When the restoration project for my first album has to be paused for a while

One of the realities I have to accept when curating my music (especially my first album) is when I have to put it on pause. Beyond that artistic world, I’m a person with responsibilities. However, since I have complete control over that world, I’ve decided to sacrifice speed and take an indefinite amount of time to ensure a job well done, the best possible legacy I can leave for future listeners.

To be honest, I didn’t think this way several years ago, when I was writing these songs. I seriously doubt that any emerging musician at that age is aware of these issues, and the upside is that it removes restrictions on free creation. In this early stage, what matters is learning about and developing skills in performance and composition, not spending all that time on engineering, mixing, or mastering (unless an experienced third party helps, but that removes one key element: it wouldn't be music made 100% by a child or teenager).

In any case, the serious release of a studio album (especially following traditional industry standards, which are generally reflected in releases on physical formats and streaming platforms) requires that it be properly mixed and mastered. The potential for the best sound depends on how and with what the album was originally recorded. If the original mix wasn’t good (or never was made), the stems are very valuable and the solution is easier. The controversy begins when the stems either didn’t exist or were lost.

In the case of “Hernán Reyes Vocal,” there are original stems for almost everything recorded in 2020 and 2021, but the rest (and especially the material from 2018 and 2019) was never saved that way. There wasn’t even any volume control. What could I do about it?

I was amazed with the results of the first half of the album, but when I tried working directly on the older original files for the other half, it was impossible to enhance the experience and maintain the overall quality. I kept telling myself, “I just need to separate the vocals from each instrument…” But of course! The instruments were nothing more than soundfonts playing without quantization, but anyway, you could replicate every note and it would sound practically the same.

In my personal opinion, AI stem separation is a tool that should be used in moderation: I couldn’t mix using all the stems it provided, since there are inevitably artifacts and elements that I know weren’t separated correctly, especially when I was the creator and know how it works inside. However, it’s a missed opportunity when you discard the isolated vocals, especially when these are better than how the original sounded. I don’t think you can claim high fidelity if you were just using a microphone app that recorded in MP3 and captured a lot of wind and ambient noise that wasn’t supposed to be there. As long as there aren’t any annoying artifacts or missing syllables, it’s a useful tool. There’s even more manual work to be done: reviewing each phrase to clean up the audio during silences and correct specific instances of clipping.

In a previous blog post, I explained how I’m approaching the restoration of digital instruments and what I still have left to complete the project. Today, I wanted to talk about those times when you can’t make quick progress (depending on external factors, which are always present). When I say it took me, for example, half a year to finish it, it doesn’t mean I spent four hours every day working on it. I might have worked like crazy for two weeks, but then had to wait another three and tackle other, more tedious aspects calmly, and so on until it was finished. The first goal is that I’m completely satisfied and no longer feel the need to correct the mix ever again, at least not for that particular UPC. Any other modification would be a remix (unnecessary if it isn’t transformative enough compared to the original release and doesn’t have its own distinct direction).

This whole experience is both complex and fascinating, as it puts various strengths to the test and can extend to other areas unrelated to music, but which are also personal goals rather than just obligations.