lyons
Full Member
Posts: 28
|
Post by lyons on Sept 1, 2024 17:59:23 GMT -6
Hi all,
I know this a question that is posted all over the internet (so I do apologise) but lately the information I’ve been reading online has just wasted a lot of time as whether it’s right or wrong information all I can say is that it hasn’t seemed applicable. (To be fair the more I progress the more I feel like the majority of what is spouted online doesn’t make much sense).
I’m going to try and explain my thought process in my own words: I’m working on an acoustic style vocal that I don’t really want to mess too much with the dynamics of as volume in certain phrases is used for artistic expression. The dynamic range is intended. But there are peaks (very short spikes sometimes) which most of the time seem to do nothing but hurt the listeners ears. I’ve always read how I shouldn’t really use a limiter in this case and to that end I’ve been trying different compressor plugins but it invariably is always doing more than I want or not enough. And even the compression is gentle it doesn’t catch very harsh peaks. I tried a limiter (fabfilter) to just take down the peaks around a max of 2-3db and then followed it with gentle compression just to bring up some detail. To my surprise it’s the best result I’ve had all week.
What are the arguments against using a limiter in this scenario? Am I going about this wrong way? Im not looking to colour the vocal at this stage especially.
Thanks
|
|
|
|
Post by skav on Sept 1, 2024 18:19:38 GMT -6
Some say if it sounds good, it is good..  Have you tried lowering those huge peaks with volume automation, before applying compression? |
|
|
|
Post by drumsound on Sept 1, 2024 18:38:02 GMT -6
What Skav said
|
|
|
|
Post by tonycamphd on Sept 1, 2024 19:58:12 GMT -6
I think you might be going about it wrong, I agree with above comments for what you articulated, clip gain is almost fool proof, a plug like vocal rider by waves makes it easy to volume dip those peaks and could cure what's bothering you, if it's more frequency specific/harsh you could dial in a d-esser to attenuate those freq's
The diff between compression and limiting is
traditional compression lowers the loudest peaks regardless of frequency by the ratio you set, 1:1, 2:1, 3, 4, 5, 6:1 etc, that means every db you go above the threshold you set, you compress(lower) db's corresponding to the first number, so 5:1 would mean 1db above threshold would lower volume 5db. Clearly the higher the 1st number the more obvious the compression regardless of knee, and applying makeup gain also raises the noise floor as well as the loudest remaining peaks
Limiting is basically a hard clip above the set threshold and is usually much more aggressive sounding
I hope that makes sense and helps you out
|
|
|
|
Post by Johnkenn on Sept 1, 2024 20:18:12 GMT -6
Yeah, if I don’t want to make something louder and just want to stop the peaks, I use a limiter.
|
|
lyons
Full Member
Posts: 28
|
Post by lyons on Sept 1, 2024 20:40:11 GMT -6
Thanks all for the replies. Yeah, if I don’t want to make something louder and just want to stop the peaks, I use a limiter. I’m definitely not looking to make it louder. Just bring down the spikes that are not serving any purpose. It’s an hour long live set - I’ve used clip gain on quite a few but they’re still there. |
|
|
|
Post by tonycamphd on Sept 1, 2024 21:22:32 GMT -6
Thanks all for the replies. Yeah, if I don’t want to make something louder and just want to stop the peaks, I use a limiter. I’m definitely not looking to make it louder. Just bring down the spikes that are not serving any purpose. It’s an hour long live set - I’ve used clip gain on quite a few but they’re still there. maybe try a d-esser and dial into the offending frequency |
|
|
|
Post by maldenfilms on Sept 1, 2024 21:54:23 GMT -6
I would say that this is basically exactly what I use an 1176 for. I usually track vocals with an Audioscape 76D set to 4:1, with attack at noon and release all the way fast. It's usually just triggering on the types of peaks you're talking about.
When I first started out, I was so afraid of "over-compressing" since most of what I read on the internet warned me not to compress more than say 3db on the way in. After spending some time in major studios and watching experienced engineers boldly compress much more than that on the way in, I decided to get a little braver.
