• 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!

How does Qobuz measure loudness for level normalization: not following EBU R128 standard?

Prius Factus

New Member
Joined
Feb 18, 2022
Messages
3
Likes
1
Location
Germany
This is my first post in this forum, so hello to everybody!

I am using Audirvana Studio for playing my local music and streaming from Qobuz. Audirvana makes it easy to see how much boost/attenuation Qobuz applies to each song to level out inter-track loudness differences (of course only if you choose to use that feature). Strangely, in some cases this replay gain seem to differ quite a lot from loudness leveling derived from industry standards (i.e. measuring the integrated loudness as defined by EBU R128 / ITU-R BS.1770 for each track and taking the difference to some reference loudness as replay gain).

I thought it might be interesting for you technically knowledgeable people to read about that in more detail, so the following is a re-post from what I wrote at the Audirvana forum.

=========================

The latest AS version 1.11 has introduced the option to use Qobuz’s replay gain values to level loudness differences between tracks. However, with some albums/tracks there are significant differences between the gain computed by Audirvana’s built-in replay gain analysis tool and the values provided by Qobuz. Such tracks, downloaded from Qobuz and loudness-analyzed by Audirvana, result in replay gain values that can differ from Qobuz values by up to 10 dB for otherwise seemingly identical streams – that’s huge! The good news is that for music listeners preferring the most popular genres (with limited dynamic range) this is not relevant in practice. So if that is you you won't miss much if you stop reading here.

Replay gain is calculated by Audirvana in compliance with industry standards (EBU R128 / ITU-R BS.1770). Apparently Audirvana uses a reference loudness of -18 LUFS (I cross-checked with other EBU compatible analyzer software and can confirm this). Qobuz’s gain values tend to differ from this standard if music has a high dynamic range, typically with a DR value of 12 or more. Any music with DR < 10 will result in gain values practically identical to standard.

Out of curiosity I checked 10 albums with more than 100 tracks in total. It cannot be excluded with certainty that some of the analyzed local albums have been changed in Qobuz’s repository since I downloaded them, e.g. were replaced by a re-mastered version (or downloaded files could be different from their streamed version). However, I think this is rather unlikely (for reasons I will not elaborate on here).

As examples here are two pop albums, significantly compressed in line with current genre standards and in consequence with small DR indices. Audirvana and Qobuz state the same gains for these tracks.

Happier Than Ever_ Billie Eilish.jpg
Californian Soil_ London Grammar.jpg


Moving to the other end of the spectrum, classical albums often are very dynamic and typically are (almost) uncompressed. Two exemplary albums reveal DR indices up to 18. The differences between Audirvana and Qobuz are clearly significant.

Silvestrov - To Thee We Sing_ Latvian Radio Choir.jpg

Stravinsky - Histoire Du Soldat_ German Chamber Philharmonic Orchestra.jpg


The chart below summarizes this little analysis. Each dot represents one track. It is visually apparent that the bigger the DR index becomes, the more Qobuz’s loudness calculations deviate from EBU/ITU standard. This correlation is also expressed statistically in terms of R^2, i.e. the correlation coefficient for a linear regression.

replay gain variance & DR index.jpg


(For completeness it should be mentioned that a similar but much weaker correlation exists between replay gain variance and loudness range as defined by EBU R128.)

The big question now is how Qobuz internally calculates replay gain/level matching, anyone with a clue?
 
OP
Prius Factus

Prius Factus

New Member
Joined
Feb 18, 2022
Messages
3
Likes
1
Location
Germany
That's right, EBU R128 is the standard everybody should follow when dealing with loudness. Qobuz claims that they use R128 as well with a reference loudness of -18 LUFS for loudness normalization. However, something apparently goes wrong as their normalization sometimes deviates from -18 LUFS significantly. Happens only with music with less than -18 LUFS program loudness and Dynamic Range of 15 or more.

I emailed them and they wanted to double check.
 

theREALdotnet

Major Contributor
Joined
Mar 11, 2022
Messages
1,181
Likes
2,036
I’d be interested to hear what they come back with. I usually listen to classical and jazz and find that sometimes the volume levelling (based on Qobuz reported gains) is just weird. It annoys me enough to turn off volume levelling altogether now.
 

