• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

My Custom Loudness Control Tools

grisys

New Member
Joined
Jul 30, 2021
Messages
3
Likes
20
Hello Audio Science Review Community,

Today, I am excited to introduce projects I have been working on for two weeks: 'FIR-Filter-Maker-for-Equal-Loudness' and 'APO-Loudness'. These projects are in their early stages and are far from perfect, but they are crafted with my passion for audio and a desire to contribute to our community.

'FIR-Filter-Maker-for-Equal-Loudness' is a Python script that generates Finite Impulse Response (FIR) filters based on the ISO 226:2003 equal-loudness contours. These filters are designed to maintain a balanced tone at lower volume levels and can be used with Equalizer APO, APO-Loudness, or other convolution hosts.

'APO-Loudness' is a Windows application that utilizes these filters to control loudness via Equalizer APO. It offers a simple interface for users to adjust the volume and target loudness levels with ease.

ApoLoudness_BCva9pvB3A.png


While they might still be a bit rough around the edges, these projects are open-sourced, allowing you the freedom to enhance and adapt them as you see fit. I warmly invite you to explore these projects on GitHub and give them a shot. Here are the links to the projects:
Thank you for your time and interest.

Grisys
 
Hi grisys, very cool work and very nice sharing!

With Equalizer APO, how does the FIR file swap out between Loudness contours work?
Any switchover delay? Or is switchover inaudible, other than the new tonality?
 
Hi grisys, very cool work and very nice sharing!

With Equalizer APO, how does the FIR file swap out between Loudness contours work?
Any switchover delay? Or is switchover inaudible, other than the new tonality?

1. The script calculates the SPL delta for every pair of equal loudness contours that might be used, creating 2761. If necessary, the script can be modified to calculate the desired SPL delta. (https://github.com/andrewjhunt/equal-loudness/tree/main#spl-deltas) A decibel meter is required for accurate use. I was surprised at how good the Apple Watch is.
2. There is a switchover inaudible of about 0.7 seconds.
 
Last edited:
ohh boy. Yesterday I created a complicated structure of config files in EqAPO to switch between compensation curves more easily. I just thought, "let me share this here". which topic? "let me search for a topic".
I will try this out and give you feedback
 
1702728351241.png


it sits at -13.9dB because of normalization.

1702728393897.png


I sugest sugesting a pink noise at -14LUFS. a true pink noise on Youtube would be at that level. There are a lot of fake/wrong pink noises on Youtube though.
 
The script automatically generates 400 curves between 60.0 and 100.0 phon based on the ISO 226:2003 standard.

guess this should read "60.0 and 90.0"?
Personally, I would like to generate/use 60 to 100 or 50 to 90 since I use -35dB compensation sometimes. my system gets very loud and I use a flat target for my room EQ (try this out btw).

will wait for a reply on that one before actually trying it out
 
Hello Audio Science Review Community,

Today, I am excited to introduce projects I have been working on for two weeks: 'FIR-Filter-Maker-for-Equal-Loudness' and 'APO-Loudness'. These projects are in their early stages and are far from perfect, but they are crafted with my passion for audio and a desire to contribute to our community.

'FIR-Filter-Maker-for-Equal-Loudness' is a Python script that generates Finite Impulse Response (FIR) filters based on the ISO 226:2003 equal-loudness contours. These filters are designed to maintain a balanced tone at lower volume levels and can be used with Equalizer APO, APO-Loudness, or other convolution hosts.

'APO-Loudness' is a Windows application that utilizes these filters to control loudness via Equalizer APO. It offers a simple interface for users to adjust the volume and target loudness levels with ease.

View attachment 330577

While they might still be a bit rough around the edges, these projects are open-sourced, allowing you the freedom to enhance and adapt them as you see fit. I warmly invite you to explore these projects on GitHub and give them a shot. Here are the links to the projects:
Thank you for your time and interest.

Grisys
Incredible work! Does it work with Roon?
 
Just some silly questions by the way:
- Is music mixed in such a way that it sounds good at a certain volume?
- At what volume are the blind mono listening tests done at Harman?
 
Just some silly questions by the way:
- Is music mixed in such a way that it sounds good at a certain volume?
- At what volume are the blind mono listening tests done at Harman?
1. Somewhat yes: 85 dB SPL for the average RMS level for stereo music. This is used in the K-system by Bob Katz that many engineers use, or so I read online. To be more precise: you said mixed. I know many mixing engineers, I've never seen it K-system used in real life, and I've never been in a studio where 85 dB SPL was the average working level (always lower). But it can be entirely true that a lot of music has been mastered to sound best at 85 dB SPL. 85 dB SPL is also the agreed standard in film.
 
Just some silly questions by the way:
- Is music mixed in such a way that it sounds good at a certain volume?
From an interview with grammy award winner John Hanes:

"Listening levels: about 85dB for critical listening. 95dB is loud listening for me to feel it, 65-70dB is softer listening for longer periods of time for basic levels, learning the song, editing, printing parts. Almost never doing any mono listening, but I am looking (correlation meter) for and listening for phase issues.

Headphones for bass checks, listening for noise, another perspective. We designed and installed the room treatments ourselves. We have not had them tested and calibrated. Working almost exclusively on the nearfields and having long experience in the rooms with them is the key to consistency."

- At what volume are the blind mono listening tests done at Harman?
I realise now this wasn't your question, but the Harman headphone stuff was done at 82dB SPL
 
I thought about an experiment in which you have to take the volume precisely into account. Hence all the questions about volume levels:
The Experiment:
Since the Harman headphone target curve is indeed a target, it would be interesting if the task for a test listener would be to tune a loudspeaker in a room with the help of an equalizer so that it sounds exactly like an absolutely neutral headphone.
What would the room curve look like then?
 
I am curious about the differences between this tool and the built-in Loudness correction in APO.
 
I am curious about the differences between this tool and the built-in Loudness correction in APO.
 
Back
Top Bottom