While I totally get that an 1176 is technically adding "color", I'll be honest... the actual compression/limiting is pretty transparent. In other words, it sounds to me like that's just the way the singer sang it in the first place.
|
|
|
|
Post by bgrotto on Sept 1, 2024 22:43:17 GMT -6
This is a job for a limiter.
|
|
|
|
Post by EmRR on Sept 1, 2024 23:28:49 GMT -6
A layered approach, very low ratio compression to shrink overall range slightly before a limiter. Will make the limiters job easier if dialed in well.
|
|
|
|
Post by copperx on Sept 1, 2024 23:58:42 GMT -6
Unfortunately, the methods mentioned here are highly destructive. Limiters will distort the audio by abruptly cutting off peaks. Additionally, they might unintentionally affect other parts of the waveform. Ideally, we need a tool that reduces peak levels without altering their shape, as if we were just bringing them down in volume.
If you're a Reaper user, the built-in Zero Crossing Maximizer JS plugin offers a free solution. It lowers peaks without changing their shape. However, it also increases overall volume to compensate. You can mitigate this by reducing the 'ceiling' setting. Alternatively, to maintain the original volume, edit the plugin's code and change the line volume = ceiling / thresh; to volume = 1;.
For non-Reaper users, UrsaDSP Boost provides similar functionality. Remember to adjust the output level downwards after processing. It's important to note that the attack/lookahead and release settings in UrsaDSP Boost don't function like a compressor or limiter; it's simply a zero crossing maximizer.
Done.
|
|
|
|
Post by tonycamphd on Sept 2, 2024 0:23:22 GMT -6
Unfortunately, the methods mentioned here are highly destructive. Limiters will distort the audio by abruptly cutting off peaks. Additionally, they might unintentionally affect other parts of the waveform. Ideally, we need a tool that reduces peak levels without altering their shape, as if we were just bringing them down in volume. If you're a Reaper user, the built-in Zero Crossing Maximizer JS plugin offers a free solution. It lowers peaks without changing their shape. However, it also increases overall volume to compensate. You can mitigate this by reducing the 'ceiling' setting. Alternatively, to maintain the original volume, edit the plugin's code and change the line volume = ceiling / thresh; to volume = 1;. For non-Reaper users, UrsaDSP Boost provides similar functionality. Remember to adjust the output level downwards after processing. It's important to note that the attack/lookahead and release settings in UrsaDSP Boost don't function like a compressor or limiter; it's simply a zero crossing maximizer. Done. Please explain how clip gain, d-esser or a vocal rider plugins are “destructive”? None of which are compressors or limiters |
|
|
|
Post by niklas1073 on Sept 2, 2024 9:17:19 GMT -6
I would agree with the very first posts regarding automation. Riding (prefer manually) the vocal is essential. It will not only even out the fluctuation of level but it will be needed to sit the vocal track in the song and it’s dynamics. Same applies for guitars etc. after that you bring in the sound and even it out further with a compressor or two or three in stages. That’s another piece of science. Never even crossed my mind to limit a vocal despite i know many do. Rather fix the issues at their root while you can 😉
|
|
|
|
Post by Johnkenn on Sept 2, 2024 10:07:03 GMT -6
Thanks all for the replies. I’m definitely not looking to make it louder. Just bring down the spikes that are not serving any purpose. It’s an hour long live set - I’ve used clip gain on quite a few but they’re still there. maybe try a d-esser and dial into the offending frequency Dynamic eq. The Pulsar has a cool sidechain eq graph that allows you to dial in where you want more or less compression. |
|
|
|
Post by robo on Sept 2, 2024 10:29:32 GMT -6
