Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Matt]

Last night, I had a strange problem.  I had Replay Gain enabled with 'Automatic based on current playlist' set as the adjustment.

I played from a list that included all the audio in my library.  The volume was really quiet.

It turns out this was because I have a recording of a rain storm that has a +17 dB Replay Gain.

This means the 'Automatic based on current playlist' wants to turn everything down 17 dB (14 dB actually, since the peak level of 70% allows 3 dB of headroom before clipping).

So a Metallica song with a Replay Gain of -20 dB becomes -37 dB when rain happens to be in the playlist.

The volume levels between the rain and Metallica are the same, but the extreme attenuation required to achieve this isn't what I wanted.  I wasn't listening to the rain, it just happened to be in the playlist.

I wonder if it would be better for 'Automatic based on current playlist' to have some cap, and allow really quiet tracks to play quieter than other tracks.  The cap could be an absolute decibel threshold (say +3 dB or some percentile so 5% of tracks would play quieter than others).

Thoughts?

[MrC]

I'm not versed in this well enough to have any recommendations.  Here is a thread I recall commenting on with the same issue.  There may be suggestions there:

http://yabb.jriver.com/interact/index.php?topic=58941.msg398661#msg398661

[gappie]

The volume levels between the rain and Metallica are the same, but the extreme attenuation required to achieve this isn't what I wanted.  I wasn't listening to the rain, it just happened to be in the playlist.

Thoughts?

maybe take the file out of the playlist or set the wizzard of radio so that it leaves files like that out.   

I'm not versed in this well enough to have any recommendations.  Here is a thread I recall commenting on with the same issue.  There may be suggestions there:

http://yabb.jriver.com/interact/index.php?topic=58941.msg398661#msg398661

thanks mrC i searched for the same thread, because that was what popped to my mind to. from the header: http://yabb.jriver.com/interact/index.php?topic=58941.0
 
gab

[rick.ca]

...or some percentile so 5% of tracks would play quieter than others.

It seems to me removing the outliers from such an adjustment is always the right thing to do. I can't think of any situation where one could successfully argue an extreme adjustment required for 5% of a population should be applied to the whole. The correct answer is always going to be to ignore the 5%, or treat them differently.

[cncb]

I never saw this option until now.  Why would you use this instead of "Fixed" (which seems to avoid this problem)?  Just curious.

[Matt]

I never saw this option until now.  Why would you use this instead of "Fixed" (which seems to avoid this problem)?  Just curious.

Automatic uses the highest gain possible that won't lead to clipping and that keeps all the songs the same volume.

If you play a pile of Metallica songs that range from -15dB to -20dB, there's no good reason to turn them all down a lot.  You could just turn the -20dB down -5dB and leave the -15dB tracks at 0dB.

Depending on lots of things, turning down the volume can hurt the SNR of the final output which is why this is relevant.

[tunetyme]

Matt:

I certainly agree.  I found this to be a problem as well.  I just turned the feature off.  I would think you need to set the adjustment in a variable fashion.  If the range is -20db to +20db and 0db being the mid point.  Without studying the matter in greater depth, I would guess that the adjustments should bring it down to a range of -5db and +5db.  This will allow for songs that are loud to play a little softer and the ones that are soft will still be a little louder.  Of course you still have to be concerned about clipping.  I realize this would create a programming challenge especially with some classical music like the 1812 Overture or Beetoven's 5th where you have a wide range of volume within the song itself.  Some Led Zeppelin songs come to mind as well that start off softly and end intensely.

For myself, I enjoy some variation in the amplitude of a playlist.  I strive to move the playlist from soft and a slow beat to a louder and faster beat and back down again.  I find this helps keep my interest and I find it to be more enjoyable than an hour of intense 140 BPM -20db songs rocking the foundation of the house or an hour of very soft music that will surely put me to sleep.

Another thought is to do leveling based on the 2 or 3 previous song and the following 2 or 3 songs which will provide a more dynamic leveling.

I have a lot of interest in this topic.  I would like to stay in the loop on this.  I have a very wide range of music that I would be happy to run tests on.

Tunetyme

[pcstockton]

Dont use replay gain!!!!  That is what "Volume" is for.