dyross
Senior Member
Sounds great! What was the LUFs range before/after the compressor? I don't hear any pumping, so if this is indeed set-and-forget, pretty handy.
Can you screenshot your settings?
Share your thoughts on this new compressor for orchestral
- Thread starter Buzz
- Start date
Can you screenshot your settings?
mbishop
Senior Member
Sorry, I wasn’t very clear! That’s the raw music I’m using this compressor with. I posted it so you could try it with the APU compressor and tweak the settings hopefully as a test bed so you could see the effect and why I might have chosen the longer length. I can post the settings that worked for me though…
dyross
Senior Member
Too transparent
mbishop
Senior Member
Sorry it took so long for a response.
I've come around to the idea that this thing is really good for mastering classical pieces (or music with a large dynamic range). I might be saying things that are true about other compressors because I haven't used a wide variety of them, but here's what I'm finding very useful about this.
Let's say you have a piece of classical music that has both extended quiet periods and periods at the top of their volume. And let's say you are trying to hit an average LUFS target for Apple or Spotify (different standards). Well, first, you can have this compressor learn the LUFS range of your piece by playing the whole thing through it in "learn" mode. Then copy the learned range to the target range so they are the same.
Then, you can raise the top of the target range and play back the peak of your piece. Adjust the top of the target range so you aren't going above 0 db. And have a limiter after it to help with errant peaks that you might miss. But get it as close as you can.
Now, set the bottom of the target range to about 1/3 of the way to the top of the target range. Set the lookahead, attack, and release to some longish value. You want it long enough that you can't really detect that it's happening. You want to preserve the local dynamic differences as much as you can. Mine is set to 500ms and I don't really notice it. The narrower your target range, the more time you are going to need for the gain-riding to be less noticeable since there is more ground to cover as the compressor adjusts.
Now, play your piece all the way through, using some tool that will give you the average LUFS (I'm using Sonible true:level). If you need a lower average LUFS, drop the bottom of your target range which will bring down the overall average LUFS. Do the opposite to increase it.
Here's an excerpt of my mockup using three different target ranges.
Here's the widest target range:
View attachment SA Wide.mp3
Here's the narrowest target range:
View attachment SA Narrow.mp3
And here's the middle ground (where I was able to get my target LUFS)
View attachment SA Mid.mp3
I probably could have made the attack/release/lookahead times even longer for the narrow version because I can still perceive that something is "raising the volume knob".
The reason I like this better than entering in the fader automation is because changing the bottom of the target range is so tweakable to hit that LUFS target. Then all the gain riding just readjusts automatically.
At the end of this process, you will have the highest dynamic range (in db) for the exact target LUFS.
I've come around to the idea that this thing is really good for mastering classical pieces (or music with a large dynamic range). I might be saying things that are true about other compressors because I haven't used a wide variety of them, but here's what I'm finding very useful about this.
Let's say you have a piece of classical music that has both extended quiet periods and periods at the top of their volume. And let's say you are trying to hit an average LUFS target for Apple or Spotify (different standards). Well, first, you can have this compressor learn the LUFS range of your piece by playing the whole thing through it in "learn" mode. Then copy the learned range to the target range so they are the same.
Then, you can raise the top of the target range and play back the peak of your piece. Adjust the top of the target range so you aren't going above 0 db. And have a limiter after it to help with errant peaks that you might miss. But get it as close as you can.
Now, set the bottom of the target range to about 1/3 of the way to the top of the target range. Set the lookahead, attack, and release to some longish value. You want it long enough that you can't really detect that it's happening. You want to preserve the local dynamic differences as much as you can. Mine is set to 500ms and I don't really notice it. The narrower your target range, the more time you are going to need for the gain-riding to be less noticeable since there is more ground to cover as the compressor adjusts.
Now, play your piece all the way through, using some tool that will give you the average LUFS (I'm using Sonible true:level). If you need a lower average LUFS, drop the bottom of your target range which will bring down the overall average LUFS. Do the opposite to increase it.
Here's an excerpt of my mockup using three different target ranges.
Here's the widest target range:
View attachment SA Wide.mp3
Here's the narrowest target range:
View attachment SA Narrow.mp3
And here's the middle ground (where I was able to get my target LUFS)
View attachment SA Mid.mp3
I probably could have made the attack/release/lookahead times even longer for the narrow version because I can still perceive that something is "raising the volume knob".
The reason I like this better than entering in the fader automation is because changing the bottom of the target range is so tweakable to hit that LUFS target. Then all the gain riding just readjusts automatically.
At the end of this process, you will have the highest dynamic range (in db) for the exact target LUFS.
