Truth behind Waves L2

[ahjteam]

So, THE reason why I still prefer Waves L2 over say Slate FG-X is because it doesn't make my mix sound distorted, and I wondered why for long, so I decided to make a scientific test. I made a really short clip of 22khz squarewave at -6dBFS level and then made 20 samples go to -0.0dBFS.

If someone wants to try the test out with your own weapon of choice, the file is here:

http://dl.dropbox.com/u/1338211/sqr2.wav


























Then I used L2 with these settings and recorded the results:
Otherwise default settings except threshold at -5.9dB and dither off

In the ideal world, the result would be a straight square, no distortion.



























And the result was this

I checked from the plugin settings that the plugin has 64 samples of latency and when I zoomed in, that the ducking the L2 did before and after the threshold is exactly 64 samples long. So how long is 64 samples? at 44.1khz samplerate it's 1.45ms, so not noticeable.





























For reference I did the same test with Slate FG-X with seemingly same settings (dither off, comp off, Gain +5.9dB), and I also noticed immidiately why I don't like. The picture on the left is L2, the picture on the right is FG-X

It causes distortion, and when I was using it on musical context, it was really audible.




























And here is the same test with IK Multimedia T-Racks Classic Clipper (gain +5.9 and slope 0.0 (hard), output 0.0). It became pretty close to what I was looking for.

And with the slope on the other edge (slope -12.0 aka soft slope):

And same settings except +12dB gain

And that includes my rant from this evening. Thank you.

 

[acappa]

Man, you break my ass with this....I´m an entuthiastic FG-X user.....now I start questioning why I hear distortion sometimes in my masters...

 

[NSGUITAR]

Slate= NOT>Every+all the time

 

[mephisto_loh]

Pretty awesome tests. thank you for the time +1000.
In fact, i propose it can be a nicely task to make benchmark test for plug ins in that detail.

real vs expected and rate it out. Not to harm the reputation of one company or another, but for the sake of all the people who have a daily basis use of these tools.

 

[egan.]

Can you generate an FGX wave that actually looks like the others?

 

[Morgan C]

I thought L2 was the one that murdered snare drums? Or was that L1?

 

[egan.]

Both do IME. You can see why is the pics actually. The anticipation and release attenuate the entire attack and not just the offending transient.

 

[Kazrog]

Limiting = Compression = Distortion. Pick a flavor that works for the musical context.

People haven't been using L2 for years because it sucks - it's a great plugin. Abuse anything and it will sound awful, and the L2 has been used to ruin plenty of otherwise great mixes. FG-X adds another tool to the palette, which has a totally different sound to it, and obviously produces very different results - which in many cases I find to be more desirable than peak limiting. If you want a peak limiter, though, it's hard to beat L2 - there's a good reason why it's been such a staple.

BTW - a 22 khz square wave at 44.1 kHz sample rate isn't going to be able to produce its harmonic spectrum properly because it will lack the headroom (since any of the upper harmonics would be above Nyquist.) Probably not the best test - try using a much lower frequency (such as 440 Hz) and you might get more interesting data.

 

[egan.]

All good points Kazrog. With the L1/2/3 I was referring to attempts to use it as a primary limiter to get stupid modern levels. A limiter (no matter how fast) is a different animal than a clipper by definition. Obviously the L series was THE choice for a decade before we even saw all of these clippers.
Literally anything beyond a sine wave at 22khz isn't going to sound correctly at 44.1 but I think it does work to illustrate what Ahj was "bench" testing. It's obviously a terrible listening test choice for the reason you outlined.

 

[Morgan C]

Limiting = Compression = Distortion. Pick a flavor that works for the musical context.

That's true. It's funny, every time I mention clipping on another forum, people act as if I'm absolutely destroying my music and distorting it to pieces in the name of volume, when in fact clipping is the most transparent tool for loudness in a lot of cases.

 

[derfbonker]

.

BTW - a 22 khz square wave at 44.1 kHz sample rate isn't going to be able to produce its harmonic spectrum properly because it will lack the headroom (since any of the upper harmonics would be above Nyquist.) Probably not the best test - try using a much lower frequency (such as 440 Hz) and you might get more interesting data.

Not being a dick, and eel free to correct me if I'm wrong, but I didn't think a square wave would have harmonics. Isn't it a single frequency?

 

[B36arin]

A perfect square wave has an infinite number of orders of harmonics. If a tone doesn't have harmonics it's a sine wave, everything else has harmonics.

 

[egan.]

Not being a dick, and eel free to correct me if I'm wrong, but I didn't think a square wave would have harmonics. Isn't it a single frequency?

Yeah, you have it backwards. Sine waves are the only pure waves and all others can be created (or at least approximated) by the the addition of sine waves across the natural harmonic series.

 

[amarshism]

Have you tried with the ozone limiter?

 

[ahjteam]

Limiting = Compression = Distortion. Pick a flavor that works for the musical context.

People haven't been using L2 for years because it sucks - it's a great plugin. Abuse anything and it will sound awful, and the L2 has been used to ruin plenty of otherwise great mixes. FG-X adds another tool to the palette, which has a totally different sound to it, and obviously produces very different results - which in many cases I find to be more desirable than peak limiting. If you want a peak limiter, though, it's hard to beat L2 - there's a good reason why it's been such a staple.

