Quote:
Originally Posted by
V4nger
Well... I just don't have the cash right now for DS1 so I've made myself become more happy with Elevate! ProL2 for clipping before limiting (dynamic mode is awesome for retaining transients.) and then Elevate. I can achieve similar loudness with less pumping as there is more control over the actual limiting process. But the compressor in DS1 is just insane. So damn good. Ultimate control. Gonna save up...
***EDIT*** Sorry to kind of highjack this thread with my own ideas/opinions. I'll consider removing my very long idea for a huge chain of several layers of limiters in parallel and in series... and posting that else where.
Much props to Weiss and Softube. Without even testing this on my own DAW, I've determined that it has made my top 4 list and is possibly the best. Extra props for the ability to be old school and still widely considered the best. My top
#1 is actually Elevate, even though it will be better once it gets implemented in better ways... ways that kind of do similar functionalities of both DMG Limitless and Weiss. Pro-L 2 is on my top 4 list, but I'd probably never use it (but will do a test/comparison with it soon). I used it, for a few months, about 4 years ago. Ozone replaced it for a number of reasons.
I haven't had a chance to try out the Weiss one... but I am strongly considering demo'ing it when I think I have audio material fro, clients, that would be best with it. I hear the beauty of its compressor isn't the compressor alone, but the way the algorithm is complimentary and sort of an inverted system when compared to the clipper/limiter algorithm.
But, we can probably find a single band compressor with a MB compressor (or both built into one) and figure out how to set it so that it compliments DMG Limitless and/or Elevate... to get max transparency, preservation of peaks & clarity/brightness, and the highest amount of gain reduction for the not-so-finished loudness wars. The issue here, however, is that the compressor within Weiss and the clipper/limiter are automatically going to compliment each other's behavior in algorithms, where as trying to mimic this with singular plugins that aren't made specifically for it but can be dialed in for it... would take a lot of time. Once the AI systems become more adaptable and advanced, then a system of complimentary functions within different compressors, deessers, clippers and limiters will be available.
Elevate probably has the most variations of styles and behaviors that it can possibly achieve, compared to anything else. I used it for a few days and decided that there were always 10 different coloration types that I loved & found it hard to decide.
I feel that Eventide's system of bands is the future, and once they work out things to take it to higher levels, they will be the best... for example, like how Weiss works under the hood so that the compressor and the clipper are like 2 opposite halves of one transparently adjustable algorithm, as well as the ways in which Limitless achieves transparency by meshing together broadband and multiband behaviors with pre-fast and post-slower limiters. Really, the magic will happen once they figure out the best ways to make their banding system and their multiple modules work together with AI adjustments for retaining character. Really, they get AI working so that Equiv is both static and dynamic, compressor, clipper set up as pre clips and post clips, and limiter (should be 2 limiters) are all working together and being adjusted according to user end of dialing knobs as well as intelligently reading the audio signal... and then dialed in with broadband functionality for each processor (like how DMG Limitless does with its limiters)... they will be surpassing the competition and the older styles of doing things... but then, when they also introduce 4 to 7 band systems, combined with the 27 or whatever, with broadband... no looking back.
If other companies are able to use the broadband "human ear" style of many bands that Eventide does (or improvements on the idea), then those companies might win unless legal issues arise.
Also, AI systems that can be turned on and dialed in from 0% to 100% achieving max transparency, to 200 or 400% so that the AI adjustments do too much and cause the opposite "shaping behaviors" of the user end interactions... this will be far superior to the human engineer... but we'll probably hit a music revolution at that point and then within a couple of years, we will be begging for re-introduction of human error and the AI systems will figure out ways of doing that (such as facial recognition to see if the engineer frowns or smiles, and then doing more or less of the behavior that either caused good or bad reactions).
I haven't compared DMG Limitless to Pro-L 2 yet, but I see arguments for it being better and being transparent while also able to handle a larger amount of GR. Limitless is nice and the only reason I haven't switched from 2 Ozone Maximizers to Limitless is because Ozone takes me about 10% of the time to dial in (because I know it well). I determined that DMG Limitless can be set with more GR (louder) and more transparent than 2 Ozone's... but I still haven't switched from using 2 Ozones to using Limitless because I already know my Ozone's well enough to do quick dialing in... where as with 30-40 hours of using Limitless... I find the need to go back to the forums and spend another 200 hours in order to get it so that it doesn't either take me 1-2 hours to find very good settings... or to not worry so much about improving its quality of resulting audio by 1 or 2% with tons of time being eaten up. This is the same reason why I didn't immediately switch to Eventide Elevate (which is the one that I feel is "the best" for my personal preferences and how my processing can achieve the most superior results for the widest range of clients and styles). CPU and hours of comparing tons of settings that all sound pretty damn good, and then it being so much coloring that the clients will have varying opinions every time... makes me think that I'll spend 20 hours dialing it in for my own music, but stick with what gets done in the least amount of time while yielding very acceptable and high quality results (Ozone). Limitless could end up replacing it, but I am testing others first.