Unfortunately, the methods mentioned here are highly destructive. Limiters will distort the audio by abruptly cutting off peaks. Additionally, they might unintentionally affect other parts of the waveform. Ideally, we need a tool that reduces peak levels without altering their shape, as if we were just bringing them down in volume. If you're a Reaper user, the built-in Zero Crossing Maximizer JS plugin offers a free solution. It lowers peaks without changing their shape. However, it also increases overall volume to compensate. You can mitigate this by reducing the 'ceiling' setting. Alternatively, to maintain the original volume, edit the plugin's code and change the line volume = ceiling / thresh; to volume = 1;. For non-Reaper users, UrsaDSP Boost provides similar functionality. Remember to adjust the output level downwards after processing. It's important to note that the attack/lookahead and release settings in UrsaDSP Boost don't function like a compressor or limiter; it's simply a zero crossing maximizer. Done. If your goal is a clean-as-possible recording or to compensate for bad mic technique, then yes. If your goal is to create the illusion of a dynamic musical performance without big volume jumps, then the artifacts of compression/limiting/saturation can add intensity and excitement to compensate for the lack of dynamic range. Not trying to be pedantic, just saying us peak-squashers aren’t cavemen. |
|
|
|
Post by Dan on Sept 2, 2024 11:27:06 GMT -6
Unfortunately, the methods mentioned here are highly destructive. Limiters will distort the audio by abruptly cutting off peaks. Additionally, they might unintentionally affect other parts of the waveform. Ideally, we need a tool that reduces peak levels without altering their shape, as if we were just bringing them down in volume. If you're a Reaper user, the built-in Zero Crossing Maximizer JS plugin offers a free solution. It lowers peaks without changing their shape. However, it also increases overall volume to compensate. You can mitigate this by reducing the 'ceiling' setting. Alternatively, to maintain the original volume, edit the plugin's code and change the line volume = ceiling / thresh; to volume = 1;. For non-Reaper users, UrsaDSP Boost provides similar functionality. Remember to adjust the output level downwards after processing. It's important to note that the attack/lookahead and release settings in UrsaDSP Boost don't function like a compressor or limiter; it's simply a zero crossing maximizer. Done. That’s still destructive. It lowers the gain before a peak, distorting more of the audio in time, pumping the audio randomly based on the future, to not distort the brief peak. This is what happens with a too long lookahead and it sounds funny (Powair) to awful (set a few ms long long ahead in Limitless or Pro L2). |
|
|
|
Post by copperx on Sept 2, 2024 12:35:06 GMT -6
Unfortunately, the methods mentioned here are highly destructive. Limiters will distort the audio by abruptly cutting off peaks. Additionally, they might unintentionally affect other parts of the waveform. Ideally, we need a tool that reduces peak levels without altering their shape, as if we were just bringing them down in volume. If you're a Reaper user, the built-in Zero Crossing Maximizer JS plugin offers a free solution. It lowers peaks without changing their shape. However, it also increases overall volume to compensate. You can mitigate this by reducing the 'ceiling' setting. Alternatively, to maintain the original volume, edit the plugin's code and change the line volume = ceiling / thresh; to volume = 1;. For non-Reaper users, UrsaDSP Boost provides similar functionality. Remember to adjust the output level downwards after processing. It's important to note that the attack/lookahead and release settings in UrsaDSP Boost don't function like a compressor or limiter; it's simply a zero crossing maximizer. Done. That’s still destructive. It lowers the gain before a peak, distorting more of the audio in time, pumping the audio randomly based on the future, to not distort the brief peak. This is what happens with a too long lookahead and it sounds funny (Powair) to awful (set a few ms long long ahead in Limitless or Pro L2).
I don't think that's how it works. It really knows the future, which is why both plugins have a significant lookahead. Whenever it detects a peak exceeding the threshold, it lowers it as if it were clip gain. From one zero crossing to the next, it multiplies every sample by a fixed constant to lower the volume, so it does not introduce nonlinearities because there is no envelope. However, it might create slight discontinuities near the zero crossings, but in practice, these are inaudible.
