One Synth Challenge #159: Surge XT (j5v Wins!)

[empphryio]

I guess Reaper will just tell you the LUFS in the newer versions when you render so you don't even need Youlean.

[Abstract Egg]

Congrats to the top scorers, well deserved all around.

Concerning my own entry (this month and last month) I've got to finally admit I'm stumped about the mechanics of loudness.

Why does my stuff sound so much quieter than everyone else's?

In both this month's and last month's OSCs I was constantly struggling with the loudness _showing up_ as way too loud in Reaper (i.e. in the visual displays of loudness), but not actually _sounding_ that loud. What common beginner's error am I probably making and need to figure out how to fix?

I of course know that compressors and limiters exist heh. Particularly with limiters, part of the problem I experience is that I end up just distorting everything if I use them, but if I don't use them everything's even quieter.

Have you tried high passing everything but kick and bass instruments? Depending on the instrument and context, you can highpass from 80-100hZ and upwards, for hihats you can sometimes go up to 250Hz. Kick and bass I would start high passing from 20 to 40Hz, again, depending on the context of the track.
Not only will this remove rumble, but it will also make your compressors and limiters work less hard. Low frequencies carry the most energy, so a compressor will slam harder if that energy is present, even if it's redundant, ruining your carefully crafted mix balance in the process.

The thing with limiters is, they're basically compressors with a short attack and infinite ratio. Whatever you feed the limiter, it will clip everything above a set threshold. If that audio has a lot of
hard transients and low frequency information, your limiter gets triggered more and will try to balance the overall level by clipping more than it should, which causes distortion. Well, clipping is actually a type of distortion and hence distortion is a form of compression (since distortion diminishes the dynamic range, which is the definition of compressing).

Bare in mind there's both upward and downward compression, limiting and expansion.
As a general rule of thumb, I will start a mix just with setting levels and pans and applying high pass filtering on everything but kick and bass (be it a Fender bass, a synth or a tuba). Then I'll apply EQ and compression judiciously, and I'll sum most of my instruments to busses (group tracks), which will usually receive some processing too. This helps glueing the mix together, alongside using reverbs and delays for the sound placement (distant vs. close) This way I will try to aim my master level as much in the ballpark as possible. Lastly I will apply a master buss compressor and limiter, making sure these barely get tickled by the master track's audio. Also, I NEVER mix with a master buss compressor and limiter engaged, those come in to play as the final touch.

There's much more I could add, such as the difference between absolute and perceived loudness, but I am not a mixing engineer by any means, and there's tonnes of information to find on youtube about each and every one of these topics. Hope my explanation helps, I tried to convey what I know to the best of my abilities.

One last but crucial tip: during this whole process I carefully manage my gain staging, meaning every instance of every plugin or effect has to more or less preserve the level of volume. You can add a little (or a lot of) volume along the way, but I make sure that no single plugin makes the audio clip in said chain, because that's a recipe for a distorted master. Google gain staging and you will find a slew of tutorials.

Warren Huart has a great tutorial on how to get louder mixes, on his fantastic channel "Produce Like A Pro":

You'll get there in the end, I struggled with the same issues in the past and even to this day sometimes. All the best

[Abstract Egg]

Congrats to the top scorers, well deserved all around.

Concerning my own entry (this month and last month) I've got to finally admit I'm stumped about the mechanics of loudness.

Why does my stuff sound so much quieter than everyone else's?

In both this month's and last month's OSCs I was constantly struggling with the loudness _showing up_ as way too loud in Reaper (i.e. in the visual displays of loudness), but not actually _sounding_ that loud. What common beginner's error am I probably making and need to figure out how to fix?

I of course know that compressors and limiters exist heh. Particularly with limiters, part of the problem I experience is that I end up just distorting everything if I use them, but if I don't use them everything's even quieter.

Are you using youlean meter? It tells you your LUFS. Most aim for about -14. TDR Kotelnikov or TDR Nova on the master or just Loudmax. Or limiter No6. Loudmax is the quickest/easiest. Also just normalize the master of course.

No offense, but I would NOT normalize the master. It will ruin your carefully crafted mix balance and raise the noise floor, which is usually something you would want to avoid. Then again, there's no rules, only things that work better than others in a given context. But normalizing a master is something I will never do.

[Abstract Egg]

[The Sound Of Merlin]

Congratulations to the 5 winners of this wonderful Surge XT challenge. Well deserved! Absolutely great stuff!

[doctorbob]

Congrats Top 5, most of my favs ...

Well done @z.prime for the single instance in the top 5!

dB

[empphryio]

No offense, but I would NOT normalize the master. It will ruin your carefully crafted mix balance and raise the noise floor, which is usually something you would want to avoid. Then again, there's no rules, only things that work better than others in a given context. But normalizing a master is something I will never do.

Works fine for me thanks.
...I suppose if it was a live recording maybe but I'm perfectly happy with normalizing vsti's.

