This is NOT an automated service. This is 100% human-controlled analog mastering using AI algorithms/tools to find the best setting for your song.
The mastering engineer has complete control over mastering process, AI only helps him find the perfect setting and make sure his opinions are correct (that he is not tired, makes any mistakes or has wrong subjective prejudices). Yes, people can make mistakes, they can be tired, have ear fatique, wrong prejudices / wrong or illusory view of the outside world / reality, or the brain can incorrectly decode input signals (for example, frequencies after long work). Even though mastering engineers are trained and have many years of experience, we are still humans. A.I., computers and machines are usually extremely accurate and correct, based on mathematical / scientific numbers, but they do not have feelings, creativity, an artistic view, ideas and imagination. We are not fans of automated mastering and we hate the sound of digital mastering in most genres. But why not combine both techniques, analog processing Human + AI + to achieve the best results, the best accuracy and efficiency of the mastering process?
Why not combine the power of both? AI + Man in analog mastering?
The best pilots of Boeing / Airbus or fighter aircraft also use various devices such as GPS, automatic systems (guidance / landing / measurement / gyroscopes / altimeters, etc.) to make your flight safer, more comfortable and work more accurate and easier. Please note that the classic compressor is also a stand-alone device, in fact the compressor is a very fast volume fader, nothing more, it behaves like a human volume fader moving up and down extremely fast, but it is much more accurate than a human one. Yes, one can move the volume fader to mimic a compressor, but who will make it more natural and accurate, human or machine (compressor)?
Another example of brain imperfection is … it probably happened to you after 3 or more hours of mixing, you imagined that the frequency spectrum of the song looked something like … and then you looked at a frequency / spectrum analyzer and you realized that you were wrong. Other examples of false prejudices are: well known “loudness is a better phenomenon”, psycho-acoustically louder recordings may sound better to many people, but when compared at the same volume levels, a louder recording usually has more distortion and problems, or another trick/illusion of mind is more expensive equipment sounds better, but when you do blind a / b tests, sometimes 3x more expensive gear sounds worse with some music material. AI / machine can effectively remove all the mentioned illusions of the brain or at least set the mirror of reality very quickly, because numbers / math are usually always correct and not deceived.
How it works?
After more than 17+ years of experience with DSP coding, we have developed a special A.I. algorithms that analyze your song in detail and take the following steps:
Analog AI suggestions – after deep analysis of your song our software show suggested settings in specialized plug-in and then engineer sets these values on the analog outboard, of course he can change values if he want – they are just suggestion.
A1 – AI Suggest ideal settings of analog compressor for best dynamics
A2 – AI Suggest best input/output levels of analog devices for best gain staging, saturation, rounding of transients etc. but also to avoid distortion, squashing of signal etc.
A3 – AI Suggest EQ settings for analog equalizer
Mastering engineer then manually set suggested values at analog outboard gear and decides if he likes the results + do his own fine-tuning / compare his settings with AI suggestion.
B1 – AI analyze your record and set correct EQ shape/EQ settings of DSP powered equalizer to balance bass/mids/highs, remove resonances or problematic frequencies in mix, same as human mastering engineer would do (but AI can create more detailed EQ curves in very small range of 5Hz steps if needed). AI can do this before and after analog chain if needed. This is just addition to analog equalization while analog excels in boosting bass, mids and trebles, digital equalization is good for chirurgical cuts and corrections.
B2 – AI analyze problematic harsh frequencies (harsh mids, harsh mid-highs), problematic muddy bass frequencies, resonant frequencies and removing them real-time, not throughout full song but only at various positions of song.
B3 – AI real-time removes some extreme peaks at selected frequency ranges at some parts of song (if needed) to avoid digital distortion, artifacts etc. and/or suggest ideal volume levels before/after analog chain and before/after digital plug-ins, limiter input etc.
The rest of the work is done by a human mastering technician with his hands and ears. The mastering engineer has complete control over mastering process, AI only helps him find the perfect setting and make sure his opinions are correct (that he is not tired, makes any mistakes or has wrong subjective prejudices). The mastering engineer uses AI tools as well as compressors, equalizers or limiters – he works with them in symbiosis.
Why AI + Human mastering?
We are not fans of fully automated mastering and we hate the sound of fully digital mastering because it produces sharp digital transients, harsh mid-trebles, often ugly digital artifacts, flat bass, etc. Automated AI masters have no feelings, they can’t feel art, emotions, don’t have creativity, visions. AI / Machine learning is still like a robot, precise thinking and accurate but not warm like humans. On the other hand, to be honest, people can make mistakes when mastering because of ear fatigue, prejudice, the fact that every brain decodes the world and audio-visual inputs a little differently and subjectively, or because of the well-known fact that the brain has something like an “automatic equalizer Inside”, which means that after more than 30 minutes of listening to one track, you are less sensitive to problems or frequency imbalances, the brain performs “auto-equalization”, so the song seems to be fine, while in reality have too much trebles mids or bass or some resonant frequencies. Therefore, pauses every 30 minutes and frequent comparisons with reference records are important during mixing / mastering.
so what are advantages of AI tools in classic human mastering?
– Time saving, usually correct manual alignment is 60-70% of the whole mastering process, our AI algorithms analyze your track and make their own curves before / after the analog string + suggest ideal analog EQ settings. This process usually saves more than 30 minutes of mastering time for the engineer, so it can save a lot of time and energy on more important sound details, engineer is less tired, has less ear fatigue and is better focused.
– Set the ideal volume levels, input / output from the AD / DA converter and individual analog devices such as compressors, equalizers, saturators, etc.
– Accuracy: numbers are always accurate and unchanging, clear rules apply in mathematics. The brain can be deceived in many ways, but scientific numbers can never be deceived
– Avoid engineer overall mind fatigue, ear fatigue, avoid wrong prejudices – an engineer often has to spend 40 minutes just leveling and doing a lot of experiments with only one song, but AI can do it in 2 minutes. It saves lot of time, energy and focusing ability.
– The engineer must check many positions / parts of the song to set the correct levels of AD / DA inputs / outputs, inputs / outputs of each analog device, inputs of digital plug-ins such as the limiter, etc. AI can analyze the track and set the correct volume levels in 2 minutes or less automatically, saving the engineer time and energy so he can focus on the more important details of the sound.
Benefits of AI + Human analog mastering for you:
– Better price and quality due to increased efficiency (technician has to do a lot of experiments and variations to find the ideal sound, but AI can suggest the best EQ / compressor / gain values after 2 minutes of analysis based on accurate scientific numbers)
– Safe and correct results thanks to mathematical calculations
– Constant quality without or with a minimum of human error