Producers often underestimate the creative advantages of using delay units on return tracks. This technique seems to get obscured by the fact that most delays have an adjustable dry/wet parameter which, when on anything other than 100%, makes the unit a parallel effect. Nevertheless, this implementation is limiting. Dropping a delay in serial will not allow you to further process the delayed signal independently of the dry signal (outside of FX within the delay unit itself). Using return tracks instead solves this issue and opens up infinite possibilities. Take a look at the schematic below that lays out the signal flow of these two setups.







1) This first example is a recording from Loopmasters called ISeeTheSunlight_Dry with our delay unit inserted onto the signal chain. I was able to filter and modulate the wet signal a bit because our unit contains those features. However, any processing I do after the delay will affect the entire output, the mix of dry and wet. There is only one path for them to take to the output.


Vocals: insert (pseudo-parallel) delay


2) In this example, you’re hearing the dry and wet signals processed separately, like in the diagram to the right above. On the wet signal after the delay, I dropped an automated filter and a stereo imager called S1 from Waves. Furthermore, notice how the send/return setup has allowed me to deepen and widen the sound.


Vocals: parallel delay, further return processing




scem 2


In this example, you are hearing the dry signal as well as two wet signals which are hard-panned and processing individually. I used slightly different delay times on the two channels to simulate a realistic room (in an acoustic space, reflections from different surfaces take different amounts of time to get to you). I also did some automated filtering like in example 2. For this setup, you will need to create two return tracks. Then, send the source track to both of them and pan them hard left and right. Set the returns to “pre-mixer” mode. This mode will allow you to adjust the level of the dry signal without changing the amount of signal that is being sent to the returns – opening up even more possibilities for automation and balance.


Vocals: separate left and right delay processing




To conclude, it makes perfect sense to use delays as inserts. This is because they’re generally relatively CPU‑intensive effects — so it makes sense for multiple channels to share the same instance of a plug‑in. In addition to this, it comes in handy to use multiple source channels to share the same virtual ‘space’. You might also use this ‘send-multing’ approach to layer and sculpt sounds. In either case, you’ll find there are definite benefits to experimentation with those send effects, whether for tonal shaping or for problem solving.