LA2A, Fairchild, 1176LN, Neve 1093, ... plug in, plug away

[mangel1234]

The Liquid Mix seems to be getting good reviews in the magazines lately, though I thought there were some issues reported early on. It sounds great and there's a new lower-end version without the screen coming out in the near future. Looking into this and loving that it has Distressor emulations.

From what I understand, there were some issues with the first drivers that are supposedly much better now.
The main thing I have found sonically, is that I can actually pick out most of the "real" versus "liquidmix" just listening on cheap computer speakers.
The LM versions of eq still sound great to me (via cheap computer speakers) but have a little bit too chipper a sound. The compressors I can tell immediately bc they all have one overall quality to them that says "sound engine" to me, and a part of that being that it is a little nicer sounding, a little less bent out of shape... in a good way.

But, with the new smaller version, it is hard to beat $500 for a very nice emu of all comps and eq's, even if they were just a software collection powered by a dongle+, quite a great bargain. Nebula is still the way to go if long-term absolute quality is the interest, along with free additional programs and inexpensive commercial collections!!!

[Compyfox]

To revive this thread, there's a new addition to the plugin series.

For example, Abbey Road finally released their Mastering Pack. So did Sonalksis (which is not really an emulation of something from the specs, but worth a mention nonetheless).

And additional to IK Multimedia's already known analog plugin set creation, with no relation to real hardware, T-Racks. This set got an overhaul. T-Racks3 now features an emulation of a Fairchild 670 Compressor, and a Pultec EQP-1A Equalizer.


There're also a couple of emulations redone by GSi (formerly known as Soundfonts.it), but these are mainly instruments, a reverb and a delay. However that topic would go beyond the scope of this discussion.

[Per Lichtman]

Slight clarifications.

Regarding Compyfox's comments:
Abbey Road's TG Mastering Pack set of EQs were recently ported to include more plug-in formats (previously RTAS only), in addition to their compressor and specialized box emulations that had been ported already.

As mentioned, none of Sonalksis' plug-ins model specific gear, and the new mastering suite is designed with a rather "modern" style.

Regarding the Liquid Mix, it models the "behavior" of the different compressors and EQs rather than the "coloration", in contrast to several UAD and Nebula programs.

Also, in case they weren't mentioned already (it's been months since I last contributed to this thread) the URS' Saturation plug-in models several pieces of gear, including a Motown preamp and the URS M Series EQ bundle models a Motown eq and a some sort of "vintage cinema eq".

The Waves JJP Collection models Fairchild 660/670 and two Pultec EQs as well.

[Compyfox]

Thanks for contributing.

Seeing KVR as mainly an infopool for VST/AU plugins, I thought it's okay just just say "finally available", but you're right - they're finally ported to the native realm and were available for a while already for the ProTools Platform.


Can you maybe go a bit more in-depth what the preamp models in the URS Saturation plugin are? And what do you mean by "vintage cinema EQ"?

And do you maybe know offhand, what the two Pultec EQs are in the Waves JJP Collection?

[jupiter8]

And do you maybe know offhand, what the two Pultec EQs are in the Waves JJP Collection?

Pultec EQP-1A, and the Pultec MEQ-5
as stated at
http://www.waves.com/Content.aspx?id=7306

[Zaphod (giancarlo)]

Regarding the Liquid Mix, it models the "behavior" of the different compressors and EQs rather than the "coloration", in contrast to several UAD and Nebula programs.

I liked very much your post!
Only I don't understand this passage, my question is: what do you mean exactly, what's the difference between coloration and behaviour in equalisers?
I try to explain better our procedure: devices were sampled and checked for each possible combination of knob position, even comparing harmonic distortion graphs and using state-of-art converters. Kernels in nebula are very long (this is a limitation in liquidmix), till several SECONDS, it means that in nebula a close emulation of narrow q is possible now (it's the first time I use this expression: "a close emulation"). You could add the dynamic variable if you want (and it was done sometimes even for equalisers). I guess (but I didn't check it) that parametric equalisers in liquid mix have some limits in terms of dynamic behaviour, simply because there cant' be so much RAM/ROM in the hardware world.

[bmanic]

Regarding the Liquid Mix, it models the "behavior" of the different compressors and EQs rather than the "coloration",

Actually, it's the other way around. At least from my experience. It models the "coloration" of compressors quite well but NOT the "behavior" (the way gain reduction happens is far from perfect). The EQs model "both" quite well.

Cheers!
bManic

[Chopper]

Regarding the Liquid Mix, it models the "behavior" of the different compressors and EQs rather than the "coloration",

Actually, it's the other way around. At least from my experience. It models the "coloration" of compressors quite well but NOT the "behavior" (the way gain reduction happens is far from perfect). The EQs model "both" quite well.

Cheers!
bManic

mmmmh... I'm somehow between you both, guys.

the eqs always have the overall tone of the emulation to me... as if there would be some sort of other eq coloring whatever I want to do with my settings, but the colour in itself doesn't "change" the way the original hardware's colour would change with different settings. I don't really know how to explain it.
it's not a bad thing, just different.
as far as the compressors go, they definitely got different tones, and although it doesn't always behave like the hardware it is supposed to emulate, the results are noneless very good and the possibilities amazing. but it behaves very diffently than either plugins or hardware for me. it's in a world of its own.
still one of the biggest bargain in the DAW world.

[bmanic]

Actually, I'm not sure how they create the programs for the EQs in the Liquidmix. I know how they are done in the Sintefex FX series (which the liquidmix technology is based on). There they simply sample every possible setting. It takes many weeks but the end results are that you get dynamic "coloration" (that is, it's dependent on input gain) and every possible EQ setting so in essence, it is a virtual copy of the original EQ. It must be a bit different in Liquidmix because you can blend various EQ models with each other.

