Dynamic Range Metric - Track and Title - Feature Request

Please add the ability to calculate track and album DR and store in metadata with ability to filter / sort against this metric using “Focus”. Thanks!




That’s a really great idea!

+1, yes indeed great idea.

+1, love this idea to see the DR

I suggested this over a year ago when testing very early alpha builds. At the time, I didn’t realize how potentially flawed the DR measurement is: I was at a Bob Stuart (of Meridian) demo where he played a track which had terrible measured DR; but the track had been carefully built up with multiple layers of sound, all of which had extremely good dynamic range (there was no dynamic range compression involved anywhere).

An extreme example perhaps, but food for thought. There are “standard” DR measurement methods which could be implemented nonetheless.

I agree that a lower DR does not always mean an inferior recording/mastering job.

But it is a useful measurement for referencing, but some have taken the DR as being a be all end all of a good sounding album.

The algo as implemented in TT DR Offline Meter is terribly flawed IMO, and I previously posted about this quite a while ago on Computer Audiophile. I was going to link that thread but it seems that it was deleted?! You can see how they implement it from the in-app help splash screen. I don’t know if this is some sort of industry standard or not, but I guess it would come down to compatibility with that database vs. following some sort of better implemented standard.

To say that DR doesn’t directly correlate to quality is a bit of a “red herring”. We can filter sort against sampling rate even though there are some 44.1KHz recordings that sound better than some 192KHz. This seems to come up every time… I’ll give a few cases where it is great to have this info: 1) When one is comparing different masterings of the same recording and 2) When you are trying to figure out which mastering you have vs. the DB. Also, there may be times you don’t want to be shouted at and want something with a bit less “wall of sound”. I’d even love to have this info written back into the metadata comments section, but that would open up a whole new can of worms (modifying metadata). But I guess it would be possible with the Roon managed data directory. Just thinking out loud here…

I’ve had some tough times extracting 3-4 versions of a recording from my library, putting them in their own folders, running the DR analyzer on them, reading the log output files, then cross-referencing against the DB. Totally painful. And in many cases I have identical duplicate masterings, or I give them a listen and typically tend to prefer the original higher DR recording, deleting the supposed remaster (Demastered).

Knowledge is power, and this is a nice piece to have to analyze ones own library. If anyone can chime in on how TT DR does it vs. any industry standards that would be very interesting. Or should a better algo just be used from the get-go? Also, I like knowing if the music clips (“over”), and what is the RMS loudness but again just throwing it out there for discussion.



1 Like

Is there a well accepted measurement algorithm in this space?

I’d check this out Danny - (EBU R128):


1 Like

noted and investigating. it seems that r128 might also be better than replaygain at sparse and bass heavy tracks.


Danny, I just started using Roon (loving the experience so far!) but have been using JRiver for about two years. JRiver uses R128 for its audio loudness analysis and I can say it is a huge step up from ReplayGain. I also often inspect the dbTP (decibel True Peak) derived from R128 analysis along with dynamic range. I am also an audio engineer so I am a nerd about these things but that is certainly something that makes me think I’ll tab over into JRiver for that in the meantime. I love using Roon with JRiver via the WDM driver for DSP (“headphones” setting/crossfeed in my case) and their extremely transparent volume control. I don’t even really mind using two apps for the time being. But it can only process what it is being fed, and R128 would be the best way for consistency. My library is still being analyzed in Roon so I’ll see how well ReplayGain works soon.

Many thanks.


1 Like

When we make this change, it’ll just reanalyze in the background and start using new values


This would ba a fantastic addition, and is currently the only reason I still dip into JRiver occasionally.

+1 On the DR Meter.

1 Like

The ability to compute and filter DR range would be a great addition.

Maybe Roon should try to incorporate sth like MusicScope’s approach (or join forces with them) :

I really like this tool and use it quite often to evaluate digital copies and hi-res files.


I agree this would be a great addition.

+1 On the DR Meter.

@danny @brian , here’s an implementation of the Dynamic Range calculation: https://github.com/simon-r/dr14_t.meter that demonstrates the algorithm applied.



Seems this DR rating is quite widely accepted by Germany