APU Software - new approach to Compression/Expansion?

[mibby]

This looks really interesting approach to dynamics control and very reasonable at $19!

https://apu.software/compressor/

And the Meter plugin is FREE!

https://apu.software/meter/

Manuals here:

https://docs.apu.software/

 

Edited May 6, 2023 by mibby

[Niky Serrano]

 

[mibby]

Doh!  I guess I needed to look further back...  🤔

[TheSteven]

Did you try it?

[Bruno de Souza Lino]

There's not much info about, videos or otherwise. By videos, I mean stuff done by people which aren't the person who created the plugin. I was also under the impression that the only people who know how LUFS work are the EBU and the people that develop meters that display it.

Edited May 7, 2023 by Bruno de Souza Lino

[Yan Filiatrault]

I thought it was written BAPU software… Would have been a new approach to expansion, expanding a chain of thousands of plugins. 

[Aaron Robinson]

👀

LUFS is an industry standard, and fortunately there is an open source and commercially friendly implementation (libebur128). The main motivation behind the compressor was to make use of this in a compressor/expander since RMS and Peak are completely ubiquitous in the industry for those tools yet they don’t take human loudness perception into account.

I’ve thought that was a glaring blind spot for the past few years and that ended up being enough motivation to start doing plug-in dev after getting laid off about a year ago at the same time as having some health issues.

To be honest, I’m pretty new to audio engineering in general. But I worked on video codecs for a decade and this path felt right.  Hoping to improve over time. It’s super exciting to see people start to use the tool and give feedback. Love the criticism as much as the compliments.

 

- Aaron (apu software solo founder)

Edited May 7, 2023 by Aaron Robinson

[User 905133]

I don't need the compressor, but I was mesmerized by the visualization in the video enough that I accepted the generous offer of the freebie loudness meter.  I dared to turn the input on just so I could watch the visualization while I listened to music on youtube.  It was great.  At one point I laughed and realized my usb mic was on and realized the loudness meter's feedback was adding a nice subtle echo delay to everything.  (I listen through speakers.)

That's probably not how the meter is intended to be used, but I just had to share my initial experience.

Thanks, and best wishes to you! 

[Bruno de Souza Lino]

I mean, maybe one could get someone like Dan Worrall to make a video about it or something?

[mibby]

53 minutes ago, Aaron Robinson said:

👀

LUFS is an industry standard, and fortunately there is an open source and commercially friendly implementation (libebur128). The main motivation behind the compressor was to make use of this in a compressor/expander since RMS and Peak are completely ubiquitous in the industry for those tools yet they don’t take human loudness perception into account.

I’ve thought that was a glaring blind spot for the past few years and that ended up being enough motivation to start doing plug-in dev after getting laid off about a year ago at the same time as having some health issues.

To be honest, I’m pretty new to audio engineering in general. But I worked on video codecs for a decade and this path felt right.  Hoping to improve over time. It’s super exciting to see people start to use the tool and give feedback. Love the criticism as much as the compliments.

 

- Aaron (apu software solo founder)

Thanks for popping in Aaron! I emailed you today after buying your plugin.  This forum is a fantastic way to get in front of some of your customers and answer some questions.  You might also like to sign on over on the KVR forum if you haven't already. Some Devs have utilized that site as a support forum for their products.

 

[kevin H]

Just got this and loving it so far.  Can’t wait for your synth

[Niky Serrano]

3 hours ago, Aaron Robinson said:

👀

LUFS is an industry standard, and fortunately there is an open source and commercially friendly implementation (libebur128)...

Add here the GitHub repos of these implementations if helps someone

https://github.com/jiixyj/libebur128

https://github.com/jiixyj/loudness-scanner

[bitflipper]

I normally greet all dynamics processors with a big yawn and treat claims that any of them does anything magical with great skepticism. I long ago settled on a couple compressors for everything and haven't ever wished there was something better.

But I gotta admit this one's piqued my curiosity.

Seems this would be best suited for the master bus, since it's only in the context of full-range material that LUFS makes sense as a leveling reference. I can't imagine, for example, that it would be particularly useful on a vocal or bass track. Only one way to find out: I've downloaded the demo. Kudos to Aaron for offering a 30-day uncrippled demo.

I have a question for Aaron - what's the music used in the demo video? I like it.

[Aaron Robinson]

1 hour ago, bitflipper said:

I normally greet all dynamics processors with a big yawn and treat claims that any of them does anything magical with great skepticism. I long ago settled on a couple compressors for everything and haven't ever wished there was something better.

But I gotta admit this one's piqued my curiosity.

Seems this would be best suited for the master bus, since it's only in the context of full-range material that LUFS makes sense as a leveling reference. I can't imagine, for example, that it would be particularly useful on a vocal or bass track. Only one way to find out: I've downloaded the demo. Kudos to Aaron for offering a 30-day uncrippled demo.

I have a question for Aaron - what's the music used in the demo video? I like it.

True, although you can also set the loudness type to RMS or Peak (or True Peak) instead of LUFS. I’ve been thinking about adding some kind of hybrid mode to use both simultaneously. But that’s just an idea so far, haven’t started digging into it. 

The meter and compressor have different demo songs used.  Both are licensed through Audiio. 

compressor: Philip Dansor - Bloom (Instrumental)

meter: Annasara - A Lifetime Rolls By

[Niky Serrano]

I have a question here a little bit out of topic, but about LUFS in general

I try to publish my songs on Spotify about 14 LUFS always... but I have seen that Spotify has songs of famous artists about 8 LUFS and similar...

