If you’re mastering and want it to sound better, try aligning the phase, expanding the transients on the side image, and upward compressing the mids using a parallel send. You can also do some experimental steps like normalizing a parallel track instead of using parallel compression.
Before we get started, keep in mind that these tips are in no particular order - some of them are a little experimental, so introduce them to your workflow with some discretion.
Before you begin to process your master, consider aligning the phase of the track using Izotopes RX plugin. If the phase rotation of your mix is off, it won’t sound bad; however, it can cause the computer to misread the samples at a high level, resulting in unwanted clipping.
Within RX, use the Phase modulation, click 'Suggest' to measure the phase rotation, then select Render to change the phase rotation. Nothing about the overall sound of the mix will change, just the alignment of the samples.
Recently Eventide released an EQ with which you can affect the transients of your side image - but if you don’t want to purchase this plugin, there is a free alternative. First, duplicate your master, and place the free Voxengo MSED plugin on both signals, then solo the Mid and Side image.
Now we can affect or mid and side separately - which is useful in and of itself. I’m going to use a transient expander on my side image, to make my stereo image more detailed.
Although SplitEQ by eventide is more convenient than this, this method results in less phase cancellation.
I tried this on a master recently, and it worked well at bringing the mids to the forefront - first I created a bus/auxiliary track for my mix. Then I introduced a low-latency linear phase EQ, and isolated the mid frequencies on this parallel track, followed by upward compression.
This way I can isolate my mids, control their dynamics from the noise floor up, and increase their overall detail and perceivability.
If you have Soothe 2 and the Gullfoss EQ, try them with very subtle settings at the beginning of your mastering chain. Soothe 2 will attenuate resonance frequencies, and then Gullfoss will dynamically alter the frequency response to balance the mix, making masked frequencies easier to hear.
Even if you choose to remove these from your chain, which I find I do about 50% of the time, they help you hear the mix from a different perspective.
Although this is a little unorthodox, a parallel normalized track is a good alternative to parallel compression. First I’ll duplicate my mix and cut it into sections based on the peak level of that section - this usually means the verse and choruses will be separated; then I’ll use peak normalization.
Now all sections of the track have the same peak level, resulting in controlled dynamics without the need for compression. I’ll blend the normalized track in with the original, just like I would a parallel compressed track.
In Logic Pro, I can create 2 signal chains and then A B almost instantly - first I’ll double-click the track, and under Functions, I’ll pick Selection-Based processing. If I saved my chain I can load them in here, or I can select whatever processors I’d like.
By clicking A and B I can now quickly preview the 2 chains, making it easier to find what sounds best for the mix.
Let’s again create 2 tracks and then isolate our verse and chorus, or maybe chorus and bridge depending on the song structure, and then process the unique parts of the song differently. For the most part, I’ll keep things the same for consistency, but add some small changes.
For example, I could make the verse have a little less distortion, or I could amplify the side image on the chorus to make it more impressive. A similar thing can be accomplished with automation, but this gives you more control and makes the differences easier to keep in mind when making changes.