Thank you DarkRed, this feedback is GOLD! It is EXACTLY the kind of feedback I am looking for. Will definitely look into what I can do about these things. I have already thought about slightly pulling back the mix of the sidechain on the leads, so I might do that indeed. As for the limiter, it is just something I threw on in a hurry just in purpose of posting this draft online, the track is not yet mastered at all. Do you reckon I should push it a bit less on the limiter? I really wouldn't want my track too quiet. I know there are certain things I can do in my mix for some more loudness though, I was planning on going through and surgically EQing out some slightly harsh frequencies here and there.
And yes, I was struggling a bit to make a "fitting" chord progression in this track, considering that I started out with my melody here and not the chords. In regards to your suggestion: I wrote this track in F# minor, would you use the IV major chord still (is mode mixing what this is called?)? Or did I misunderstand?
Also I would really love if you could ellaborate slightly on the stereo image part, in terms of what I can do to improve that.
And yeah about the gain staging, I haven't really reached the point where I fine-tune that yet, but after all that was what my thread was initially about, I feel like this isn't my strongest point considering that I jsut do it so that it sounds good to my ears personally. So if you have suggestions for what I can turn down or up I would love that too.
And again: THANK YOU!!
I'm here to help, so I am delighted to hear I am able to provide some guidance.
When there are numb transients in the mix like a constant layer on the mix, it is typically due to a brickwall peak limiter that very quickly and aggressively attenuates the signal very frequently on the content, meaning the content is pretty dynamic at the current loudness level so it has a lot of peaks to brickwall peak limit. The solution is in this case not to lower the input signal to the limiter, because although that makes the limiter act less on the content and hence the numbness decreases, you now also introduce a gain staging issue. So that is a scenario of losing more than you gain.
The solution is to lower the peak to rms level of the input to the brickwall peak limiter, basically feeding it a thicker waveform and hence have it work less hard. This you can do by compressing the elements of the mix in the earlier stages more, meaning not by using a high ratio and fast attack, but rather allowing the content of the mix to "breathe" dynamically by having the ratio lower, the knee softer and the attack time longer. Once you hit the brickwall peak limiter with a thick signal that breathes dynamically, then the brickwall peak limiter can also act on the content in a less numbing way because you don't have to push it as hard to reach your goal.
Keep in mind that how you in a good way achieve this thick waveform I am talking about is actually going back all the way to recording, because it is much better to remove the dynamic issues during recording than having to deal with them later on. And in a pro situation you should not accept that kind of a recording to mix, especially if it is the vocals that are like that, it should go straight back to recording. But if for whatever reason that is not an option and you have the kind of content that is very demanding to thicken, meaning all mix elements jump in levels a lot, then it becomes incredibly important to resolve those issues in the first compression stage, meaning when you deal with the dynamics of the individual sound sources early on in the mixing process. Many approaches work, some better than others. You can for instance prepare for peak compression in this case by first compressing the signal in rms mode to stabilize the overall rms level, alternatively volume ride the signal, or use a volume rider fx. Then once the overall rms level is not jumping like crazy you can take care of the peaks using a compressor with certain dynamic characteristics that can handle those situations well, for instance an 1176 or a distressor, works great for instance on vocals and drums. So it is important to be aware that in order to be able to achieve a good thick signal to feed the brickwall peak limiter you sometimes need to work in two dimensions with the dynamic signal stabilization, both rms and peak, because else the compressors might attack certain parts of the input signal too much, you need to spread out when it acts on the content so that you can stay more gentle with the peaks and be able to apply more compression in total in these cases. And this is the case with all compression you apply, not only during the first stage of compression, although that is where it is needed the most. Naturally by spreading out the compression into stages, parallel processing etc. you are automatically making the compressors act on the content more evenly over time so that you can bring down the peak to rms and still have good perceived dynamics left.
