Easy Mastering

[laurend]

I know that most of the people here are doing self mastering for their tracks. Manufacturers now offer very powerfull tools to do the job. Mastering isn't only about tools. It requires high level skills that musicians and producers have simply no time to learn. As results, a lot of demos and released tracks are just slaughtered in name of the lousness war.
That's the reason I've designed an online service which can help music producers for reaching a competitive level without destroying their creations.

I've just put the version 2.0 of my processing online. It has a very clean, loud sound. Do not hesitate to test it. That's free and the job is ussally available in less than one hour.

I would be be happy to have your feedback on this preview here since the official release is officially announced on September first. Here's the official annoucement:

MaximalSound updates its online algorithm.

Since 2010, MaximalSound.com offers an online audio optimization service. Based on a proprietary algorithm, this service allows indie productions, small labels and musicians to release their productions with the best technical specifications. Because the processing uses the same pattern as the human hearing, it supports every music genres from metal to hip-hop, from classical to country music.

The limiter section and its interaction with the 32 bands de-expander have been totally reworked. This new version improves the micro-dynamic perception and the tonal balance while preserving the spacial enhancement and the original hot loudness. The safety margin has been increased in order to facilitate further lossy encoding for digital distribution platforms such as iTune, MySpace or Youtube.

MaximalSound 2.0 will be online on 09/01/2011 at 12.00 am (western Europe time)

Summary:

· State-of-the-art processing
· Unlimited number of free samples for approval before purchase
· Aggressive pricing policy based on the size of the processed files
· The job is generally available within one hour

More info at: http://www.maximalsound.com

[laurend]

How does it work?

[brettheaslewood]

legit or dubturbo?

[stompzi]

Without the human element, surely you might as well just stick 3dB of limiting on it? Or am I being super naive about how this thing works?

[laurend]

What do you mean?

[Teknicyde]

Why you pay for mastering ; so a good engineer will correct discrepancies in your mix in an acoustically treated environment, and will apply proper amounts of dynamic reduction to make given the tune at hand... when doing this, every bit of work will be totally unique to YOUR tune and YOUR context, caking the best possible product.

This thing ; uses the same formula for every tune to make it loud as fuck.

[laurend]

My algorithm doesn't have any education nor artistical sensibility. But it's adaptative to the signal. The sound is transformed in order to fit in a particular shape which corresponds to the human hearing system. It isn't supposed to correct all the flaws of a poor mix. But it improves correct and great mixes about the overall level and the tonal balance also.
Mastering engineers can fix some serious mix problems. But they can also request a better mix which the best solution for the sound quality.

[Teknicyde]

My algorithm doesn't have any education nor artistical sensibility. But it's adaptative to the signal. The sound is transformed in order to fit in a particular shape which corresponds to the human hearing system. It isn't supposed to correct all the flaws of a poor mix. But it improves correct and great mixes about the overall level and the tonal balance also.
Mastering engineers can fix some serious mix problems. But they can also request a better mix which the best solution for the sound quality.

It has an a-b eQ curve with a professional track you mean?

[laurend]

It has an a-b eQ curve with a professional track you mean?

a-b EQ is never relevant for serious audio jobs. Comparing 2 FFT analysis to generate a particular filter can't be representative of particular audio signature because it relies on the used time window. With the same reference track, the results will be different depending of the time window you used. Using 1 seconds, 10 secondes or a complete song will give different signatures.
My algorithm is based on studies on the Equal Loudness Contour. This is the only solution to obtain good results for all audio contents.

[Teknicyde]

It has an a-b eQ curve with a professional track you mean?

a-b EQ is never relevant for serious audio jobs. Comparing 2 FFT analysis to generate a particular filter can't be representative of particular audio signature because it relies on the used time window. Whith the same reference track, the results will be different depending of the time window you used. Using 1 seconds, 10 secondes or a complete song will give different signatures.
My algorithm is based on studies on the Equal Loudness Contour. This is the only solution to obtain good results for all audio contents.

