Voxengo Elephant 4.0 mastering limiter plugin released

[Dr.Gunjah]

Very weird.. are you sure you downloaded the latest v4.0 version? Those spikes when tweaking parameters happened in the old version (or at least you got mutes happening) but the new version at least on Windows running Reaper is silky smooth (also smooth in FL Studio).
I'm also a bit skeptical about the loudness thing as I've yet to hear a track where Elephant couldn't match the loudness or surpass it in a cleaner fashion than some of the competition. Of course if you are not afraid of distortion then you can perhaps get a bit louder with some of the other options but do you really need that much of it?
Care to post an audio example of the track?

Yeah I'm sure it's v4. I tried to set the automute to +3db but it still mutes too often. It happens mostly with extreme settings. I'm on reaper x64 / windows7.
As for the loudness - It's not only about the track I mentioned first. I tried some other tracks (all edm) and Elephant seems to work better with sparse arrangements and not so extreme settings. It does a good job preventing kicks from distorting, as long as you don't be greedy with loudness. Though at DR levels of 5 or lower I can't get it to sound good (yet). Pro-L tends to distort the kick though once you find a setting that works for the kick, it very often works for the rest of the track too; unfortunately this does not work with elephant. It sounds either distorted or squashed; Pro-L seems to more forgiving here. Will have to test it some more though.
I have to admit I'm a bit lost sometimes with all those different settings. Although I read the manual twice I'm not quite sure if I understood all of them completely and then there are the EL modes that are totally "black box" to me. Kind of overwhelming.

[Agreed]

Aleksey, I pretty much take your word as gospel on this because I use a number of your plugins (special shout-out to GlissEQ, that thing is BLISSFUL, but it's not the only one I've paid dues for a license ) - may I ask, why do you think a limiter would, well, misrepresent its release like that? What does the benefit the end user, exactly? I am just trying to wrap my head around why PRO-L would have that behavior... Why the developer or developers would think "This is the best way to represent our release time, by showing the halfway point of the actual release but hiding the real number from folks." What does that encourage when using the plugin that just showing the true release time would not?

I think Elephant is an excellent plugin, I will say that it is not my mastering limiter of choice but that's not because it is poor quality, I just already have other tools that I am very familiar with and "quick" testing with Elephant does not suggest that I will get anything more (which is not to say that I would get anything less, mind you, but these all boil down to questions of integration into workflow, and I am already ultra familiar with the aforementioned other tools, so learning a new limiter would be time somewhat wasted, even if it is a very good one).

If you don't want to comment out of respect for their development strategy or whatever, I can understand, but it's very curious that such an important number would be obfuscated like that on something that isn't at all a "character" device, but rather a precision tool... I expect my limiting and clipping plugins to accurately represent themselves, I don't use PRO-L and hearing that I am frankly glad that I don't because it does not make sense to me.

[Aleksey Vaneev]

I do not know why PRO L displays release in at least its Allround mode like that. The test is simple - boost by 9 dB a -2dBFS sinewave whose loudness is quickly reduced by 12 dB after 1second of playback, you'll see the release shape clearly.

[PatchAdamz]

Oh please! Check the demo guys. I don't want anybody to be taking only my word for it.. but yeah, it's an awesome limiter!

Good idea

[Shy]

This two files are Pro-L and Elephant.. IMO they are comparable and I cannot tell a big difference. For better comparison both are in a lossless format (Wave)

Elephant
Pro-L

Please post the original. In both examples, the "snare" is ruined, but much less so in the Elephant example.

[electro]

The least destructive method for maintaining Snare is the straight clipper which introduces distortion. Brickwall limiting will always changes the sound.

[Shy]

A good and suitable brick-wall limiter can maintain the original dynamics better than a plain clipper, which at that amount of gain reduction, would "flatten" the dynamics (transients specifically) more in comparison and result in a sound that's farther from the original, and likely with more overall perceivable distortion in the mix, so it would be more destructive.

[4damind]

This two files are Pro-L and Elephant.. IMO they are comparable and I cannot tell a big difference. For better comparison both are in a lossless format (Wave)

Elephant
Pro-L

Please post the original. In both examples, the "snare" is ruined, but much less so in the Elephant example.

This snare has not much attack and has a distorted release...it needs an envelope-shaper to form a bit the attack and make the release a bit shorter). Anyway, this is not the effect of some of the limiters, it's the original sample of the snare.
Elephant can sound a bit snappier/punchier so this could be the result that the Snare sounds a bit different.