Last edited:
dyross
Senior Member
Sorry it took so long for a response.
I've come around to the idea that this thing is really good for mastering classical pieces (or music with a large dynamic range). I might be saying things that are true about other compressors because I haven't used a wide variety of them, but here's what I'm finding very useful about this.
Let's say you have a piece of classical music that has both extended quiet periods and periods at the top of their volume. And let's say you are trying to hit an average LUFS target for Apple or Spotify (different standards). Well, first, you can have this compressor learn the LUFS range of your piece by playing the whole thing through it in "learn" mode. Then, you can raise the top of the target range and play back the peak of your piece. Adjust the top of the target range so you aren't going above 0 db. And have a limiter after it to help with errant peaks that you might miss. But get it as close as you can.
Now, set the target range to the same, top and bottom. Then, set the bottom of the target range to about 1/3 of the way to the top of the target range. Set the lookahead, attack, and release to some longish value. You want it long enough that you can't really detect that it's happening. You want to preserve the local dynamic differences as much as you can. Mine is set to 500ms and I don't really notice it. The narrower your target range, the more time you are going to need for the gain-riding to be less noticeable since there is more ground to cover as the compressor adjusts.
Now, play your piece all the way through, using some tool that will give you the average LUFS (I'm using Sonible true:level). If you need a lower average LUFS, drop the bottom of your target range which will bring down the overall average LUFS. Do the opposite to increase it.
Here's an excerpt of my mockup using three different target ranges.
Here's the widest target range:
View attachment SA Wide.mp3
Here's the narrowest target range:
View attachment SA Narrow.mp3
And here's the middle ground (where I was able to get my target LUFS without peaking)
View attachment SA Mid.mp3
I probably could have made the attack/release/lookahead times even longer for the narrow because I can still percieve that something is "raising the volume knob".
The reason I like this better than entering in the fader automation is because changing the bottom of the target range is so tweakable to hit that LUFS target. Then all the gain riding just readjusts automatically.
At the end of this process, you will have the highest dynamic range (in db) for the exact target LUFS.
This is an awesome post, thanks so much for writing it up and showing the examples.
I asked the plugin author some questions in the Gearspace thread on this page and they chimed in with some answers: https://gearspace.com/board/product...-apu-loudness-compressor-12.html#post16986231
I agree with your assessment - this thing probably doesn't sound much different than what other compressors could sound when dialed in properly. But it makes dialing in for this use case really convenient.
By the way, have you experimented with dry/wet parallel compression in this scenario? While it makes dialing it in a bit trickier, and probably doesn't hide the gain riding itself much, I wonder if it hides the attack / release artifacts at all.
wing
Senior Member
Really cool mockups @mbishop ! I think they are a great test of what this plugin can do, and thanks for sharing so much detail. I do a lot of mockups in Dorico and usually slap a limiter on the output bus to raise everything from the default NotePerformer levels which are rather quiet, just to make it more presentable on YouTube/Soundcloud etc. Seems like this plugin might be a good one to try for simply getting everything louder in a transparent way while preserving my dynamic range.
mbishop
Senior Member
Thanks for pointing me to that thread and all the interesting comments! Love that the developer is responsive.
I haven't experimented with the dry/wet knob at all and unfortunately, I really want to finish and post this mockup this week (!) so I probably am not going to have time for a lot of experiments until after then. However, feel free to take the audio that I posted of the "wide" version and see if you think it makes a difference!
If it does, that would be pretty interesting. I might slow down my attack/release times even more. I want to leave the transients completely unaffected. And if the listener notices the volume knob going up or down, then I feel I haven't set it correctly.
The only reason I even went down this path was because I didn't want to have to change the volume knob in my car while I was listening to my mockup (and my car has noisy tires!).
Thanks for the kind comments on the mockup! I think you are right that it could be of use with NP. NotePerformer already has a 1 second delay, and this would be adding more, but in some ways, I probably wouldn't mind actually since I'm not playing anything live with NP anyway. I'll try it on my next piece!
jbuhler
Senior Member
I've often thought there should be players that do car mixes with this sort of compression. For that matter even playing it on the stereo at home when someone is mowing or leaf blowing in the neighborhood... I often have to do quick compressed remixes for film screenings in class, because the blower for the heating and cooling system is so loud that it makes anything at even a moderately low dynamic level impossible to hear...
tc9000
Absolute Member
APU Software also has a Dynamic Optimizer plugin:
The APU Dynamics Optimizer is a loudness optimization tool designed to support modern loudness types (LUFS, True Peak). Underneath the hood, the optimizer uses the same compression engine as the APU Loudness Compressor, but benefits from the full audio waveform instead of real-time measurements. This allows the gain envelope to be meticulously optimized across the entire waveform.