I understand fletcher munson, hes a problem though - things dont belong at perfectly equal loudness, when we write music we make instrumentss quieter both perceptually and technically for creative reasons constantly.

Take 'mud' by loefah and make the mid of an equal perceived loudness as the sub... the tune will be thoroughly ruined.

(Im sorry if I seem critical, I just dont accept that mastering can be done by a machine, you need a good engineer.)

[laurend]

- The equal loudness curve tells to amplify the bass to make them as loud as mids.
- Is it a real drummer on your tracks, or is it machines?

You just have to click on my sinature to test the algorithm for real. I would be happy if you give me your feedback here.

[KONVEX]

Anyone here had a go with this? Seems like a cool idea, I just don't see how it works in reality :/

[laurend]

I just don't see how it works in reality :/

Here's the mastering chain.

Harmonic enhancer:
The harmonic enhancer, this old invention of the Aphex company, has the advantage on an equalizer to not apply a fixed amplification on the top of the audio spectrum. This technique allows to naturally clarify the sound without fatiguing constant high frequencies boost. In the MaximalSound algorithm the even and odd harmonics are generated separately using dynamic content signal depending on whether the attack or the body of the note are processed.


32 bands processing:
A 32 bands crossover filter splits the signal to be processed by 32 de-expanders before being rebuilt in broadband by summing all the bands individually processed. This produces a dynamic equalizer where each part of the audio spectrum is magnified in the low levels according to a psycho-acoustic model standard.



De-expander:
A de-expander can also be called reverse-expander. A de-expander amplifies the signal below a threshold in contrast to an expander which attenuates the signal below this threshold. This choice is fundamentally different than techniques related to compression. The signal is amplified when it is weak and not planed when it becomes too high (above threshold). The difference may seem subtle, but it is significant when the attack time of treatment is taken into account . A compressor will act always too late (attack time) on the peak values, forcing the limiter to affect the transients which are so valuable to the perception of stereo space. The de-expander just amplifies the signal after the attack time, avoiding any unnecessary limiting.

Limiter:
Unlike the analog domain, digital doesn't support any overload, even briefly. The limiter is there to handle situations where the summation of all processed bands produces a signal exceeding the allowed maximum value (0dB FS) in the digital domain. The limiter is an essential part of treatment since at this stage the signal is full-band and any flaw may lead to tonal unbalance by favoring or attenuating certain parts of the audio spectrum. In limiters the attack time is null or negative (look ahead) to prevent any violation of the limit value. The art is to find an optimum release time, in order to preserve the frequency content and the perception of the dynamics of the original signal regardless of the complexity of this signal.

I hope you have all the info you need.

[blinx]

Well i will give it a whack tonight when i get home. It sounds interesting enough and like any other technique im sure i can find its rightful place in a producers arsenal of tools.

[-[2]DAY_-]

de-expander is basically like maxxvolume by Waves, yes?

[laurend]

I don't know the exact design of the maxxvolume or if it does some upward compression.

[laurend]

Without the human element, surely you might as well just stick 3dB of limiting on it? Or am I being super naive about how this thing works?

That's more than 8 dB dB of gain. A simple limiter can't do this while improving the sound. All my clients have compared their own mastering to my online solution. MaximalSound won.

[sunny_b_uk]

id like to maybe hear some comparisons of original mixdowns of songs, then how it sounds after with the mastering effects added. im kind of sceptical at the moment lol.

[laurend]

id like to maybe hear some comparisons of original mixdowns of songs, then how it sounds after with the mastering effects added. im kind of sceptical at the moment lol.

The source
Full quality file here.



The processed file using the MaximalSound 2.0 algorithm with a 4.62 dB pad in order to allow a direct A/B compare with the exact same RMS level.
Full quality file here.


Do not hesitate to test the algorithm with your own files. That's fast and free.
You could also have a big surprise when comparing your current mastering chain with the MaximalSound treatment.

[-[2]DAY_-]

it's maxxvolume... but probably a lot cheaper.. Sounds good.