JanesJr1

Addicted to Fun and Learning
Joined
Jun 9, 2021
Messages
505
Likes
450
Location
MA
This is my first post in this forum, so hello to everybody!

I am using Audirvana Studio for playing my local music and streaming from Qobuz. Audirvana makes it easy to see how much boost/attenuation Qobuz applies to each song to level out inter-track loudness differences (of course only if you choose to use that feature). Strangely, in some cases this replay gain seem to differ quite a lot from loudness leveling derived from industry standards (i.e. measuring the integrated loudness as defined by EBU R128 / ITU-R BS.1770 for each track and taking the difference to some reference loudness as replay gain).

I thought it might be interesting for you technically knowledgeable people to read about that in more detail, so the following is a re-post from what I wrote at the Audirvana forum.

=========================

The latest AS version 1.11 has introduced the option to use Qobuz’s replay gain values to level loudness differences between tracks. However, with some albums/tracks there are significant differences between the gain computed by Audirvana’s built-in replay gain analysis tool and the values provided by Qobuz. Such tracks, downloaded from Qobuz and loudness-analyzed by Audirvana, result in replay gain values that can differ from Qobuz values by up to 10 dB for otherwise seemingly identical streams – that’s huge! The good news is that for music listeners preferring the most popular genres (with limited dynamic range) this is not relevant in practice. So if that is you you won't miss much if you stop reading here.

Replay gain is calculated by Audirvana in compliance with industry standards (EBU R128 / ITU-R BS.1770). Apparently Audirvana uses a reference loudness of -18 LUFS (I cross-checked with other EBU compatible analyzer software and can confirm this). Qobuz’s gain values tend to differ from this standard if music has a high dynamic range, typically with a DR value of 12 or more. Any music with DR < 10 will result in gain values practically identical to standard.

Out of curiosity I checked 10 albums with more than 100 tracks in total. It cannot be excluded with certainty that some of the analyzed local albums have been changed in Qobuz’s repository since I downloaded them, e.g. were replaced by a re-mastered version (or downloaded files could be different from their streamed version). However, I think this is rather unlikely (for reasons I will not elaborate on here).

As examples here are two pop albums, significantly compressed in line with current genre standards and in consequence with small DR indices. Audirvana and Qobuz state the same gains for these tracks.

View attachment 187740View attachment 187741

Moving to the other end of the spectrum, classical albums often are very dynamic and typically are (almost) uncompressed. Two exemplary albums reveal DR indices up to 18. The differences between Audirvana and Qobuz are clearly significant.

View attachment 187742
View attachment 187743

The chart below summarizes this little analysis. Each dot represents one track. It is visually apparent that the bigger the DR index becomes, the more Qobuz’s loudness calculations deviate from EBU/ITU standard. This correlation is also expressed statistically in terms of R^2, i.e. the correlation coefficient for a linear regression.

View attachment 187745

(For completeness it should be mentioned that a similar but much weaker correlation exists between replay gain variance and loudness range as defined by EBU R128.)

The big question now is how Qobuz internally calculates replay gain/level matching, anyone with a clue?
I am thinking of switching to Qobuz. A basic question: am I correct from the above that if the user wants NO loudness normalization, they can choose that option in Qobuz?
 
OP
Prius Factus

Prius Factus

New Member
Joined
Feb 18, 2022
Messages
3
Likes
1
Location
Germany
I’d be interested to hear what they come back with
Me too, will report back.

I usually listen to classical and jazz
Same here, and I am not using loudness levelling either. I am quite keen though on adjusting volume manually to keep loudness consistent when doing critical listening.

am I correct from the above that if the user wants NO loudness normalization, they can choose that option in Qobuz?
I haven't been using Qobuz's native apps for ages since I am using the Audirvana player. In Audirvana the loudness matching is an 'opt-in' feature for all streaming and local sources, including Qobuz. So if you don't actively switch it on you will have no normalization.

As for the Qobuz native apps, I took a look into the app settings (iOS and Windows) and I could not find a similar switch there. However, from what I can tell by ear is that there is no volume normalization going on per default (it's easy to check by playing two songs with a loudness difference of 10 LU or so o_O ).
 
Top Bottom