Voxengo Elephant: best algorithm for solo piano?

[D2sX9ek8w3]

Hi,
I am trying to get a solo piano piece to playback nicely when distributed at -14dBLUFS integrated with -1dBFS max peaks.

I have balanced the bass and treble to sound good in my studio, home entertainment system, car stereos, etc., but the percussive treble melody starts to blare or splat very subtly on cheap headphones when the mp3 version is driven into the too-loud territory.

I usually work with large ensembles or at least mellower-sounding solo instruments, so I have rarely scrutinized a solo instrument with the sort of detailed analysis demanded of a solo piano.

I think it may be helpful to studiously audition the output of the different Elephant algorithms, so I wondered if anyone else has already been down this road.

I'd like to finish off this project by making the tone idjut proof in case some one turns up their mp3 player's cheap headphones too loud.

Thank you!

[Ploki]

-14 LUFS can be enough that the end system itself starts to distort.
in any case, Elephant would not be my first choice for finishing a solo piano piece, i usually go for Pro-L2 in such cases.

I'd go with EL-5 or AIGC tho.

[D2sX9ek8w3]

Thanks for sharing your thoughts.

I've used Voxengo Elephant since 2010, and FabFilter Pro-L/L-2 since 2013.

I tend to use Elephant on the master and have generally been very happy. For reasons I can not clearly articulate, I rarely use Pro-L2, and when I do, it is usually on an individual track. FWIW, I rely on Pro-Q and Pro C and could be considered a Fabfilter fanboy.

I agree that solo piano probably should not be distributed at -14LUFSi but if that is how it is going to stream I want to do my best to send it out clean. If this were going to a CD I would just make this song quieter.

I recalibrated my studio playback for this project from 83dBSPL to 77dBSPL, so I could listen at a level nearer to what I imagined the intended audience might enjoy.

I have 11 other solo piano pieces in this album, and they are all doing ok.

This particular piece has a funeral dirge pacing. There is a lot of space and sustain, punctuated by chord blocks and expressive melody.

The voids in this particular music seem to be adversely influencing the LUFS integrated measurement, so when I increase the overall levels to make the arbitrary -14 readout, the punctuated portions seem too loud.

I know loudness leveling always works like this, but I think this is an exaggerated example of the circumstance.

I had been using EL-4 out of habit.

I just purchased an upgrade to Elelphant V5 and am trying EL-5, and AGIC as well.

If that does not help, I'll give FF Pro L-2 a try.

Thank you.

[cdbaksu]

I found Apogee's Soft Limit to sound beautiful on solo piano.

[soundmodel]

I agree with some of the other posters. Elephant is not really a transparent limiter, but a very characterful one.
I would not consider it as a first option for acoustic instruments. It may work, if you chain several of them at very mild settings.

OTOH, a solo piano sounds much more like the ballpark of iZotope Ozone, Waves L2/L3 or any LA-2A model.

[Frankie.T]

Voxengo Elephant can be transparent as hell if you know how to use it, is probably the best limiter out there and there is no reason to don't use it everywhere a limiter is needed on any track and genre, you just need to learn how to, since it's very deep.

For a cleaner limiting try EL-UNI, then use the cog to access the advanced mode enable DRC on RLS mode, turn the Tr Shape to -1 place the channel link around 50%, the knee at 100 and Rls shape between 1 and 2.
Set Trans as Classic, enable lookahead, at middle should be good enough

I hope this helps

[D2sX9ek8w3]

Hi Everyone,
I think I met my match on this song.

I went back and studied some of the previous mix renders and accepted that I just cannot get this particular song to -14dBLUFSi with the mix I wanted.

The song is in D minor. The sympathetic resonance of the piano's strings and soundboard tends toward a discordant character.

In my previous mixes, I used a lot of reverb but found that the resonance of the strings decayed into strange comb filter effects. So, on my latest mixes, I pulled down the reverb and minimized any movement in the tail of the reverb.

As the mix dried out, it became very difficult to achieve the -14dBLUFSi level without having the transients push against the ceiling. The reverb had been acting as filler and satisfying the LUFS meter. The distortion I was hearing was not obvious until you pushed the playback levels and listened for it, but it bugged the snot out of me.

I have settled for restoring the reverb, and adjusting the mix level. Now, the transients sound graceful, and the sympathetic resonance is what it is.

This has been quite an exercise for me. I usually work with content that has a full arrangement, so achieving -14dBLUFSi is fairly easy. In the past I felt at liberty to simply set the levels on a selection in an album where it seemed appropriate. Streaming distro makes for new challenges.

Thank you.

[D2sX9ek8w3]

...For a cleaner limiting try EL-UNI, then use the cog to access the advanced mode enable DRC on RLS mode, turn the Tr Shape to -1 place the channel link around 50%, the

knee at 100 and Rls shape between 1 and 2.
Set Trans as Classic, enable lookahead, at middle should be good enough

I hope this helps

Hi, Thank you for the suggestions. I will give this a try, as I want to use this opportunity to study the character of Elephant's different modes more closely than I have considered them in the past. Previously, I have been pleased to use whatever Aleksy offered as his latest greatest.

I have come to believe that the nature of the program material caused my concern more than the character of the limiter.

Nevertheless, the material provides a good opportunity to see how the limiter treats the signal.

Thank you!

[dyross]

-14 LUFS can be enough that the end system itself starts to distort.
in any case, Elephant would not be my first choice for finishing a solo piano piece, i usually go for Pro-L2 in such cases.

I'd go with EL-5 or AIGC tho.

Which Pro-L2 algorithm would you use?

[Ploki]

-14 LUFS can be enough that the end system itself starts to distort.
in any case, Elephant would not be my first choice for finishing a solo piano piece, i usually go for Pro-L2 in such cases.

I'd go with EL-5 or AIGC tho.

Which Pro-L2 algorithm would you use?

I’d start with safe and transparent.

Agressive can be extremely transparent as well despite the name

[bmanic]

The Safe algorithm in Pro-L is exactly what it implies in it's name. It is virtually impossible to get it to distort. However, this does mean that it is also impossible to get to huge amounts of dynamic range reduction. It has a very clear "ceiling" where it simply stops.

It demonstrates the "physics" of limiting very elegantly. You can sort of use the Safe algorithm to judge if your mix can go loud or not.

For solo piano at -14 LUFS I would definitely stick with the safe algorithm. Also, make sure you understand LUFS correctly. For instance, it is NOT recommended to use 'Integrated' LUFS to measure and compare two different pieces. Especially if one is a quiet piano piece for most of the run time if it has just a few fortissimo passages. It's integrated LUFS will be completely skewed when compared to another piece that plays moderate loudness throughout. Instead you should be comparing the loudest passages of each and check the short term LUFS measurement.

As for Voxengo Elephant, I suspect the AIGC algorithms should be able to be really transparent too without any risk of distortion, if you tweak them accordingly.

[Ah_Dziz]

I've usually had the best luck buying some extra decibels on solo piano pieces with a combination of very subtle compression and taming some resonance with a nice surgical EQ. I've not had great luck with "modern limiting" on piano unless it was for a specific "over the top" effect that was called for, or to make something sound "degraded" on purpose (like smash the hell out of it like it had been mastered straight to vinyl).