Hi, guys.

I would like to share with you a tool that I have been developing for the past 3.5 years in open source. Instant Reference Mastering Tool called Matchering.

My repo: https://github.com/sergree/matchering

This is not a commercial project with a zero budget, except for me, 2 other people have been involved in the project for a while, and one of the HTC employees made a huge contribution to the project about 2 months ago.


I hope you or your colleagues will find a use for this.

 

mmm

I think match target to master is a wrong concept, the idea is based on a wrong assumption.

Please spend programming hours in something more useful.

 

What's wrong with this idea?

Some music producer made a song (target) and wants it to sound like Linkin Park - Numb (reference), for example.
My tool matches and masters it in a matter of seconds. Frequency response, stereo width, etc.

Does Ozone Match EQ have the wrong concept too?
Or this: REFERENCE

It is unlikely that our previous online service would have made 15 thousand automated masterings if it was the wrong concept.

 

You ever heard of the Golden Rule: 'If you want to sound like XY, you gotta be XY'?

IMHO, yes. It could help you to find some issues but I wouldn't let it make final decisions.

So you think because more or less people are doing something, it must be right? Debatable.

 

The mixes aren't the same. You are squeezing together audio to be in the same range but the content of the audio is not the same. It's like making a cat sound like an elephant, the frequencies and volume will be the same but the cat is going to sound like a loud distorted cat with too much bass, not an elephant. For this to work you need to balance every channel in a mix not try to equalize and compress the master. This is no different than if a human used Ozone and just squeezed the shit out of his bad sounding mix to make it loud.

 

Yes, I understand your point. This is very professional and correct.

But there are many hobbyists who can't afford professional $$$ mastering services or spend their time learning pro mastering plugins and techniques.

I know what I'm saying because I've been working in the EDM industry since 2006.

Perhaps this forum is not the right place to talk about this tool, sorry for that.

Yes, for such questions, I have a great FAQ section on GitHub on how to prepare your mix and choose the right reference track.

 

This forum is the right place, some member will find it useful, some don't and some will argue about it.
But obviously it's a good place to inform about your project.

Your program have a clear goal and it can be very useful for the ones who don't have time and or knowledge to do their own mastering.
It also can serve as a test, so I could take one of my tracks, use your program in it and see how it would sound in the mastering style of xxx then decide if if I want to go that way or not.
I find this quiet useful even if it's not perfect.

 

@Infextro Thanks for sharing your 2-year project with all of us. As @Olymoon stated, it can be useful for the ones who don't have the time or knowledge to do their own mastering. Therefore, do not get discourage by the comments of those that do not find it useful or to be the right concept. To the contrary, as a programmer you need to use criticism to make improvements. Do not give up!

 

This forum is the right place, please accept criticism.
Why not ask here:
"I'm a programmer what do you think could be an useful audio DSP project to develop?"

Regarding your project it recall me a guy who desperately wanted to looks like Michael Jackson.
Do you think the result is good enough?

 

Thank you, guys!

 

hi you could compile Matchering 2.0 to win 7 x64 ? its a third option for me..

 

Why should a hobbiest even think about it? It'd be like a beginner thinking about buying a Stradivari.

You've come to the right place. We've just different opinions, nothing wrong with it.
And I appreciate your modesty.

 

@Infextro :
Is Matchering different from Ozone Match EQ or IK MasterMatch?
What I mean is : does it bring something new or is it just for people who doesn't have those tools?

 

@Infextro hey..!!! i always like new concepts, what about making the try process easier..? people don't need to be a programmer to install and try your idea, simple to try that's my point because i'm programmer too and sometimes we forget about the end-user , but i like the concept is cool to have alternatives..!!! imma give it a deep test he he he.. peace

 

Sorry.. Matchering web app is very coupled with Docker and it's modern Desktop app supports only Win10.
It's possible to use Matchering python library via console on Windows 7 using https://www.anaconda.com/distribution/ and https://github.com/sergree/matchering-cli but its too nerd way i think

The process is much faster.
Using Ozone, you need to manually select the sections of tracks that will be captured (using which matching eq is configured), you need to listen these sections in real time. Moreover, ozone makes one matching eq for both channels.
My tool selects these parts automatically and analyzes the mid and side channels separately. this part of the algorithm gives a zero cost stereo width matching

Thanks
I tried to make the installation process as simple as possible. the user needs to install the Docker, enter a single command in the console - that's all. Docker in this case is the easiest way to add cross-platform functionality to a Python application that uses platform-dependent libraries such as numpy, scipy. of course, if it was a paid product, it should all be rewritten in c++ and made separate binaries for all platforms. but I only know Python and dsp.

 

Hi Infextro! Thx for the software. I think in principle it's a good solution for someone who is not so experienced for "manual" mastering.
Now I have a question! I'm doing a PhD and I will be using some of my findings to create a midi generating software. I would really like this software to be a VSTi, it's hopefully going to be helpful for other producers as well. The problem is that I only know Python. Do you know if there's a way to make VSTi-s that way? Any workarounds? Thanks!

 

@Infextro i was trying matchering, hummm what about the reverb..? for safety i will make the low end in mono, but interesting results. something is happening in the highs too, way too crunchy he he he but if the user can control the reverb, and the brighness, i beliee it will be cool to process something like effects or something i like a lot

 

Hi! I know only Python too and i don't make vst plugins, i think you need https://juce.com/ but its c++

thank you, i think reverb is more arrangement- or mixing- side, IMHO
Now it's just a track structure analyzer, a complex automatic mid-side matching equalizer, a matching amplifier, and a brickwall limiter basically.
with an idea: if you want to change the result, select a different reference (or edit it in some way with other software). no knobs or faders. this is a mental shift

 

Can the "analysis" stage be used separately to the "processing" stage?

 

thaaaankss a million BROTHER !!! love this tool