I have a feeling that once I test Weiss... then I'll start saving up the rediculous $600 for it... and unless I get more clients for mastering work, that just isn't possible.
The pre-clipper stages of both Elevate and Limitless were not complete failures, but I prefer separate ones that are higher quality results. If we want clips reintroduced after the first or second limiter, we can do parallel processing with the clippers.
And, StandardCLIP & KClip 3 Pro got replaced by freebies! AirWindows OneCornerClip (broadband, preserves clean state) and ADClip7 (reintroduces clipped bass signal and smooths out highs.. this one better for the parallel re-introducton of clips at post limiter stages (I actually used 2 Ozone Maximizers and 1 Limitless, with clips coming back just before the final true peak handling & bass massaging slow Ozone 8 Maximizer IRC IV Modern. The CPU gets smashed, though. For more transparency, just OneCornerClip before Limitless... done after tweaking Limitless for an hour (after 30 hours of testing it).
My go-to compressor became AirWindows ButterComp2 but its been 2-3 years since I started avoiding compression in mastering chains at all costs (dyn EQ and smoothing harmonics, etc, handles things).
I feel that other limiters are right up there in the top10, and was actually shocked yesterday when I tried out AOM Invisible Limiter and Voxengo Elephant together
Big idea time:
My guess is that this type of system might work best on some material, but arguably it could be doing too much processing and cause not only CPU issues but also way too much math going on:::
2 Parallel chains in series, with:
Parallel chain A) [max transparency, probably does more GR than chain B]:
The best smooth & transparent compressors which are complimentary/oppositely shaped than Limitless... 1) broadband, 2) multiband, then 3) Limitless with clip off
and then, placed in series (or perhaps in the same single parallel chain), after parallel chain A), we'd have:
Parallel chain B) [more tone shaping but possibly can be as transparent]:
1) PSP E27 and/or 2) Hornet 31 (or similar) with 3) Elevate, perhaps with 4) Equivocate being used either also on parallel or placed in series before the E27 and/or Hornet 31.
To further over-process with many things that do less amounts of GR...
Optionally, add pre-clipping before the parallel chains and re-introduce some of the clips after the parallel chains... and also add pre clipping with post massaging back in of clips on each parallel chain, doing much less than the first one that is in series before the parallel stuff.
Even further, follow the whole mess with Weiss DS1-MK3... and to add max rediculousness, we could do further clipping in between the parallel stuff and the final Wiess... perhaps not "loudenator" clippers but more along the lines of Sonnox Oxford Inflator.
It probably would not work as well as just 1 combo that uses the ideas behind what Weiss does with Limitless or Elevate... or Weiss.
The parallel thing could eventually have AI systems in place so that, if its 100% on, when you dial in a change on one parameter then its going to change stuff (especially anything multiband) on other things so that it stays in this "complimentary functioning" with whatever you alter... but if we dialed down the AI amount to say 50%, then we'd still be able to add some desired coloration from the human error and judgment of taste. If we, say raised the pre-clipping stuff, then it'll adjust multiple things, like post-massaging back in of clips, as well as limiters attack/release based on the peak to RMS ratios and the true peaks to short-term and momentary LUFSratios, as well as the RMS to integrated LUFS ratios and the multiband shaping of the audio signal... so that things don't get totally whacked off by dialing up extreme clipping.
The AI stuff has a long way to go, though, to be even close to where I'd want it to be, in order to handle automation of multiple plugins inside of DAW, based on reading each one and adjusting them singularly while also adjusting them based on dependency of working together in ways that preserve maximum levels of transparency.
I would also guess that the argument true, when people speak up against such over-processing chains and how they'd be more damaging that they would be good. For example, chains with too much processing will end up causing so much math that too many BITs become lost and therefore the final resulting quality of audio becomes somewhat damaged. Also, we're not quite there with getting our tech built in ways so that MB compression can work with absolutely minimized amounts of the bouncing behaviors caused by the cross-over systems of splitting up things into multiple bands with slopes and a bunch of math to try to remove bleed through of signal. I think that we'll eventually get there, though... perhaps even having to MB compressors would work together in parallel or in series, to remove the bouncing thing (forget what thats called)... have them alter the cross-over points of the 2nd one, and slopes/effect wet amounts/etc, so that the bouncing from one becomes completely canceled out by the opposite bouncing from the other. We've got a long way to go, for such intelligence of these things.
Also, our computers and audio processing capabilities will eventually become much greater... so perhaps a system similar to the above idea, in which "doing more is most likely going to do less," in most case... might actually turn around with high quality bit depth/sample rate processing and larger CPU processing capabilities.
A note on the AI stuff... so far, its not taking over the world yet. Once we give it the ability to make decisions based on how it feels at the moment, or if we some how tell it to do A if B and C happen... but some D comes along and exploits a glitch in the math... then we're screwed when it suddenly decides to print 3d guns and robots in an effort to destroy the annoyance of mankind... or if it some how decides that preserving the Earth is more necessary than babysitting human necessities... then were screwed, very fast.