DEALING WITH EXTREMES

The effects of many plug-ins may also be extremely frequency and level sensitive, so don’t get too hung up on what appears with one test signal – this may change significantly at other frequencies and other levels (try sweeping the test frequency for example). Also, remember to check frequency extremes when putting your plug-ins under the microscope. At the low end, look out for significant content at 20Hz and below (even DC may appear with a few plug-ins). You’re unlikely to hear this through your loudspeakers/headphones, but it will nevertheless show up in an analyser, and it could damage loudspeakers on large systems if ignored. If a plug-in is adding any significant content at 20Hz or below, put a high pass filter after the plug-in to remove it. On the other hand, don’t bother too much testing how plug-ins react to high levels of high frequencies – bear in mind that above 8kHz real world levels are likely to be at least 20dB down on the rest of the spectrum.

Overall, remember the popular audio maxim ‘if it sounds good, it is good’, and always use your ears rather than your eyes when choosing plug-ins. Yes, it may be tempting to follow the herd and buy or sell a plug-in because someone praises or rubbishes it online, but if your ears aren’t sure then a few one-off analyser tests on a particular plug-in may provide all the confirmation you need. Making music is a constant learning process, and it would be foolish to abandon a particular plug-in simply because we are abusing it in ways its developer never intended or subjecting it to tests that are highly unlikely to happen with music.

GO FORTH & MULTIPLY

Hardware or software gear can also be cascaded in series, such that each device adds a little to the overall effect. I’ve heard this technique called ‘sonic varnishing’ — adding layer upon layer of more subtle effects to hopefully achieve audio nirvana. This can, for instance, work well with compression, by letting each device in the chain tame a part of the sound; or with reverb, where you can end up with more complex and individual acoustic environments by combining different aspects from each. Additionally, the software plug-in can be used in a way that’s less feasible with hardware, and that’s to keep adding multiple instances to each and every one of your playback channels. For most of us this would be prohibitively expensive to do with hardware, but it’s the perfect opportunity to let software plug-ins shine.

The most obvious examples are the proliferation of ‘console’ plug-ins that add some ‘flavour’ to your mixes. These will typically roll off the frequency response at low and high extremes, add a dash of benign harmonic distortion for added richness (particularly at the low end), and while the difference may be extremely subtle on a single channel, once added to every channel, the sonic difference is more pronounced. There are lots of such plug-ins on offer, many of which recreate the entire control set of famous console designs, complete with sections for EQ, compression and so on, while some also offer other analogue subtleties: such as small random variations in frequency response from one instance of a console channel EQ to another, to mimic the inevitable spread of electronic component values that you get in real-world circuitry. This randomness can ‘flesh out’ electronic sounds in particular, and such variations between the left and right channels of a stereo EQ will convert a straight down the middle mono centre signal into a more fleshed-out central mono mass.

Nevertheless, you don’t necessarily need all this functionality to explore the essential analogue console ‘sound’. If you want to dip your toe in the console waters without spending any money, why not download Airwindows’ Channel7 (www.airwindows.com/channel7), which offers a choice of three basic console types along with a drive control that morphs from perfectly clean, through very pure saturation, to more obvious fatness. I’ve mentioned Chris Johnson of Airwindows before in this column, because his plug-ins often achieve subtle yet beneficial results due to novel algorithms of his own design. He deliberately avoids the slavish modelling of real-world audio artefacts in favour of novel combinations of his own mathematical operations, and, as always, the truth of the algorithm is in the listening.

THE NEXT STAGE

Another fascinating example from the Airwindows stable is the recent Interstage plug-in (www.airwindows.com/interstage), which aims to sound like running your audio through an ‘optimal analogue stage’. It offers no adjustable parameters at all, but performs its magic by reshaping the low end, similar to that of a capacitor-coupled circuit, while the highs runs into ‘active component electronic limits’, by ‘restricting treble slews based on the general amount of energy in the circuit’. Essentially you’ll hear no change if you don’t need what it’s doing, but Interstage will kick in when it detects excessively digital bass and treble in your audio, especially once audio levels are approaching 0dBFS. Once again, the results are subtle, but nevertheless fascinating, and when switching Interstage in and out blind, I’ve often found myself preferring it in circuit due to its goal of converting any overtly digital sounds into more analogue equivalents.