The distortion which is created by the dynamic convolution technology is pretty far off from the original gear though. To get more accurate distortion modeling you need a system like Acustic Audio is offering with Nebula.

I'll try to post some examples at some point of the API eq, which is available for both the Nebula plugin and for Sintefex FX series. I'm sure it's also available for the liquidmix.

Cheers!
bManic

[shogger]

I'll try to post some examples at some point of the API eq, which is available for both the Nebula plugin and for Sintefex FX series. I'm sure it's also available for the liquidmix.

Cheers!
bManic

Oh yes, please do so. That would be very interesting!

Shogger

[Per Lichtman]

Okay, allow me to clarify my comments about the Liquid Mix and Nebula.

Focusrite has other products, such as the Liquid Channel and the Liquid 4Pre, that model gain stage emulation and use very expensive systems to do so. If you look at the Liquid Channel it models the transformer behavior by changing power variables internally among other things. Part of the reason that Liquid Mix costs so much less per channel than either of these is that it very explicitly does not model such functionality. This was confirmed for me by a Focusrite representative last year when I was still trying to decide whether to get the Liquid Mix. This eliminates a huge part of the coloration and distortion from the sound. I've used the UA 610 and 1176 in a channel 6176 channel strip and compared it to the Liquid Mix and you just can't get the same sort of coloration.

Now in terms of the compressors, the Liquid Mix cannot get the exact behavior of all the gear modeled and the more extreme the settings you are using, the more you'll notice it. But it does model a wide variety of compressor behavior and they tend, to my ears, to be a little brighter and clearer than much of the vintage gear modeled (likely in part because of the lack of gain stage emulation). This means that I actually tend to reach for them when I am looking for less coloration than, for instance, some of the UAD plug-ins such as the full Neve 33609.

Some Nebula programs intentionally leave harmonic distortion out as well, but many model this coloration and there is an entire section of the presets dedicated just to modeling gain stage coloration, with and without distortion.

The Abbey Road plug-ins also have some gain stage modeling. In one of the EQs, for instance, raising the gain knob produces a different sound than raising the volume in using the fader in your DAW either pre or post EQ.

[Per Lichtman]

Regarding the Liquid Mix, it models the "behavior" of the different compressors and EQs rather than the "coloration", in contrast to several UAD and Nebula programs.

I liked very much your post!
Only I don't understand this passage, my question is: what do you mean exactly, what's the difference between coloration and behaviour in equalisers?
I try to explain better our procedure: devices were sampled and checked for each possible combination of knob position, even comparing harmonic distortion graphs and using state-of-art converters. Kernels in nebula are very long (this is a limitation in liquidmix), till several SECONDS, it means that in nebula a close emulation of narrow q is possible now (it's the first time I use this expression: "a close emulation"). You could add the dynamic variable if you want (and it was done sometimes even for equalisers). I guess (but I didn't check it) that parametric equalisers in liquid mix have some limits in terms of dynamic behaviour, simply because there cant' be so much RAM/ROM in the hardware world.

Thanks for responding to my post. The difference between coloration and behavior in equalizers, as I use them, are primarily as followed.

Behavior is modeling the curves for each of the bands, filters and shelves. This includes "indirect frequencies" such as how some vintage models would modify frequencies that "had nothing to do with the target". For instance in the Liquid Mix EQ model "Old Tube 3", every band you activate affects the very highest and very lowest frequencies just by being activated but the exact behavior varies depending on which band you are dealing with. The plot of the EQ curve shows them both dropping off more and more with each band you activate, even if you haven't applied any gain or attenuation. In the example of "Trany 5" boosting the bells won't just peak near the target frequency and then attenuate until the reach the neutral "0 db" line, but rather create two more peaks, one above and one below the target frequency. Sort of "secondary bells" that are tied into the main ones. All of this is behavior.

Coloration are the additional variables that affect the sound you run through a piece of gear. Harmonic distortion is one of the most popular ways to differentiate the sound. In the case of digital gear, ADC and DAC behavior may also be part of the equation. Noise floor can also be part of the equation. There a lot of potential variables, many of which are not modeled or sampled much of the time. Morgana's 112 db models some of these variables in the sampler emulation to achieve the "low-fi" sound they wanted.

[Zaphod (giancarlo)]

a ok, thank you

[Per Lichtman]

No problem.

It can be a little confusing because on gear that isn't thought of has having a curve (for example compressor) sometimes coloration can refer to simply applying a static curve to either the input or output. And in the case of EQs where you don't have the option of switching bands on and off, then it sometimes can seem to refer to the difference in the sound depending on whether the unit is on or off, as opposed to when any of the bands are modified. So there is that big huge gray area. But the definitions I used are meant to highlight the "black and white" as opposed to gray areas.

[bmanic]

Well you mean exactly the same thing as me then. Behavior = the results the unit is theoretically capable of (eq curve, compressor transfer curve and attack/release shape etc).

Color = harmonic distortion, additional compression, hiss/noise/hum, additional dynamic frequency changes (tube being cold/warm) etc.

So, in the case of compressors on units that use Sintefex patented dynamic convolution, you get a semi-accurate transfer curve (it's not 100% because the transfer curve itself can be program dependent, like on a vari-mu compressor) and not at all accurate attack/release curves. Basically it seems to be a standard compressor that "reads" a look-up table. After this you get the dynamic convolution process which has additional compression happening and the dynamic frequency changes (if there are any).

Conclusion, the "color" part is more accurate than the "behavior" part, at least in my opinion/experience. Anyhow, it's still a useful technology and I use it all the time. I do see it becoming completely obsolete though.. once Nebula technology matures some more.

Cheers!
bManic