PG Music Home Forums General Discussion Recording, Mixing, Performance and Production Normalizing a waveform based on loudness

This might be a long-shot question.

I have a ride cymbal recording generated by me live playing a drum pad. Because of the sensitivity of the pad and my limited drumming skills, it's difficult to produce a raw stem with consistent loudness for the full duration of the recording. With any live (or BiaB generated) recording, my first step is to "normalize" the waveform. But even after normalizing, the loudness is still noticably variant.

I have Studio One Pro and understand that it's normalizing algorithm is based on "peak" (presumably to prevent clipping), rather than "RMS" which would address loudness.

Does Studio One contain a loudness/RMS normalization command?
If not, how else might I process this waveform to achieve a relatively consistent loudness?
I'd prefer to stay within Studio One but if that's not possible I'll use another tool.

Maybe a compressor could be the answer?

not sure how the loudness/normalise would work with a compressor in a single pass but if you export the orignal wav with a heavy compression to smooth it out then normalise the compressed wav it may solve your problem.

you might need to experiment with the compression so you don't lose dynamicsand attack

Perhaps Gregor Beyerle from Presonus can explain this for you.
Normalize Audio - Yay or Nay?

Hmmm, a compressor.

Compression reduces the overall dynamic range of a piece of audio by detecting when it exceeds a specified level, and then attenuating it by a specified amount. Put simply, it narrows the difference between the loudest and softest parts of a track so that it’s more consistent in level.
https://splice.com/blog/what-is-compression/

Even though I'm trying to even out the loudness of the cymbal (not necessarily narrow the difference between the loudest and the softest), this still might work and is worth a try. I'm sure lots experimenting is needed, but I'm game

Thanks Bob.

Thanks Brian, I'm having issues connecting to that video but will try again later.

I've watched many a Captain Gregor video, in his "starship studio", he knows his stuff.

Steve, can the drum pad audio receive and send MIDI?
If so record your playing as MIDI. There are a number of ways to even out velocity in Studio One Pro. Then send the modified MIDI back into your drum pad and record the audio from it.

Another option would work if your quietest cymbal volume is loud enough. That would be to us a limiter on it. That would lower the loudest only sounds.

I use a plugin called Pure Limit. It has AI and makes it extremely easy to get a good sound quickly. It should do what you need if you can put it on just the drum kit part.

I would do the normalize and see if that works for you. You can always UNDO it as long as you don't close the program. If the normalize doesn't work.... it will set the highest peak in the song as it's shall not exceed point and essentially amplify everything else by the amount needed for that one peak to be at the set point...... you can then use a compressor to even out the rest of the lower hits and bring them in line with the rest of the louder hits.

If I was doing this I would likely record it as midi so I could swap the cymbal samples to get the best overall sounding cymbal. Then you convert that to audio after you have set the velocity and quantized it to your satisfaction.

Another vote for compressor here ..
to understand the concept, I once again defer to Mac ..
http://audiominds.com/compression.html

/it's an ancient site but has lots of nuggets worth harvesting
// it may be one of the last caches of Mac knowledge now
///I still miss Mac, learned so much from him

Status Update:
Thanks to all who provided ideas. I was wrong. I initially thought this would be a long-shot question because I've seen no others here talk about recording percussion; be it acoustic or electronic. But I now see that audio is audio and sound is sound with lots of processing overlap/commonality across different instruments.

Brian, I watched the video and it confirms that S1 is using Peak Normilization. The Gain Staging he talks about at the end I don't think will address loudness variation within a stem but addresses variations across stems.

Mario, Yes, I have a boatload of outputs on the drum pad, MIDI is included. So far I haven't needed to deal with MIDI yet. Partly because of the satisfaction I get from capturing what I play on a physical instrument in my final recording, even if my non-MIDI recordings are far from perfect. Yes, I do apply effects and volume automation so it's a matter of personal philosphy; at least at this point in time. A year from now things might be different. As for the Limiter, I think I'll explore that as-needed. It's looking like the Compressor (although not perfect) is giving me an improvement.

Matt, I looked at Pure Limit. It looks like a solid plug-in. I'm hoping that the next major rev of S1 will contain that kind of AI capability.

GuitarHacker, Peak normalization in S1 doesn't solve this problem but the compressing seems to improve it somewhat.

rharv, thanks for the AudioMinds webpage. I bookmarked it. There's a lot of good stuff packed into that article.

Here is what compression can do to this recording. The first half is with no compression, just peak normilization. The second half is with compression per the parameters in the attached screen shot. I also included an 86 BPM click as a consant volume reference. One other detail: The act of compressing seems to have reduced the overall volume of the recording. So in order to compare apples-to-apples, I decreased the non-compressed waveform by 3dB so that the variations within it wouldn't be percieved as exagerated.
Compression Comparision

My ears are telling me that compressing does in fact reduce the loudness variation. As I have applied it, does it eliminate variation? No. But someone coined "Never let perfection get in the way of good enough". Besides, some may say that a little "humman errorr" is not necessarily a bad thing.

One this is for sure, you could spend a lifetime studying mixing and audio processing, it's a huge subject. And I could spend days with just fiddling with the Compressor parameters. For me, I just want to know the bare minimum to construct pleasing songs; playing music is a higher priority than processing sound. That said, I'm grateful for all that responded, the learning goes on and the compressor now has a place in my toolbox . . .

Steve, in the gain window of the compressor is a makeup knob. Increasing that will increase the overall gain of the compressed track. If you already know this then sorry for the redundancy.

Compression and limiter (I also like the Pure plug-ins, EQ, limit, compression, and reverb - thanx Matt!) are almost always in my workflow.

My first thought upon reading your description was, compressor. Compressing, normalizing, and limiting can all overlap.

AI simply makes it easier for those of us who are not doing audio engineering daily. You are right; I wouldn’t be surprised if the various DAWs adopt these functions.

By design compressors decrease loud parts and increase quiet parts while a limiter simply reduces peaks to stop clipping. I would think light compression (perhaps ideally with a multiband compressor) would work. FWIW.

Bud

Hey Mario, yeah I know that. I tweeked each knob from zero to full scale just to get a feel for the effect and what I posted was the best I could produce inside a few minutes.

But there’s always more to the story that can be revealed from a little more research.

It turns out that Audacity actually has 2 normalizations available “Perceived Normalization” and “RMS Normalization”. This is from its manual.

Normalize
There are two available options "perceived loudness" (default) and RMS:
 perceived loudness: the default -23 LUFS (the EBU standard) will produce audio that is approximately 25% of full scale.
 RMS: This will change the amplitude such that the result has the desired RMS level The default setting is -20dB which will also produce low level audio.
Both LUFS and RMS normalization ensures that different audio projects come out at a relatively uniform volume.

I won’t know how well Studio One’s Compressor solution will work until it’s embedded in the mix with everything else. But if it doesn’t work that well it’s nice to know I’ve got Plan Bs.

But what is baffling is why Presonus didn’t bother to include RMS Normalization in S1. It’s easy-peasy for their programmers to do. On a scale of 1 to 10 this complaint registers only about a 3 for me but is probably the first weakness I’ve come across in Studio One. I wonder if anyone has requested this to Presonus.

PS To the Audacity programmers: Keep pushing your envelope; competition in the marketplace of products is a good thing!

[quote=Bass Thumper............................
I won’t know how well Studio One’s Compressor solution will work until it’s embedded in the mix with everything else. But if it doesn’t work that well it’s nice to know I’ve got Plan Bs.

........................ [/quote]

That is true for any effect and/or a sounds tonally qualities, i.e. what sounds great alone may suck in a mix. But I'll bet you already know this also.