The Song Fight Bulletin Board

I'm starting this here so that I don't sidetrack either of the other two discussions. To get caught up you should probably read this thread:
https://songfight.net/forums/viewtopic.php?f=11&t=11553 - Loudness management

And this post by furrypedro in the pocket full of bones review thread:
https://songfight.net/forums/viewtopic. ... 40#p212340

So in response to furrypedro:

Now on Song Fight! ... most stuff sounds quiet

Louder things sound better as a rule. In order to appropriately compare two mixes, one would have to normalize the volume so that both tracks are at the same perceived loudness to truly compare.

The question is should I then bring the level of my mix down so that it sounds normal amongst the song fight crowd. Frankly, I'm not gonna do that. Like most people, I won't touch my track again after the fight is finished and I'm more concerned about how it'll sound out in the normal world alongside the music that most people listen to rather than how it'll sound for one week in a song fight

Firstly, if you're "releasing" a track elsewhere, putting it on your soundcloud or youtube or spotify or anything, then you make a very valid point. You should be mixing so it sounds "good" in whatever setting you're gonna ultimately put the music. But most (definitely not all) services have some sort of normalization for play back. If you read that first thread, for instance, you'll see Spotify plays back at -14 LUFS. So if you have a track that is -8 or -6 or whatever, they are gonna simply reduce the volume until it plays back evenly with all the other tracks. Similarly if you are at -16 or -20 they are gonna raise the volume. To me it makes more sense to mix/master/render at that level so that there is less chance of over-compression and you'll have a good reference. I, personally, aim for -14 LUFS and will error on the quiet side.

Secondly, a louder mix generally sounds better than a quiet mix because human ears/brains are weird. So the same mix rendered at different levels (assuming you don't have to compress to get it louder) will sound better/worse depending on the level even if the mix is the same. However, if your song is quieter than the other songs or louder than the other songs, I'm simply gonna use the volume knob to adjust it (or if it's too loud, I'm gonna skip it half way through on repeated listens because I'm susceptible to ear fatigue, especially if I'm listening on the phone where I tend not to change the volume song to song.)

While it's more work, it's not that hard to do renders with varying levels. If it were me, I'd do a songfight render and then another render for whatever application I would release the song on. I would hate for a good song to miss out on a listen or a vote because it was too loud or too quiet compared to most of the other songs in a fight.

furrypedro said:

none of us are using thousands of pounds/dollars worth of mic/valve compressors/anything in the signal chain, and without that ridonkulously expensive gear you just can't achieve pro results sadly.

The guy over at recording revolution makes a good case that you can make very good sounding recordings with a very inexpensive set up. He has a ton of free videos and I especially liked his 5 minutes to a better mix video series (all three of them.) As well as some of his other things. Yeah, he sells his education series etc., but his free videos are worth getting into.. Here's a link to his channel in case you want to check it out:
https://www.youtube.com/channel/UCjRzsi ... HV4-2LKBtg

We didn't really get into the loudness war, but hinted at it. Personally, I'm not convinced that a louder mix has any real world benefit in most listening applications. And in songfight, having an obviously much louder mix than everyone else draws attention away from the song and to the level. That can't be a good thing, can it?

Yeah and this is why I'm kind of leaning towards the idea of having all uploads be automatically ReplayGain-tagged and recommending people to enable ReplayGain/SoundCheck in their player of choice, and at least suggest a target LUFS level in the FAQ. Not sure how to convey this in a way that's easy to follow and see though.

I think that were an upload form ever implemented, giving submissions a LAME normalization pass is a reasonable and feasible level of processing that would help the Song Fight listening experience be a little more consistent in one specific area.

JB