very good impression of the overall) we are given the whole picture, summed and averaged, which is extremely valuable when doing comparisons.

“Difference”

What sort of comparisons? Comparisons between different songs, comparisons between different audio loops or one-shots, comparisons between a piece of audio while it is clean vs after it has effects applied.

Super Nifty Modality
While you can, as with any analyzer plugin, open up two instances and compare them side-by-side, this particular plugin has a super-nifty mode which allows you to go beyond that, linking two instances together to show you the spectral difference between the first and second audio files. This works perfectly for comparing two energy contours to each other. You can also see the real time comparison between two peak or rms signals, which can be excellent for monitoring an analog signal path in real time. Note that the real time modes can appear a bit jittery (especially in the lower frequency regions) as the after “catches up” to the changes in the before. You can also use it to compare the left or mid portion of a single stereo file to its own right or side signal, an technical way to visualize imbalances in the stereo field.

For example, if both the “before” and the “after” analyzers are reading identical audio, the spectrum analyzer will show a flat line across 0dB for all frequencies. If, however, an EQ node were bumped up at certain frequency, you are shown the exact EQ bump that is being implemented.
In rms or peak mode, this will show the changes as it occurs (with some hiccuping as the instances talk to each other), while in energy mode you would want to pause and then restart playback to reset the readings (so as not to retain the previous, now obsolete, spectral info in the average).

Aha, so this is what the Carver Digital Time Lens is doing.
Kush Audio’s Omega A plugin at default settings, calibrated to unity gain. Interesting.
Code Red by Shattered Glass Audio: Real Time Before-After Analysis

A Journey of Discovery.
Before-After (aka “delta”) analysis is pretty invaluable for discovering frequency contours that analog gear or plugins are implementing which you may have noticed, but which were not explicitly described by the (naughty naughty) plugin documentation. Once you start analyzing your gear you’ll realize that this… happens… a lot. Maybe you find that the “special mojo” of your “secret recipe” processor for your vocal mic recordings turns out to have been primarily due to a particular frequency contour it always applies, and the reason it always sounded good is now revealed to your amazed cranium… and you are now equipped with the knowledge to replicate similarly beneficial results with other processors. Or etc.

The analysis is accomplished by placing the “before-and-after” template rack on an audio track in Ableton Live, and then placing effects in-between the two analyzers. Thus the first analyzer will be reading the pre-fx signal while the second analyzer will be reading the post-fx signal.

Keep in mind that the results will be based upon the audio; for example, an EQ boost applied to a frequency which doesn’t exist in an audio clip will not be shown or will be shown to a lesser degree than expected (as the EQ won’t have any—or will have less—frequencies to affect). For this reason, fullscale pink noise is useful for plugin analysis, as it captures anything that might occur and is sloped somewhat akin to human sensitivities. Before-After Analysis can be really interesting for seeing how the contour of nonlinear distortion-inducing processes (such as preamps, for example) changes with the gain. Some distortions change in tone wildly at different gain settings!

Caveat: some types effects may not register with the spectral difference display; such effects might have a distinct sonic impact that is obvious to the ears when heard, but is not as obvious by looking at the spectrum, as they don’t alter the spectral contour. An example of this might be gain-matched compression, limiting, or saturation.

For frequency-altering effects like EQ or multiband compression, before-and-after analysis can be great for gaining an insight into what exactly is happening to the sound which you certainly detect by listening, but which is more difficult to know with definite precision exactly what frequencies are being affected and how. This type of analysis is also preferable to comparing two graphs side-by-side for very small differences; with side-by-side the difference might not be easily detectable, but with before-and-after analysis, even a minuscule spectral alteration will be evident.

Freebie Download Time

I found myself scanning so many things with before-and-after analysis that i made a handy template set and racks to make the workflow almost instantaneous to do so. The template set is for being ultra-technical and finding out exactly what a plugin does when pitted against randomized LCR pink noise. The racks are for portability, able to drop into any Ableton Live set at any time to detect and display what difference is happening between A and B at any point in any audio chain, on any audio material.

!!! System Requirements !!!
Ableton Live 10
HOFA IQ Analyser VST
HOFA SYS PinkNoise VST (free)

The download zip includes…
Ableton Live Set: “Before-After Testing”. Used to test out plugins and gear to see how they affect the frequency spectrum using “before-and-after analysis” with pink noise.
Audio Effect Rack: “Before-After Analysis”. I suppose you could demo the plugin to check it out, if you don’t own it. I’m assuming most people reading this don’t own the plugin, as it’s pricier-than-average. But it’s sooooooo worth it… trust me.

“Testing” In Action. So that’s what Ableton’s Drum Buss does to the tone…

IQ Analyser—Before & After can be dropped into any Ableton Live track and used to assess what any plugin or outboard gear, or chain thereof, is doing to the spectral contour of the sound. Simply open up both GUIs, then minimize the first one (this is necessary to calibrate them), place and resize the second GUI where you like it (second monitor, mayhap?) Place the effects you want to analyze in the “drop” area between the two analyzers and hit “play”. You’re good to go!

To use the Testing Ableton Live set, make sure you have first grabbed and installed the free HOFA SYSTEM basic for its pink noise generator.

Then follow these steps.
•open the set in Live 10.
•open up the two IQ Analyser GUIs to calibrate them.
•close the first (“before”) GUI because we don’t need to look at it.
•place the effect(s) we want to test on the group track named “Test”.
•press the space bar to start testing. Make sure all three “pink” tracks show audio running through them. If only the center track shows audio, hit spacebar again. If only the left and right track, or none of the tracks, show audio, hit the p key (mapped to enable a pink noise generator).
•Adjust effect parameters, then double-tap spacebar to reset the energy reading. Now you have your spectral before-and-after difference, sweet!
• To Loudness Match (if desired): Take note of the LUFS value after it settles down. Apply output gain to the plugin to match -16LUFS, and double tap spacebar to reset. Now the total cumulative energy of the audio spectrum shown will be the same as for the dry version—excellent for objective comparisons.
•Repeat the “adjust effect parameters and reset” step as many times as you like, loudness matching each time or not as desired. See how different settings of processors affect your audio. It’s fun.

NOTE: the audio output level is set to -48dB to begin with. If you want to actually hear the pink noise (not that interesting usually, but can be for things that greatly affect the contour, such as guitar amp combos), slowly pull up the master fader. It is advised not to double-click the fader as you might get a blast of noise in your face, or possibly damage your speakers. There is a limiter placed on the final output (post-analysis) to protect from massive boosts, but this won’t protect you if you turn the master fader above zero.

If you want to test something in MONO, please LU-match to -19 instead of -16. To check a stereo source file in mono, you can enable the “true mono” rack that is pre-loaded on the “test” track, either by clicking its power button or by pressing the keyboard shortcut m.

Shoutout to audiocheck.net for pink noise test samples included in the pack, provided for educational purposes.

>>>DOWNLOAD “Before-After Testing” set and “Before-After Analysis” rack by clicking HERE<<<

or here.

I had initially intended to also briefly go over a bunch of other nifty features that the HOFA IQ Analyser has, but this entry is already relatively long, so i’ll save those perhaps for another post. Suffice to say, it has a bunch more cool features i didn’t even touch on here.

One thought on “Before & After Analysis = Awesome

  1. Pingback: Using LR Effects as MS (+ free Encoding/Decoding racks) | PerforModule

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s