WINDING THINGS UP

Another aspect of audio components that can add cumulative mojo is the sound of transformers, and in the plug-in arena I’ve been impressed by Kazrog’s True Iron plug-in (kazrog.com), which models the euphonic effects of audio transformers. Unless you push the parameters quite hard for special saturated effects, the results are again very subtle, yet once again if you add a single instance of True Iron on each of your playback channels the cumulative results slowly build into a warm roundness that fits beautifully with many musical genres. True Iron can also knock the perfect edges off clean softsynth sounds for a little more grit and attitude.

One important aspect with all cumulative plug-in use is to be careful about gain-staging. All the euphonic effects described in this column will vary considerably with input level, so it’s important to spend some time discovering the ‘sweet spots’. After all, we’re only looking for subtle changes for each instance. I don’t necessarily add exactly the same treatment to every channel either, as some (bass instruments and drums for instance) may benefit from more transformer warmth, while others blossom with console tweaks. It can be a good idea, once you’ve got everything as you want, to render the song, then temporarily deactivate all the mojo plug-ins and render it again, so you can hear the difference properly and get your ears more finely attuned to these ‘analogue improvements’.

It always makes sense to keep plenty of free disk space on your drives, to provide working space and buffers. However, now that so many of us are relying on much smaller solid state drives to host Windows, it’s much easier to fill them up without really trying.

For instance, I recently ran into a ‘no drive space available’ error part way through a long and involved batch process, sample rate converting a huge collection of audio files. I was a bit surprised, especially since when I started this task I still had 8GB available on my solid state drive. Now it claimed to only have a miniscule 18MB available, insufficient to complete the task; which is why my audio application had bombed out.

CCLEANER IS BETTER

The first thing to do in such a scenario is to run some sort of cleanup routine. I still periodically use the freeware version of the CCleaner utility (www.ccleaner.com), which not only removes a wide variety of temporary files from your drives, but can also scan for and remove issues in your Windows Registry, manage any System Restore points, as well as seeing what programs get run each time you PC starts up (sometimes sneakily added by developers whether you want them or not), and what plug-ins have been set to run each time your Internet browser is launched. I ran this as usual, but it only found and then removed a few hundred megabytes of temporary files, mostly used by my browser. CCleaner can also analyse your drives in a basic fashion so you can see what’s taking up most space — the output is a simple bar chart with different colours displaying the proportion of your chosen drive devoted to Pictures, Music, Documents, Video, and Other files. This can be a really useful function, but for a musician it can still occasionally feel like you’re searching for a needle in a haystack. My next port of call was Windows’ own Disk Cleanup utility, which apart from the usual temporary files also searches for any Service Pack backup and Windows Update Cleanup files, lets you delete your most recent System Restore point (which CCleaner by default wisely ignores for system safety), and disposes of any Windows Error Reports. However, this didn’t really find much that CCleaner hadn’t already dealt with, so I had to explore somewhat more deeply than normal.

WINDIRSTAT TO THE RESCUE

Now at this point I reach for the free third-party utility WinDirStat (windirstat.net), which is described as a disk usage statistics viewer and cleanup tool for Windows. It’s known to run from Windows 95, through Windows 98, ME, NT4 2000, XP, Vista, 7, 8 and 8.1, and although the website doesn’t specifically mention Windows 10 there are plenty of people using it on this latest version, as well. Its opening graphic display doesn’t impress, but once you’ve chosen one or more drives in your PC it will scan through all the files and present its results in three main graphic areas. This can take a few minutes, especially for a large drive, but it’s well worth the wait. At top left there’s a familiar-looking directory tree displaying all the folders and subfolders on your drive, now sorted from biggest to smallest by how much space they take up. To the right of this is the extension list, which provides a neat breakdown and statistics for all the various file types (each of the most common is displayed in a different colour, along with totals for how much space has been used by each type of file and the percentage of the total).

