DDMF Plugindoctor by Sound-Guy
Plugindoctor from DDMF
DDMF make a number of fine audio processors and utility plugins, but I found only the clever Metaplugin utility reviewed on Gearslutz so far. This is surprising considering that I often use DDMF compressor and limiter plugins, particularly their NoLimits lookahead limiter and the NYCompressor, and they recently introduced a fine “high end” variable-mu compressor, the MagicDeathEye. I may get to reviewing some of those sometime soon, but in this review I’ll show a useful utility that has recently been updated with some great new features.
What is It?
The Plugindoctor is not a magic cure for your sick plugins, however, it may help figure out what is wrong if a plugin doesn’t work right. It enables you to measure in extreme detail how a signal is processed by an EQ, compressor, console emulation or distortion plugin. Since I test and review a lot of software, I’ve found Plugindoctor indispensable to analyze qualities like frequency and phase response, harmonic (and non-harmonic) distortion, and the dynamic response of compressors and limiters. The latest version has really expanded its capabilities.
The Plugindoctor is not a plugin – it’s a standalone program that analyzes VST, VST3 and (for OSX) AU plugins. It comes in both 32 bit and 64 bit versions, and the latest release (1.4.7 as I write this) includes the ability to load two plugins at the same time for comparative analysis. And it’s not just useful for software audio processors – if you have a good bidirectional A/D-D/A interface you can actually use Plugindoctor to evaluate hardware. So if you have a real 1176 compressor and some software emulations of it, you can compare them in great detail.
What Can It Do?
You can load a specific plugin either directly from a disk drive or via Plugindoctor’s very handy built-in plugin browser that can scan a folder to set up its own list of available plugins. In the current version you can load either one or two plugins, even a VST and VST3 at the same time (I have no way to test the AU format), to see how they compare. Its main menu has six test types: Linear Analysis, Harmonic Analysis, Hammerstein (really!), Oscilloscope, Dynamics and Performance.
Linear uses a “delta-peak” signal to measure frequency and phase response. It can be informative to see how different EQs actually process a signal over the frequency range, and the phase response can show if your “linear phase” device is actually linear (no phase shift) and how phase shift occurs in minimum phase systems and hybrid designs. You can even view an impulse response, how the device under test (DUT) responds to a very fast pulse (how much does it overshoot, undershoot, pre-ring and post-ring?).
There is a camera icon that takes a snapshot of the Plugindoctor display, opening a file browser to save a
PNG file (which I used for this review). And a handy “Store” button that will freeze a displayed curve which will continue to be displayed when you change parameters of the plugin under test or load a different plugin.
T-RackS 5 Mastering EQ enables separate settings for left-right or mid-sides audio signals.
Note that the frequency and phase plots are made for each channel of a stereo device and typically plot as a single line if they are identical, but if there are differences in the channels, typical with real hardware and some “authentic” plugins, you will see the left plotted in cyan and the right in a pink color.
Phase response of the T-RackS 5 Mastering EQ settings shown in the frequency plot.
Harmonic Analysis includes a traditional single frequency mode (usually 1 kHz, but you can dial in any fundamental frequency you want) and a swept frequency mode. Using the single frequency mode you can visually see integer (and non-integer) harmonics, including aliasing. I’ve used Plugindoctor to determine that some plugins I’ve bought (and some free ones) that had some “oddness” to their sound have rather disturbing levels of aliasing and I’ve retired these from use. Note that harmonics and aliasing may be viewed down to -200 dBFS and to be fair, any lower than -125 dBFS are not going to be detectable even by those self-proclaimed, snobbish “golden ears” whose articles you can read in some (hilarious) hi-fi magazines.
In addition to the graphic display there are two measures available of the distortion generated by the plugin shown in the upper right corner of the graph: THD and THD+N. THD is Total Harmonic Distortion and includes only integer harmonics while THD+N is Total Harmonic Distortion plus Noise and includes any spurious harmonics, aliasing and noise. These are standard measures you will often see listed as percentages for amplifiers and other electronic gear. Here they are listed as dB levels which is actually preferable – and in case you’re not familiar with how dB relates to percentages, -40 dB is 1%, -60 dB is 0.1%, and 80 dB is 0.01%. Of course if you are testing a distortion plugin or an emulation of an analog device that adds harmonic distortion you may likely see (and want) levels well above 1%, even 10% (-20 dB) or more.
Harmonic Analysis includes IMD (Inter-Modulation Distortion) that uses two frequencies simultaneously, typically 60 Hz and 7,000 Hz, and if there is intermodular distortion you will see several peaks at occurring at 7,000 +/- N*60 Hz, where N is an integer. Ideally there will be no such peaks, but both hardware and software will usually show some level of these. Again, if lower than -125 dBFS you’ll never hear them, but I’ve seen intermodulation peaks with some gear sticking up to -40 dBFS or higher, which is not a good sign!