This tool is especially well suited for mastering and post-production workflows, where the goal is to optimize the loudness of an entire waveform. Once the gain envelope has been optimized, you can playback the result in real-time while adjusting the character of the sound with the same collection of parameters from the compressor plug-in.
It's odd that it's not linked or mentioned anywhere on the site. I only found it via the updates page in the plugin.
The APU Dynamics Optimizer is a loudness optimization tool designed to support modern loudness types (LUFS, True Peak). Underneath the hood, the optimizer uses the same compression engine as the APU Loudness Compressor, but benefits from the full audio waveform instead of real-time measurements. This allows the gain envelope to be meticulously optimized across the entire waveform.
This tool is especially well suited for mastering and post-production workflows, where the goal is to optimize the loudness of an entire waveform. Once the gain envelope has been optimized, you can playback the result in real-time while adjusting the character of the sound with the same collection of parameters from the compressor plug-in.
It's odd that it's not linked or mentioned anywhere on the site. I only found it via the updates page in the plugin.
dyross
Senior Member
Wow, this looks cool! Guessing the dev isn't quite ready to release it, but seems imminent.APU Software also has an optimizer:
The APU Dynamics Optimizer is a loudness optimization tool designed to support modern loudness types (LUFS, True Peak). Underneath the hood, the optimizer uses the same compression engine as the APU Loudness Compressor, but benefits from the full audio waveform instead of real-time measurements. This allows the gain envelope to be meticulously optimized across the entire waveform.
This tool is especially well suited for mastering and post-production workflows, where the goal is to optimize the loudness of an entire waveform. Once the gain envelope has been optimized, you can playback the result in real-time while adjusting the character of the sound with the same collection of parameters from the compressor plug-in.
Ut's odd cos it's not mentioned anywhere on the site. I only found it via the updates page in the plugin.
mbishop
Senior Member
Way cool! That’s probably what I should have been looking for when I tried out the compressor. I will be watching this and maybe even try it out (but after my deadline is over!)
Pier
Senior Member
Compressors are really for handling peaks and transients.
Maybe instead of using a compressor just draw the volume curve? It's more tedious but you have full control. In the analog days this was called gain/fader riding.
Another option would be a gain riding plugin which adjusts overall volume but doesn't handle transients. There are tons of options like Waverider etc but it's never going to be as good as doing it manually.
dyross
Senior Member
That's generally true, but Loudness Compressor isn't a classic compressor. It has, for example, momentary and short-term LUFs detection circuits rather than traditional peak/RMS detectors.
I think there's an argument that an automated approach is more flexible in that it's easier to tweak (without having to input in a whole new gain riding "performance"). The workflow @mbishop posted above demonstrates this flexibility.
Exactly, I think you can think of Loudness Compressor, in a mode like this, more analogous to something like Waverider, but hopefully better sounding.
caustik
APU Software
yep, I posted it on the beta site to start getting some quick feedback. Found a couple issues which should be fixed by tonight or tomorrow.
The new version will fix some audio glitches which tended to happen around lower thresholds, and there is now a built-in limiter. The limiter can be configured for either peak or true peak, with configurable look-ahead separate from the compress/expand look-ahead. It works by sitting between the compress/expand envelope and the attack/release ballistics, adjusting the gain ramps the minimal amount necessary to avoid the peaks.
Glad to find this site.. it's snappy and clean and you all seem cool.
mbishop
Senior Member
Welcome! You've written a really useful tool.
wing
Senior Member
Welcome! By the way in addition to these cool plugins, I love all the artwork of robotic owls listening to headphones on your homepage
As they say in Twin Peaks...
caustik
APU Software
I think it would still tend to be "directionally correct" at the macro scale, because the sign of the gain delta ("attack" or "release") would be the same. It just wouldn't "get there" on time. It would probably also tend to be disrupted by the small scale dynamics around the gap duration (look-ahead time - attack time).
mbishop
Senior Member
What about the other way around, is there an advantage to having a longer look-ahead time than attack time? If not, it might be kind of nice to be able to optionally link those two settings together. I find myself moving both of them in tandem but I didn't know if there were reasons why I might not want to.
caustik
APU Software
Interesting idea. I do still use faster attack/release pretty often with longer look-ahead, but I've never done any strict sort of experiments to fully understand what the consequences are. It usually seems to liven up the dynamics a bit more, probably some of that is basically noise though.
Maybe something like little "[A]" and "[R]" toggles to the left/right side of the look-ahead text box?