BTW - a 22 khz square wave at 44.1 kHz sample rate isn't going to be able to produce its harmonic spectrum properly because it will lack the headroom (since any of the upper harmonics would be above Nyquist.) Probably not the best test - try using a much lower frequency (such as 440 Hz) and you might get more interesting data.

The square wave at pretty much fastest possible frequency (well, I could've used 22.05khz) is good for realizing the attack and release times as each period is only 2 samples long and the best part is that you can get that up to RMS -0dBFS I can do the next test in similar fashion with like 7350hz sine (6 samples per period), 4410Hz (10 samples per period) or 3675Hz (12 samples per period) 3150Hz (14 samples per period), 1575Hz (28 samples per period), because I want to keep the wavelenghts as short as possible so that you can see the distortion on sample level.

Have you tried with the ozone limiter?

Nope, but the link for the audio file is at OP, you can try it out yourself.

 

[Slate]

Hi ahjteam, thanks for posting. The FG-X (specifically the FG-Level module) is a tool, but its not a robot. You must understand its controls and how they interact with the audio. For one thing, there was no mention about the ITP value and there should be since I wouldn't use the default value on a square wave.. but I also don't usually master square waves since my current musical preference is saw waves (easy listening for early morning workouts). Furthermore, the ITP process was built and tested on modern musical mixes, and therefore, I'm not sure I understand how processing test tones with it will reveal anything practical about it's affect on actual music (although ahjteam does say that he has heard artifacts on music and if he posts the mix in question I can probably figure out why). Having said all that, I'm up for the following challenge:

Upload a music mix (with headroom) that needs loudnessizing. I'll make a master with FG-X, and you make a master with L2. My master will have no undesirable distortion. And it will sound more like the original mix than the one you used with the L2.

Wow that was a pretty arrogant and smug thing to say... but I've got the gloves on and I'm willing to go 12 rounds, so lets do this

 

[wishtheend]

I just got add that I love Steve's attitude when it comes to a challenge Someone should take him up on this, I woud love to hear a direct shootout on a sneap forum mix.

 

[Jarkko Mattheiszen]

Hi ahjteam, thanks for posting. The FG-X (specifically the FG-Level module) is a tool, but its not a robot. You must understand its controls and how they interact with the audio. For one thing, there was no mention about the ITP value and there should be since I wouldn't use the default value on a square wave.. but I also don't usually master square waves since my current musical preference is saw waves (easy listening for early morning workouts). Having said all that, I'm up for the following challenge:

Upload a music mix (with headroom) that needs loudnessizing. I'll make a master with FG-X, and you make a master with L2. My master will have no undesirable distortion. And it will sound more like the original mix than the one you used with the L2.

I really want to see this happen, since I'm considering the FG-X as I really liked the demo. I noticed it needs to be fed a balanced mix to get the most out of it, but, well, isn't that how it's supposed be?

Wow that was a pretty arrogant and smug thing to say... but I've got the gloves on and I'm willing to go 12 rounds, so lets do this

It's your product man, you should be arrogant about it (to some extent anyway)

 

[crillemannen]

Well the L2 sounds like shit anyway, distortion or not. Smothers out the whole fucking mix, destroying transients. '

Izotope Ozone is the one I like the most. Sounds waaaaaaay better then L2

 

[ahjteam]

Hi Slate, I used the default settings on FG-X with exception of the listed I said above.

I accept your challenge, I bounced a short 38 second clip without anything, with L2+L3LL ultramaximizer and with my settings of FG-X.

Dry 24-bit wav: (I actually think that the drums are Slate btw )
http://dl.dropbox.com/u/1338211/ahjteam/SLATE/ahjteam_water_v02_dry.wav
-20dB RMS with plenty of headroom

L2 + L3LL: (L2 threshold -12.2, out ceiling -0.5, dither off; L3LL treshold -3.0, out ceiling -0.3, "Loud and Proud" profile)
http://dl.dropbox.com/u/1338211/ahjteam/SLATE/ahjteam_water_v02_L2.wav
-9.2dB RMS

FG-X: (comp off, lo punch 1.0, detail 1.0, gain 12.1, dynamic perception 1.3, ITP -0.4, ceiling -0.5dBFS)
http://dl.dropbox.com/u/1338211/ahjteam/SLATE/ahjteam_water_v02_FGX.wav
-9.2dB RMS, and I hear audible distortion either on the kick or some other element with other ITP settings. Reason most likely _IS_ in non-optimal settings

I used audacity > analyze > contrast for analyzing the RMS values for the whole clip, the RMS reading on FG-X was higher.


edit: I just noticed that there is a missing crossfade and you can hear a pop on the left channel around 00:08 and a few times after that. my bad

edit2: hmm, I just checked the session and I didn't miss a crossfade as all the tracks except intro and leads were single full takes. I noticed that there is a pop on the miced guitar track, so it came there when tracking. Must've been a hardware problem with buffer, synch or something, so try to live with it.