Are the LUFS an estimation only, or is a strict rule that you have to follow for each platform individually?

I now that the question is general, but after publish some song successfully in the different platforms, is a concurrent question without a clear answer not only for me but also for a lot of people

To avoid problems, I continue publishing at 14 LUFS, but I think that break the rule is valid too

[mibby]

1 hour ago, Niky Serrano said:

I try to publish my songs on Spotify about 14 LUFS always... but I have seen that Spotify has songs of famous artists about 8 LUFS and similar...

I think when a mix is too loud, it'll just get turned down to whatever the normalized level is.  That 8 LUFS mix is giving up some dynamic range for loudness.  If you can make a great sounding mix that comes in at 8 LUFS (like the pro mix you mentioned ) then you're ahead of the game.  It'll just get turned down, but will still sound good - AND louder overall than a more dynamic mix.  (At least this is my current understanding anyway.)

[antler]

2 hours ago, Niky Serrano said:

I have a question here a little bit out of topic, but about LUFS in general

I try to publish my songs on Spotify about 14 LUFS always... but I have seen that Spotify has songs of famous artists about 8 LUFS and similar...

Are the LUFS an estimation only, or is a strict rule that you have to follow for each platform individually?

I now that the question is general, but after publish some song successfully in the different platforms, is a concurrent question without a clear answer not only for me but also for a lot of people

To avoid problems, I continue publishing at 14 LUFS, but I think that break the rule is valid too

Plugins like Mastering the Mix Expose and ADPTR Streamliner will help you with target loudness levels for the various platforms. From the Expose release notes, it looks like Spotify is -14 LUFS ±2dB.

My advice (for what it's worth) would be to aim for -14 LUFS. That way you'll have a good amount of headroom to preserve dynamics. Dynamics are important because they are what keeps your song sounding punchy where it needs to be, and also preventing ear fatigue. I have a few CDs from circa 2000 during the loudness wars. They sound great when you start playing them because of a psychoacoustic phenomenon that makes louder sound 'better'. Problem is, the constant loudness starts to grate after a while and it's not that pleasant to listen to the whole album even though the music's good.

If it's too quiet, it'll get turned up. If it's too loud, it'll get turned down. But nothing can restore the dynamic range. If it's mastered too loudly, you'll end up with a squashed and quiet song that will sound weaker than one that fully utilises the available dynamic range.

[antler]

1 hour ago, mibby said:

I think when a mix is too loud, it'll just get turned down to whatever the normalized level is.  That 8 LUFS mix is giving up some dynamic range for loudness.  If you can make a great sounding mix that comes in at 8 LUFS (like the pro mix you mentioned ) then you're ahead of the game.  It'll just get turned down, but will still sound good - AND louder overall than a more dynamic mix.  (At least this is my current understanding anyway.)

I don't think that's the way it works unfortunately. Once a mix has given up it's dynamic range, it's gone. Consider two mixes:

• Mix A targets -8 LUFS. The kick drum peaks with the rest of the (squashed) mix.
• Mix B targets -14 LUFS. The kick drum peaks higher than the rest of the mix.

When played back at the same volume, the kicks in mix B will be louder than the rest of the mix and will really let you feel the oomph. You wouldn't be able to do that in Mix A because the mix is already squashed and there's no further headroom. You might be able to do some ducking on the kicks or something else that's clever; it might work, but it's likely to sound less natural than simply having more headroom to emphasise the peaks.

[mibby]

2 hours ago, antler said:

I don't think that's the way it works unfortunately. Once a mix has given up it's dynamic range, it's gone. Consider two mixes:

• Mix A targets -8 LUFS. The kick drum peaks with the rest of the (squashed) mix.
• Mix B targets -14 LUFS. The kick drum peaks higher than the rest of the mix.

When played back at the same volume, the kicks in mix B will be louder than the rest of the mix and will really let you feel the oomph. You wouldn't be able to do that in Mix A because the mix is already squashed and there's no further headroom. You might be able to do some ducking on the kicks or something else that's clever; it might work, but it's likely to sound less natural than simply having more headroom to emphasise the peaks.

I'm not trying to be contentious, but I think perhaps you misunderstood my point.  And let me preface my reply with - I had no idea what the LUFS was going to be on this track. I use it frequently as a reference.

I used a YouTube converter site to download a .WAV file, then ran it through the YouLean online analyzer.

It's a "modern" mix by Serban, one of THE most in demand mix engineers today ( https://serbanghenea.com/).  Even turned down on Spotify, you cannot objectively say this mix lacks "oomph!" In fact,  I think you'd be hard-pressed to find a better mix in spite of the lack of dynamic range in this one. (material dependent of course!)

Here are the results:

My real point is, that, in spite of the lack of dynamic range in this particular mix, it is a phenomenal mix and not lacking much of anything. You can still get a superb mix with a LUFS of -8.37 if you know what you're doing as Serban does. 

 

 

[antler]

44 minutes ago, mibby said:

I used a YouTube converter site to download a .WAV file, then ran it through the YouLean online analyzer.

It's a "modern" mix by Serban, one of THE most in demand mix engineers today ( https://serbanghenea.com/).  Even turned down on Spotify, you cannot objectively say this mix lacks "oomph!" In fact,  I think you'd be hard-pressed to find a better mix in spite of the lack of dynamic range in this one. (material dependent of course!)

Be careful with that - YouTube and Spotify (probably) do different things with the audio. According to this, YouTube could be applying compression to the audio which could affect the results of analysing audio from YouTube.

In either case, I agree it's a great sounding mix - if that's the sound you're after, it's definitely worth using that as a reference track.