When it comes to rounding the stereo image, like taking it from a "+ shape" to an "o shape", this is done similarly to how you deal with frequency overlapping by adjusting volume faders and EQ knobs in between tracks to pull together the right perception of the frequencies of each sound source in the context, similarly when you build the stereo image you work with the volume faders and pan knobs in between tracks to pull together the right perception of each sound source in the stereo field, you want to know what frequencies are played on each hit and what you need to do in terms of timing in order to form what is required to create your desired perceived stereo image, because also the stereo image is the result of a perception, whatever that creates a particular stereo perception is what is of importance. By roughly knowing the fundamental frequencies of each sound source and the fundamental frequencies of the sound sources on each hit, you can design the stereo image accordingly, even on a per song section basis. This is where delay and pre-delayed reverb comes in. If you for instance just have center panned and hard left/right pan positions, you are going to create a "+ shape" stereo image as long as the mix has depth left, when the depth is gone it is going to become a "- shape" stereo image. By introducing 50%L/50%R pan positions into your mix you are by default moving towards an "o shape" stereo image, but it has to be done in conjunction with the volume faders, because when you lower the volume of the sound source towards 50% of the available voltage, it moves towards its null position in the context, meaning towards the center of the stereo image (in 3D). So for the sound sources that form the "o shape", you need for the 50%L 50%R both a bit louder volumes to establish the perimeter at the back of the ears and a bit softer volume to establish the perimeter just outside of the eyes. You also need the EQs, since you need to position the element in terms of height too, when talking about the "o shape" stereo image in 3D. Then you can balance all of those to shape the desired stereo image more towards the "o shape". You can of course also create other forms of stereo images. When you use headphones and gain increase, the whole stereo image moves towards the center of the head on a specific playback level, so that now you get more and more encapsulated by the music, the perceived circumference of the stereo image increases. This also explains why a "+ shape" stereo image does not work so good on headphones, because many other mixes you listen to with headphones sound bigger and more present in the "o shape" when they are properly gain staged. The gain staging ensures enough information density, with that comes a more clear image of the stereo field. It also helps to separate the sound sources in the sound stage, although the sound stage becomes somewhat unrealistic when you are surrounded by the band. But music that encapsulates the listener, the listeners like. You can of course try to mimic the listener perspective of a live concert by focusing the reflection frequencies towards the up front half of the "o shape" (or the other way around if you want to mimic listening at the front row), but it is not necessarily that kind of perspective you want to create or like the most, but it can be a good way of learning how you can use the stereo image in the 3D context.
In my view the takeaway thing about this though is the height dimension, it is as important as the width of the stereo image. Many mixes out there are great, except the height dimension is totally messed up creating this very strange sounding stereo image, especially on headphones. And remember that you have both the height of the mix and the relative height of the sound sources within the 3D sound stage. This is important because it impacts on such things as gain staging. If you for instance position a hi-hat outside of and close to the ears, you are going to have to reduce the gain on it, this means now it sinks into the back of the mix and helps create a more blurry sounding mix. Be aware of what quiet high frequencies does to the perception of the clearity of a mix - makes it blurry. Now that does not mean you should do hi frequency boosts all of over the mix and on all individual sound sources to heighten it, you still must keep the frequency overlapping at a minimum, you just need to be aware that some sound sources have the information density so concentrated around a specific narrow frequency range, that if it gets blurry there, it's going to put a cost on the mix as a whole. Those sound sources must be able to compete well in their frequency context and enough gain ensures that.
When it comes to gain staging, ensure you maintain good gain staging end-to-end throughout the music creation process, supported by great monitoring. Personally I think it is a bit easier to relax the mix later on rather than pumping it up later on, again some monitoring solutions support that kind of approach, some do not. (because what you do not want is to have it pumped up and then when you listen to that on some other playback system it is like 500% too powerful, that's far from a good gain structure) If you are done with the mix and everything is perfect, except when you A/B your mix is 500% more in the back and inside of the content you have various elements randomly becoming 500% too dynamic on various locations, how can you fix that without having to redo the whole gain structure. It is a work of redoing the mix. So I want to push enough of the work with the dynamics and perceived resolution as early as possible, by creating high enough perceived loudness/resolution early enough. Clipping is not what I focus on, I know how I resolve that, ending up with a mix that is totally dead when I start to master because the gain structure is completely wrong, that is a bigger issue.
The fundamental frequency of the mix in combination with the integrated LUFS should be roughly optimal when the rough mix is ready, especially in the chorus. That in combination with good gain staging supported by great monitoring is key. Especially when you have a mix with hit potential, it's all about the rough mix.
I hope this helps.