Yup, some "brickwall" limiters ain't, really. Really though, you have a couple of different problems. First, you're relying on a single processor to do too much heavy lifting; the finest compressors known to mankind are capable of about 4.5dB of increase without signal degradation, and you're going from -18dBFS to within two tenths of digital zero. The higher the ratio, the worse that's gonna sound, and limiters tend to have higher ratios than compressors.ComposerLDG wrote: ↑Sat Apr 06, 2019 4:47 amOk, so lately I've been focusing on upping my production skills, since there is so much well-produced stuff out there to compete with. I sent in a track to Taxi for a custom critique focusing on overall mix, eq, and production overall. The screener gave some very encouraging observations, and also offered some suggestions. One thing he/she said is that overall the mix seemed slightly hot. On listening closely to it, I agree. I don't hear distorting, but it definitely sounds like it's pushed pretty hard.
Typically when I track, I gain stage the individual tracks to -18 dbFS so there's lots of room. In the final mix, I'm using a limiter to raise the overall volume to -.2. I've had the Waves VU meter plugin around, but never thought of using it. I figured after those comments, I would give it a try by throwing it on my submix bus to see what it showed, since VU tends to "hear" much like the ear does. After calibrating it so that 0 was -18 dbFS, I started playing the mix to find that it's totally in the red at +3 and showing clipping at times, even though the limiter is definitely stopping it at -.2 ( have it set to true peak detection).
So right now my head's spinning because I've been listening to the track for a while. Before I do any more work on the track, does anyone have any input on what I've said above? My my calibration is off? Or is this just a classic case of pushing a limiter too hard?
Thanks all,
-Loren
So what are we doing here? The reason we look at crest factor is that the ear responds to averages more than to peaks, so if we raise the average relative to those peaks, the perception is that, 'it got louder'. basically we're trading dynamic range for 'loudness'; too much of that and the transient peaks get squished, distorted softened and you start to lose 'punch', so it's a balancing act.
When you want transparent compression, you need to apply it at multiple stages; I compress kick, snare and sometimes overheads, I may have as many as three compressors on a lead vocal, another on the drums bus, one on bass another strapped across L&R guitars, backing vocals, horns, etc. Another goes across the two-mix before hitting the mastering chain, which is a multiband compressor-limiter sandwiched between two linear-phase EQs and followed by a limiter.
But at the very end of that chain lies your second problem; you can send a mix that's peaking at -.2dBFS off to the CD pressing plant, no problem (if your meters are accurate, that is) but if you feed that to an MP3 encoder, there's gonna be trouble. Those things require a full decible of headroom, and in some cases, more. If not, the sonic gremlins eat your mix. Maybe you don't even need to start with so much headroom; if I'm doing a rock trio, everybody's going to overplay and be loud and little finesse will be required; it's different if you're shoehorning a couple hundred tracks into a country mix.
All I do now is check that I have a healthy headroom on the master bus and then use Ozone (also comparing my mix to reference tracks). I always make sure Ozone's limiting is done true-peak friendly and I have to say I am generally happy how my mixes translate to different sets of speakers and headphones... I am sure that I am oversimplifying it the way I do it but as longs as the "mix box" is not checked in my Taxi critiques I am happy 
(and thus far I never had any complaints). 