Notwithstanding, it’s the third area across the bottom of the display that provides the real revelation. Its ‘treemap’ visualisation of your hard disk contents shows each file as a coloured rectangle whose onscreen area is proportional to its filesize. Folders also make up larger rectangles containing all of their files and subfolders, so their area is proportional to the size of the subtrees. Cushion shading is also used to highlight each different area of the overall structure, and you can zoom in and out at will. As often happens, a picture is worth a thousand words, so you can see a typical WinDirStat display at the top of this column.

THE WOOD FROM THE TREEMAP

The biggest single rectangle on my C:Windows drive belonged to the Windows Pagefile, followed by various Windows installer files, but the most significant proportion of my drive was taken up with DLL files (around 30%), System files (14%), and Applications (8%). However, this very clever visual mapping makes it far easier to spot large groups of similar files that may be taking up far more space than expected, and, sure enough, I soon spotted a huge chunk of temporary files belonging to Wavelab, which had been carrying out my batch sample rate conversion. I was able to click on this area in the treemap, open the appropriate folder in the standard Windows Explorer utility, and then delete the lot of them, which turned out to be some 7GB in total. While I was about it, I examined various other areas in the treemap that looked suspiciously large, and was able to recover 8GB in total on my SSD without breaking a sweat. Once any biggies have been dealt with, you can now try clicking on the TMP file extension in the extension list to see if there are any other temporary files lurking elsewhere – most can probably be safely deleted, gaining you yet more space on your drive.

CACHE ON DEMAND

It’s also a very useful exercise to find out where your chosen audio applications store temporary and cache files – many let you choose a folder location for these, or at the very least tell you where the defaults are. Even after I’d deleted the morass of temporary files following my sample rate batch catastrophe, peering through the Wavelab Options menu helped me discover a further 770MB of redundant cache data that could safely be deleted. The CCleaner utility lets you add custom files and folders to be removed, so you can in future achieve this cleanup automatically. Auto-save functions have saved most of us at some time or another, so make sure your audio apps have theirs enabled (typically you can adjust the auto-save interval, but every 10 minutes seems to be a good compromise). This may avoid you losing work in the event of a PC crash. It’s always worth archiving your completed audio projects too, as then you can safely dispose of all sorts of temporary files that may not have ended up being used in that project, plus auto-saved backup files that you no longer need. For a freeware utility, WinDirStat can be a revelation!

WAVES & ENVELOPES

The Oscilloscope function will already be familiar to most musicians, letting you view the test sinewave at the output of your plug-in to examine how it’s changed in shape, run into clipping, and so on. Watching the waveform change as you alter plug-in parameters such as Drive can also teach you an awful lot about their design and how best to use them. Dynamics is the perfect tab for exploring the action of compressors, expanders, limiters, and saturation plug-ins. Its Ramp display of input level against output level across a 100dB range means you can easily see at what input level compression/expansion starts to kick in as you change the threshold; whether it’s a hard or soft clip, and so on. Meanwhile, its Att/Rel option shows how the audio envelope responds over time, so you can see how much of your transients get through – it’s an educational feast!

Just a couple of days before I submitted this column, yet another update appeared for PluginDoctor, not only adding more refined options to some existing functions, but also adding a completely new page devoted to Hammerstein analysis, which can simultaneously display the frequency response/level of the first seven harmonics of your test signal – a great way to explore mojo saturation devices in even more detail. This 1.2.2 update also added a Store button to the Dynamics tab, plus very handy Min/Max/Step options that let you zoom in accurately to more closely explore the clipping area of so many plug-ins, which often only occur in the upper 10dB or so of their range.

INTERPRETING RESULTS

It does take a little knowledge to interpret some of the results, so a few beginners have in all good faith rubbished some aspect of a new release because they don’t fully understand how a particular feature functions, although thankfully such mistakes tend to be quickly corrected by more knowledgeable folk. On the other hand, since plug-in developers now know that many users can more closely examine how their products work, they tend to be even more on their toes to avoid releasing new products with bugs, since these are likely to be spotted and exposed on the internet within hours of release. Hopefully, it will also help ensure that marketing departments can no longer get away with exaggerated claims for new plug-ins, as we will be able to measure them for ourselves. Personally, I love PluginDoctor to bits!

