I 💗 NY - New York Compression Explored
Updated: Sep 8, 2022
New York- style compression is a popular audio engineering technique made popular in that city in the '90s, still employed today for its effectiveness.
Basically it works in this way: in the analog domain you would send a copy of the audio channel you want to process to a bus, where you would eventually put in an heavy amount of compression to squash the signal really hot, so that you can then blend in and balance the processed version with the dry signal with the bus fader; in this way you are keeping both the naturality and dynamics of the original sound AND the power and energy coming from the compressed version.
Sounds cool, right?
Indeed it's a simple yet very effective technique that is based upon the concept of parallel processing, where you can literally blend multiple copies of the same signal with complementary sonic characteristics to achieve a result usually unachievable in other ways.
In this first tutorial and blog post of an upcoming series covering Pro Producer's Tips , I'm gonna use Ableton Live 11 (you can follow along with the same concepts in any DAW anyway) and the pretty 'I Heart NY' plugin by Baby Audio, which was distributed for free last year but you can still grab a copy of it for free in the latest issue of Computer Music Magazine, n° 303, to explore some common and more creative uses of the NY Compression Technique.
So without further ado.. Let's get started!
Back To Basics
Open a fresh project in your DAW of choice, import an audio file to an empty channel, then add in 'I Heart NY' plugin from Baby Audio as an insert. Indeed, the cool thing about this plugin is that it simplifies the routing and workflow of the technique explained before, so that you can apply straight up the processing into the channel without the hassle of setting up a bus.
Here is the signal flow scheme as found in the manual of the plugin:
As you can see a copy of the signal is routed directly to the output (this will be the dry, unprocessed version) while the other follows a parallel chain that goes through the compression stage, the amount of which is controlled by the single 'SPANK' knob and its level via the vertical slider at the center of the plugin, before being mixed again in with the dry signal and the overall result (dry + wet) balanced with the 'OUTPUT' knob:
If it sounds confusing or complicated, the aural examples provided below will clarify this:
As you can hear by listening carefully, after the processing applied, the acid loop in the first example and the acoustic guitar in the second sound fuller, more present and energetic, which was our goal afterall.
By applying parallel compression we have increased the RMS value and made it louder, even after having compensated with the output knob for a gain match so that it's a fair A/B comparison.
In the acid loop example I had access to the individual stems - drums, bass, lead - and I applied a different instance and settings of the plugin to each one so that I could have more control over the end result. Also I applied an instance of it on the bus group of them as a regular style bus compressor: with the SOLO button engaged in the plugin, the dry signal is disabled, so we only hear the processed compressed version like a standard compressor (no parallel processing applied in this case).
The result? The drums punch through more, have a more pumping groove/feeling and bite in the attack stage; the acid lead line sounds more upfront, squeaky and crispy; the bass, which had some ping-pong delay on it, sounds wider; the acoustic guitar has a warmer sound with the details of the fingers sliding between the frets more upfront: if you think about it we have achieved all of this just with parallel compression! That's the power of it, we could have achieved something on this direction experimenting with EQ, transient designers, swing on drums, widening plugins.. but we just used this single technique. How can we explain it?
Technically speaking compression increases the average volume by reducing spikes of the audio signal overcoming the threshold by an amount set with the ratio knob, speed of action set by the attack time and speed of recovery by the release time. The Baby Audio plugin has all of this under the hood, so different settings of the SPANK knob mean different ranges of action, from subtle to more extreme. In a regular compressor for instance it's the attack time that influences how transients will sound after processing, with longer settings having them more pronounced and shorter ones having them more under control. The release time is super important for the 'pumping' feeling: if set right it can enhance the groove or make the compression action sound more natural and discrete. Furthermore, with very high ratio settings together with short attack/release times you can get a very squashed and saturated result, because in these extreme cases you are altering the volume behaviour of the signal so fast that the waveform is altered consequently.
These are just some of the creative applications of compression, but the most important that you should keep in mind is that by raising the RMS value you are getting a fuller, more rounded sound because psychoacoustically the louder you get, the more flat our earing system becomes.
Compression is a non-linear process and indeed a complex tool where many parameters interact together; that's why its uses can vary wildly. I think the Baby Audio plugin is great if you want to get creative fast without getting bogged down by many technical details, that's why I chose it for educational purposes in this tutorial.
2. Stereo Trickery
'Till now we have applied the parallel compression technique in its standard and traditional way. Now, let's get more advanced with mid/side processing.
First, a little bit of theory:
the MID information is what is shared between the Left and Right channels, basically stereo collapsed to mono (we could write L + R to describe it);
the SIDE information is what is different between the Left and Right channels (we could write L - R in this case).
By having control on each of them we can get more interesting results with the technique we wanna apply. I used the free MSED (Mid Side Encoder-Decoder) plugin by Voxengo to separate the mid information from the side. I created an Audio Effect Rack you can download for free at the end of this page (continue reading!) to streamline the routing of this.
In the first of the following examples I applied a different instance of 'I Heart NY' to the Mid and Side channels of a techno drum loop, with more processing applied to the side to create the feeling of a wider stereo image. In the second example I gradually automated the fader controlling the parallel compressed volume of the side channel only, similarly to the previous example to create the illusion of the stereo image getting wider and wider and also more saturated:
3. Getting Into 'Geek' Territory
If you're still with me and have developed more curiosity and hunger of exploration, then we can get deeper and more experimental.
If you have understood the mid/side technique then this idea should be really clear: why not applying a different amount and settings of parallel compression to the left and right channel indipendently?
You would ask why on earth you would do such a thing.. Well, if you already have an audio file with a nice stereo image and various sounds panned, for example a drum loop, you could enhance it and get an interesting and complex interaction between the left and right channel, hypothetically even creating two different grooves interacting together. Sounds crazy, right?
I have created an other Audio Effect Rack for L/R splitting using the Utility plugin inside of Ableton Live for you to download at the end of the article.
For the final and more creative application of the NY compression technique I created a multiband splitting rack. Again, you can download it below and basically it splits low, mid and high bands with the Ableton's device Multiband Dynamics by soloing each band in parallel.
Then I inserted three instances of the Baby Audio plugin, one for each frequency splitted band.
In this way I could mangle the audio in a very deep way, but with a simple and fast workflow, very difficult to achieve with standard parallel compression routings. For the Low band I engaged the SOLO button to compress it fully and make it more compact, while the others had the parallel configuration for extra character and malleability. After that, I used the mid/side technique explained earlier to also shape the stereo information with our NY treatment.
The finishing touch was to apply an other instance of the plugin to the reverb return channel so that the reverb alone could be coloured and enhanced as well:
What I've not mentioned so far is that under the hood the Baby Audio plugin has an 'EQ-curve and adaptive side-chain filtering for extra tone-shaping', as written on the manual. What this means is that we don't have to care about equalizing the parallel processed signal to avoid low-end distortion for example, as you would in a typical scenario, because it was already addressed by the developer to achieve more clarity.
I truly hope you found this reading insightful and learnt something new. My ultimate goal was to inspire you and hopefully give you fresh ideas to try out with some out of the box thinking perspective. Feel free to drop a comment below if you liked it and please consider sharing it on your social media to support me for future blog posts. Also don't hesitate to subscribe to my newsletter and contact me for any feedback or suggestions about future blog posts.
Happy Music Making! 😉
P.S. Click the image below to download the three racks I used!