[Shy]

No, it's the opposite, Pro-L is the limiter that added "punch" to the snare in that sample in some parts, but not in any useful / good sounding way, just in a way that makes it sound less stable, and it made the snare quieter in other parts, again, a less stable behavior.
If you post the original sample, I could show you a much better result than both. But it doesn't matter much, I guess, if you don't mind the differences.

[4damind]

Could be the "Dynamic" mode of the Pro L which was used...
"By enhancing transients before actually applying limiting, this algorithm excels in preserving the original punch and clarity of your audio."

This example is using the "transparent" mode... (I exported both again, because I made some changes):
Pro L (Transparent Mode)
Elephant (EL 4)

[Uncle E]

I wonder how Elephant compares to the Sonnox Limiter in terms of sound, not sheer amount of options. Any ideas?

The Sonnox has a more distinct sound. It's great for when you're looking for that sound but it's one of the less transparent mastering tools around.

[Tapehead]

I wonder if Alexey would consider making a feature reduced version of Elephant? Something that would primarily be aimed for track limiting.

It should have a ceiling control that could go all the way down to -30 and another control for makeup gain and that's it. Any chance to make it happen?

[Uncle E]

I wonder if Alexey would consider making a feature reduced version of Elephant? Something that would primarily be aimed for track limiting.

It should have a ceiling control that could go all the way down to -30 and another control for makeup gain and that's it. Any chance to make it happen?

Agreed. A simplified "Little Elephant" that functions like the L2 would be a pretty neat tool, particularly now that it's got the EL-4 algorithm. A lot of major producers and engineers are using the L2 as a channel limiter now, not just on the master buss, and I could see the same thing happening with the "Little Elephant".

[bmanic]

I do not know why PRO L displays release in at least its Allround mode like that. The test is simple - boost by 9 dB a -2dBFS sinewave whose loudness is quickly reduced by 12 dB after 1second of playback, you'll see the release shape clearly.

None of the attack or release numbers make any sense in Pro-L. I've said this time and time again.. personally I was trying to get FabFilter to show the attack and release knobs in % instead of a number that makes no sense.

The algorithms in Pro-L are extremely program dependent.. at least the two I made. You can not simply "measure" them with a simple static wave nor with short impulses. There are too many dependencies going on behind the scenes. The attack and release times are basically continuously modified. This is partly the secret to the "naturalness" of some of the algorithms. A lot of trial and error and a huge amount of time went into the designs.

Also note that the loudest and perhaps most transparent algorithm in Pro-L is NOT the default one. AFAIK most people when going for loud use the 'transparent' algorithm or the 'dynamic' one. Both of those are so ridiculously program dependent that if you figure out a way to properly measure the real timings that are going on under the hood then I owe you a beer.

Anyhow, where I think Elephant v4 is superior to pretty much any limiter at this point is the way the material being limited stays intact. It just sounds "more high fidelity" to my ears. It's such a cliche and I can't even pinpoint what happens or why it happens but Elephant just sounds more coherent and more "3D".. or let me put it this way: There's less damage done to the original audio and thus during hard limiting the "3D-ness" of the sounds stay intact (front to back depth coherence).

People who can't get Elephant loud enough should try the settings I posted on the first page. They are rather brutal but if you can't get it equally loud to other limiters with those settings then there must be something weird going on. I have absolutely no problem making Elephant as loud as Pro-L and usually with less loss in fidelity.

Cheers!
bManic

[bmanic]

One more thing: If the loudness wars would continue (which they will not, fortunately!) I guess the future of limiters would be some kind of predictive model. The only way to do a proper predictive model would be to investigate what happens at the D/A converter stage.. to measure the actual analogue signal and see what is the optimal way of shaping a signal before it gets chopped off to get the ultimate volume out of the converter.

This of course is not trivial at all and would require some damn expensive analysis equipment and a lot of different converters and then build a bunch of algorithms around these different converters and perhaps narrow it all down to a few generic best cases.

.. but luckily we don't have to worry about that any more. The loudness war is going to die soon due to the aggressive legislation that is happening all over the world. Flat, clipped tracks will sound much less loud than tracks that retain their dynamics so this will most likely reverse the trend soon (well, perhaps within the next 5 years or so).

Cheers!
bManic