SPEAKERS CORNER

Having got these improvements between my ears, I demoed Sonarworks’ Reference 4 Studio Edition Plug-in, which effectively performs the same function for loudspeakers as its headphone counterpart — following a calibration process using the associated Reference 4 Measure utility and a suitable omnidirectional microphone. I already had an old Radio Shack SPL Level Meter that I’d modded to extend its response down to around 20Hz, so I used that to start with. I would rarely recommend using any form of EQ to ‘flatten’ the response of a room without first installing some acoustic treatment, because otherwise you’re just EQ’ing the direct sound while still hearing unbalanced reflective sound bouncing off all the walls. In my case, I’d already installed some reasonable acoustic treatment (10 bass traps of various descriptions plus a large ceiling cloud, and had also done plenty of Room EQ Wizard (www.roomeqwizard.com) plots to minimise the peaks, dips and decays in my waterfall plots.

The Reference 4 Measure process proved simple to run through, and far easier to understand than other room correction utilities I’ve used in the past. It builds up its accuracy from a total of 37 measurement points around your listening position, all with on-screen visual feedback of where to move your microphone next. At the end of this sequence, Measure presents you with a frequency response plot of your loudspeakers in situ, and then generates a corresponding filter profile to flatten everything out.

The frequencies of my overloud string bass notes corresponded with a couple of measured dips at 100Hz and 300Hz, and with the correction curve in place the required loudspeaker mixing tweaks now became obvious. Measure also noticeably improved my stereo imaging, which had previously been a little lop-sided due to a large window on one side of my studio, leading to various differences in left and right channel playback response between 100Hz and 1kHz. Frankly, I was shocked at the measurements, since they also largely explained why I’d missed the 10kHz air EQ in my mix – with a 6dB peak centred around 1kHz it’s hardly surprising that high-end subtlety was being masked.

CALIBRATED MIC

After some on-line investigation of just how hugely budget measurement mics can vary in response from unit to unit (there’s a sobering set of results at www.cross-spectrum.com/weblog/2009/07/), I subsequently ordered one of Sonarworks’ own XREF20 microphones, which are individually calibrated against an ANSI-certified measurement microphone so their unique response is known and more accurately compensated for. Although the Tandy and XREF20 frequency responses proved to be similar up to around 1kHz (which is why my Tandy SPL meter had already improved my low end anomalies), above this they diverged significantly. I had already suspected that the Tandy response might be falling rapidly above 10kHz, so had already limited the Sonarworks correction curve to ignore this anomaly. Nevertheless, I was surprised that the Tandy mic response had such a large low-Q bell curve between 2-8kHz, peaking at around +4dB at 5kHz. Thank goodness I hadn’t attempted to compensate for this! The correction curve of the accurately calibrated Sonarworks XREF20 mic proved that my little studio was already reassuringly flat above 2kHz, requiring just +2dB shelving lift from 8kHz up, which restored the ‘air’ I’d previously been missing in my mixes.

My top end now sounds beautifully silky and flat, while the lows are now extended to -3dB @ 37Hz (pretty good for my tiny ATC SCM10 loudspeakers). I could switch to a flat response down to -3dB @ 27Hz by changing to the Sonarworks Aggressive low end response, but that would require up to a 9dB boost at 20Hz, and I really don’t think the extra 10dB of low end would benefit my music, or the longevity of my monitor speakers. However, the most remarkable improvement is that my stereo imaging is for the first time razor sharp. Acoustic treatment should always be your first port of call, but with the basics in place I can heartily recommend the Sonarworks Reference 4 Measure software, although switching to the XREF20 mic gave me improvements in high end smoothness and stereo imaging that were nothing short of remarkable. I am now able to hear further into my mixes and feel significantly more confident in my future mix decisions. That’s a result!