Post by: natpub on May 31, 2004, 03:28:31 AM
**********************************************
"...FWIW: A mastering engineer I usually master with has abandoned using his L2 almost entirely and finds that just turning it up and letting the inputs to the AD clip, sounds better 90% of the time..."
***********************************************
It seems to me that would be like taking scissors and cutting off the top of all the music and replacing it wish harsh digital distortion?
Post by: jfrigo on May 31, 2004, 03:50:38 AM
| natpub wrote on Mon, 31 May 2004 00:28 |
I read this quote on a thead at another site: ********************************************** "...FWIW: A mastering engineer I usually master with has abandoned using his L2 almost entirely and finds that just turning it up and letting the inputs to the AD clip, sounds better 90% of the time..." *********************************************** It seems to me that would be like taking scissors and cutting off the top of all the music and replacing it wish harsh digital distortion? |
I've had a couple clients who like to just push up the master fader into the red in their pro tools rig rather than use a limiter. They like digital clipping better than analog clipping, though a bit of that seems attractive to some as well, especially in the form of some of the simple "soft limit" or "soft clip" circuits in certain A/Ds. I've also had a couple that liked this pro tools clipping approach in Sonic. I know several NYC guys (and others I'm sure) have done this on certain sessions for years. Back when a plant would bounce back a pre-master for overs, they'd reduce the master fader in Sonic by .1 dB and clip the channel faders. That way you get the clipping but never produce an over and the master doesn't get rejected. These days clipped masters are pretty much standard in some circles and plants no longer complain. I think often they don't even read the PQ log anymore.
Post by: natpub on May 31, 2004, 04:21:33 AM
Thanks,
--KT
Post by: Ed Littman on May 31, 2004, 11:18:25 AM
Ed
Post by: Jason Poff on May 31, 2004, 12:26:39 PM
Jason
Post by: jfrigo on May 31, 2004, 02:39:40 PM
| natpub wrote on Mon, 31 May 2004 01:21 |
I think I see, but isn't that just going to get you a flat top across the whole thing, with a fakey digital sounding haze around it? It seems worse to me than hyper-compression. Or am I missing the boat here? I don't know mastering, so am curious. |
You are right that in general this is a "bad" thing, but if the client loves it and leaves happy, it can't be all bad. I certainly try to avoid clipping and the resultant distortion when possible, but if that's what the job demands, I'm not too cool to light a couple red lights. However, I would not use this approach as standard.
Most people use the term "hypercompression" to encompass too much compression, too much limiting, or clipping. It's the abuse of dynamic range reduction by whichever method, so this intentional clipping can be thought of as one manifestation of the "hypercompression" problem.
Post by: bblackwood on June 01, 2004, 11:19:43 AM
I also think you'll find that if you are gonna clip the signal, doing it after is digitzed will yield better results than hammering the ADC.
Post by: j.hall on June 01, 2004, 02:23:45 PM
| jfrigo wrote on Mon, 31 May 2004 02:50 |
I've had a couple clients who like to just push up the master fader into the red in their pro tools rig rather than use a limiter. |
not sure about HD, but mix plus systems make your statement redundant
the mix buss essentially was a brick wall limiter. you really couldn't get any digital clipping at the 2 buss
it just got more and more squashed as you hit red.
Post by: jfrigo on June 01, 2004, 05:43:23 PM
| j.hall wrote on Tue, 01 June 2004 11:23 | ||
not sure about HD, but mix plus systems make your statement redundant the mix buss essentially was a brick wall limiter. you really couldn't get any digital clipping at the 2 buss it just got more and more squashed as you hit red. |
You don't think what a master fader in PT mix did when it hit red was clipping? It sure looks and sounds like clipping to me. Are you saying that a Mix system wouldn't put out full scale? I thought it would, but even if it didn't, it sure the heck can still clip. It's a brick wall clipper like any other digital gain device turned up beyond it's headroom, not a brick wall limiter, as far as I can tell.
Post by: Skwaidu on June 03, 2004, 12:57:37 PM
He very much prefers it for at least beat-heavy stuff.
He did me a demo with a L2 and some smashed converters on a club-style track and I have to agree that he has a point...