Single frequency harmonic distortion plot at 1 kHz of Sonimus Burnley 73 Preamp/EQ with moderate drive level.
Intermodulation plot of unnamed plugin with extreme distortion – might be useful if you want to really destroy a sound.
Harmonic Analysis can also be run using a sweep mode by toggling the "sweep" button. In this case the response of the tested plugin is measured over a range of excitation frequencies and THD is plotted against frequency up to ¼ of the sample rate – this limit is because the calculated THD using this method approaches zero as the sweep frequency approaches ¼ of the sample rate.
Rodgers and Hammerstein?
Hammerstein may be a famous name in American music (think Broadway musicals) but this Hammerstein is an alternate way to view harmonic distortion. The plot shows an estimate of the harmonic level versus frequency for each integer harmonic from the second (twice the input signal frequency) up to the seventh – you can set the max number of harmonics to view. This mode lets you easily see how even and odd harmonics compare over the audio range – and using the Store button you can compare different plugins although running too many plots at once can get confusing! Note this is very different from the THD sweep mode mentioned above that estimates total distortion of all harmonics versus frequency – the Hammerstein shows each harmonic’s separate contribution over frequency.
Sonimus Burnely 73 again with a high drive level – LF distortion is mainly 2nd harmonic, but 3rd predominates above 2 kHz.
The Oscilloscope is not terribly exciting, but if you’ve used an older version of Plugindoctor you’ll find the new one more useful if you want to view the signal output of the DUT. For one thing, if you use some kind of saturation plugin you can measure the distortion using one of the harmonic distortion views described above and use the oscilloscope to see what the waveform looks like for various levels of distortion. You can observe how greater levels of odd harmonics, particularly 3rd and 5th cause symmetrical distortion, squashing the tested sine wave vertically – and how even harmonics like 2nd and 4th cause lopsided results making the waveform asymmetrical. I’ve always been fascinated that even harmonics sound more musical even though the waveform looks more “distorted” than with odd harmonics (but from a musical standpoint the even harmonics are octaves, so it makes sense).
The Dynamics analysis is an excellent way to see how a compressor or expander affects signals versus input level and over time. In addition, tape machine emulations, amps and consoles may introduce compression. This plot is made for both left and right channels which you usually figure will provide the same compression or expansion, but this plot will show any differences if they exist.
A flexible compressor that can also expand (top pink plot) shown at three ratios. Product currently in beta test.
There is also an Attack/Release view that shows the output envelope over time with three adjustable signal level sections, nominally the input at -20 dBFS for one second, then at 0 dBFS for two seconds, followed by the original level for one second – but you can change these times and levels as you wish. While this won’t enable precise measurements of the times, you can see how different modes in a given compressor affect the shapes and general times. And you can use the zoom feature to expand the view if short attack or release times are used. The x and y axes will automatically scale when you change the times of the segment described above, and when you zoom in.
Attack and decay time plot for a plugin currently in beta test.
Finally there is a Performance graph tat shows the time spent in a plugin's processing callback as the number of samples the plugin is processing, a measure most useful to a software developer.
You can zoom in on all graphs except the Performance graph. Zooming is accomplished in two ways. First you can draw a rectangle with your mouse, starting in the upper left corner and moving the mouse towards the lower right. Releasing the mouse button will zoom in to the rectangle. You zoom out again by clicking and dragging the cursor towards the upper-left. You can also zoom in on the x or y axis using the mouse wheel: set the mouse cursor to the left of the graph to adjust the y-axis or at the bottom of the graph to adjust the x-axis. Scrolling the mouse wheel will move the graph along the desired axis, and scrolling while holding down the CTRL button will zoom in or out along the axis. To reset the graph you drag a rectangle from lower left to upper right corner using your mouse again.
There are several variations of analysis for each test type that I haven’t shown here, but all are described well in the user manual.
Plugindoctor is an excellent tool for anyone who needs (or wants) to analyze the detailed performance of any processor plugin. It has saved me loads of time over the years compared to setting up oscillators and spectrum displays, and can show parameters almost impossible to measure any other way. If you’re a technically-oriented engineer you’ll find Plugindoctor indispensable.
A wide range of measurements are available, some very difficult to accomplish manually.
New dual plugin test capability save time comparing devices
Store mode very handy to layer in multiple plots
Camera function excellent to save plots.
Handy plugin browser can scan folders for appropriate VST. VST3 and AU devices.
Not much to complain about – some plugins can cause problems (can shut down Plugindoctor) due to some tricky coding issues, but since it’s not a device used in a project, it’s not a big deal and when this (rarely) occurs I just run the Plugindoctor again. DDMF have solved some similar issues in earlier versions and are glad to hear from any users if they see such issues.