[z.prime]

I'm seeing some videos posted about mixing and mastering advice and I'm not sure I would recommend following their advice exactly. They're omitting the most important thing: proper gain staging. My track, for example, was reasonably loud and not distorted and only used Surge with no compressor or limiter (I did use a surge clipper at the very end, as you can see in the patch).

There's a very long thread about gain staging that has some absolutely gold information. I only made my way through the first 10 pages or so, there's probably some other great stuff in there. Even though it's a dubstep forum, the idea is effectively the same for all music genres (except maybe classical or somesuch). If you want, check it out here and pay particular attention to anything that macc says.

The idea is this: find good levels for your instruments, and just set them. This ends up giving you consistent levels every time. For example: I usually look for a kick to hit around -6db, snare a bit quieter, sub -8db or so... pads at like -20db, etc.. What you end up with, if you have properly gain staged, is something that almost mixes itself, you won't be fighting adding volume automation all over to get parts to fit the right way because something weird happens in a limiter or in a mix bus. Make sure to HPF and EQ things as appropriate before considering the levels - e.g. you might have a patch that has a lot of bass you don't really even want to mix in because it muddies up the mix.

The next thing is sorta what one of the guys was talking about: processing individual tracks as needed and then processing in groups. For example: compressing a vocal track to even out the levels is a fine use of a compressor, then summing all vocal tracks together, maybe into a vocal bus and processing this further is definitely a good structure for a project. All percussion except for kick and snare being sent to a percussion bus, all basses to a bass bus, etc..

You get loudness in the mix, not the master.

Now, once you've already gain staged things properly, and set up good instrument groups (busses), the next thing I would suggest to do is: clip. Huh, you say? Skrillex, whether you like his music or not should agree his mixes are pretty damn good. He had a quick walkthrough video of one of his tracks where he basically showed using the "clip to zero" method. Baphometrix has a series of really long videos with a lot of good information about this. If you want loud mixes, you should definitely be doing this. The gist is: on your busses (maybe individual tracks if needed), you hard clip high peaks. This makes the overall mix work a ton better because you don't have random peaks triggering limiters or compressors and preventing pushing the levels up higher. (This is why I started working on ZEEK CLIP, to make it super easy to see what peaks you're affecting.)

At the end of something properly gain staged with proper peak clipping, you now can easily have a master chain with a limiter that you just boost up 6+db to get the levels you're looking for. Always compare with reference tracks, especially paying attention to the LUFS of each to make sure you're in the ballpark for something you're looking to match.

P.S. Sound selection can also have a huge impact on this -- sometimes sounds just do not mix well together and you have too much mud or overlapping frequencies and no amount of boosting/eq/compression/etc. is going to fix it.

[Rahodees]

What does "normalize the master" mean?

[Boy Wonder]

About a year and a half ago, someone broke into my bank account and transferred $5000 out of it and left me with $80. It took me two months of back and forth with my bank to get my money back, but the ensuing nightmare left me so paranoid that now I go out of my way to have various anonymous logins and identities to protect myself.

When I went to vote in OSC #159, I didn't know I had to use my KVR login, so I created a new one and voted. Later, when I learned my vote may not count because I didn't think KVR would recognize me as the creator of "Daddy's Little Android", I logged back into OSC #159 but used my KVR login instead. Still, to be on the safe side, I didn't vote for the same tracks again. I believe my track was disqualified because I'd voted twice. Hard lesson learned.

[empphryio]

What does "normalize the master" mean?

It means turn the volume up just short of 0. In Reaper you can double left click and a window pops up where you can do it automatically. I have no idea how turning the volume up a few dbs is supposedly going to ruin the mix. You're free to turn the volume up and down on your master and decide for yourself if it's actually ruining anything. If it was a live recording you could use reafir to cancel out background noise.

[nenneb]

Congratulation to all of you for great work, esp. to the winners of course!
Personally I'm a bit disappointed, but thank you very much for 3x 5points and 21x 4points!!!! On the other hand I don't understand a one-point-rating, but it's only one ... maybe one point for one instance?
All the best to all of you!

[ElVincente]

Congrats to the winners and to everyone, what a nice round !
@boy wonder, too bad I really liked your dubby baby android! Hope it won't stop you from getting aboard in the future.

[BrainBeatMusic]

Congratulations to the winner. I personally am feeling a bit surprised at how low my ranking and score were (= 40th with 194 points) especially compared to my last entry (I came 8th with 211 points) and how many nice comments I got this time. I will try again to get back in the top 10 (ideally top 5) but not sure if it will happen with the current competition as the synths are not great. To me the competition was for sure a lot closer and good in general but still surprised by the points gaps.

Edit: I didn't realise the scoring sheets were made available now looking at it I am shocked that nearly everyone in the last 2 competitions has gotten at least 1 score of 1 and 2 for their track. I assume 1 of the ones is your own vote but the others surprise me even for the winners seems like there are some voters who are either hard markers, don't like/appreciate much, or are strategic with their voting.

There would seem to have been 2 more people who voted this time than last time (57 to 59) even though there were 70 entries last time and only 55 this time.

[Boy Wonder]

Congrats to all the winners!