Post by: Bob Schwenkler on June 04, 2004, 07:21:28 PM
Post by: bblackwood on June 04, 2004, 08:04:39 PM
| chikkenguy wrote on Fri, 04 June 2004 18:21 |
what does clipping converters do that a limiter wont do more gracefully? |
Clipping allows the transient of the waveform to remain intact until 0dBfs - a limiter will start to roll that off as it approaches full scale. This can be felt as a reduction in 'impact' especially, but the diff is usually only apparent at larger levels of GR...
Post by: Level on June 04, 2004, 09:13:28 PM
If you push L-1+ or L-2 up, the mids come up to overbearing and the highs get cloudy. The bottom gives way to "blasting"
To solve this "problem", I have one stand alone CD recorder where I can hit its analog inputs pretty hard and the circuity compresses without the large change in powerband response. It has a rather soft limiter and easy knee compression "sound" that will give him the goods.
That damned thing will actually get a +18VU if I need it.
I feel it is high time for the standards to be limited to RMS of -17..but it is all music dependant. His R&B album will end up around -12RMS ...although it WILL sound better at -17RMS.
R. Kelly sells, he wants to do the same.
I do not like the trend...but clients wishes prevail...unfortunantly.
Post by: Ronny on June 05, 2004, 02:45:03 AM
| bblackwood wrote on Fri, 04 June 2004 20:04 | ||
Clipping allows the transient of the waveform to remain intact until 0dBfs - a limiter will start to roll that off as it approaches full scale. This can be felt as a reduction in 'impact' especially, but the diff is usually only apparent at larger levels of GR... |
Two kinds of clipping here at issue and I wouldn't want chikkenguy to get the wrong idea. The clipping that Brad is talking about, "I think", is flat topping the waveform before it exceeds -0dBFS. There is no distortion just reduced peaks. Digital clipping of the converter is when the signal exceeds -0dBFS. To my knowledge a digital limiter will not roll off when it's printed, it will limit exactly where the threshold is set, if it's a hard knee it's instantaneous brickwall, as soon as the first sample exceeds the threshold. To flat top the waveform at the ADC I always thought took limiting the mic pre, how else do you clip an ADC without exceeding -0dBFS? If you exceed -0dBFS you run the risk of the worst clipping distortion of all time, digital distortion when the signal exceeds -0 and the system no longer interprets it as music. At 24 bit risking overs at the ADC is a needless worry, IMHO, just bring the gain down and make it up later where you can denoise any noise floor lost, if you feel the need. One poster mentioned digitally clipping quick transients. While it's true that fast transients that exceed -0dBFS can be inaudible to some ears, if the signal lasts only a few milliseconds over, the audio is gone just the same. If you exceed the limit too long you get digital burps that sound terrible and can ruin a take, but clipping quick transients although inaudible as individual transients because they are short, form a blanket over the sound that is quite noticeable to my ears, "if" too many transients are clipped in a row over short timeline. It's a fuzzy distortion that renders the audio degraded in my opinion. I'm sure some people perceive it as squash distortion, but it's quite different. There are better ways these days than clipping the waveform to a distortion level, staged limiting with automated release control, will get you just as hot and without the blanket on the sound that clipping quick transients results in. IN MY HUMBLE OPINION!
Post by: bblackwood on June 05, 2004, 08:00:29 AM
Post by: moogus on June 05, 2004, 12:28:43 PM
I usually mix to 1/4" tape, and most of the projects I do arent taken (at least straight away) to a mastering engineer, so I do the transfer back to digital, from tape.
My question is slightly less relevant now record back at 24/96 but Id still like to know what yall think.
How should I optimise the level coming back into the computer from tape? If I set the interface's intputs to -10, I get level that is roughly the same as a normally loud cd. But of course its clipping the A-D all the time.
If I set the inputs to +4 I get no clipping but then have to apply 10db or more of gain (usually with L2) to get to normal cd levels again.
The usual 'trust your ears' advice aside, what do you think is the best way to go?
M
Post by: Ed Littman on June 05, 2004, 02:43:29 PM
If i were you, I'd get the mix exactly the way you want it processing & all before mixing to 1/4".
Then transfer the mix tape at close to your final level at 24/44.1 through your adc(or clip the adc if your looking for loud). Once your tape mix is a 24/44.1 file you can get away with furthure processing if needed or just dither down to 16bits & your done.
Ed