For all of this to work, the affected peaks must fit within the lookahead buffer, that's why it needs to be so huge.
|
|
|
|
Post by Dan on Sept 2, 2024 12:45:47 GMT -6
Unfortunately, the methods mentioned here are highly destructive. Limiters will distort the audio by abruptly cutting off peaks. Additionally, they might unintentionally affect other parts of the waveform. Ideally, we need a tool that reduces peak levels without altering their shape, as if we were just bringing them down in volume. If you're a Reaper user, the built-in Zero Crossing Maximizer JS plugin offers a free solution. It lowers peaks without changing their shape. However, it also increases overall volume to compensate. You can mitigate this by reducing the 'ceiling' setting. Alternatively, to maintain the original volume, edit the plugin's code and change the line volume = ceiling / thresh; to volume = 1;. For non-Reaper users, UrsaDSP Boost provides similar functionality. Remember to adjust the output level downwards after processing. It's important to note that the attack/lookahead and release settings in UrsaDSP Boost don't function like a compressor or limiter; it's simply a zero crossing maximizer. Done. If your goal is a clean-as-possible recording or to compensate for bad mic technique, then yes. If your goal is to create the illusion of a dynamic musical performance without big volume jumps, then the artifacts of compression/limiting/saturation can add intensity and excitement to compensate for the lack of dynamic range. Not trying to be pedantic, just saying us peak-squashers aren’t cavemen. Brief peaks lowered and replaced with distortion is less audible than the pumping. Of course flatlining something into a blob with an a massive amount off on an 1176 won’t make it an even bob, just a pumpy and distorted blob with no peaks left because it’s not logarithmic |
|
|
|
Post by tonycamphd on Sept 2, 2024 13:19:28 GMT -6
Please explain how clip gain, d-esser or a vocal rider plugins are “destructive”? None of which are compressors or limiters They’re all volume changes of some sorts. They all produce distortion except for clip gain and the ramping of that still produces distortion. Smoother non-linear gain. Even volume automation and fades produce distortion. Good compressors are faster than your fingers and have far better curves than wobbly fingers can draw with a mouse or a fader. I accept your assertion, but it's new to me that a simple level change distorts as I can't hear any beyond the fletcher munson effect, as far as i understand fader changes bring the ceiling and floor of a track up or down simultaneously irrespective to dynamic range and hardware specs, compression and limiting bring the ceiling down and leave the floor as is constituting dynamic range change and CERTAIN distortion that i can hear. IME i've heard a so so d-esser twist things as they can qualify as band specific compression, but certainly not clip gain or a fader ride unless done very poorly. I have 9 aphex 651's with JW mods here, they are the most transparent compressors on the planet IME, but I can clearly hear that they don't sound as transparent as good fader riding or clip gain adjustments on anything i've ever done. So I'm positive your tech savvy is above my pay grade and you are probably correct, but IME for practical real world purposes there is a big diff between comp/limit distortion and fader/clip gain moves Considering the OP has been unsuccessful with everything he's tried so far, my recommendation remains that he should painstakingly scan the waveform for the offending spikes(if you can hear it you can see it) and carefully clip gain them into submission one at a time, I believe that will be the most transparent fix he will find. D-essing, comps, limiters, dynamic eq etc are all the quick way out if they work(and i love them when they do work). As for me compression or limiting isn't even a consideration for "FIX or REPAIRS", going through a kick drum track that was performed by a 1/2 rse drummer, beat by beat and clip gaining him into competency is all too common for me, doing the same for vox is a household chore that is done 100% of the time. My pal Mike told me he used to pre amp fader ride Whitney Houston's vox while tracking her as per normal and still needed to level her in post, if she needed it everyone does to some degree |
|
|
|
Post by copperx on Sept 2, 2024 13:35:27 GMT -6
Considering the OP has been unsuccessful with everything he's tried so far, my recommendation remains that he should painstakingly scan the waveform for the offending spikes(if you can hear it you can see it) and carefully clip gain them into submission one at a time, I believe that will be the most transparent fix he will find. Yes, but there's no need to do it manually. The solutions I described do the same thing. |
|
|
|
Post by Dan on Sept 2, 2024 16:35:46 GMT -6
Just limit the peaks before or after compression
Or Getting something that can level it and limit it at the same time will reduce IMD. This is how the multiple detector and “auto” compressors work. They are by far the most utilitarian compressors ever made or programmed. Massenburg, Crane Song, and Tokyo Dawn compressors are fabulous. What you can get for 10 bucks on sale now is only rivaled by things that cost thousands.
I’ve used the Oxford Dynamics to simulate them for years. Of course Weiss and Tokyo Dawn can be much more aggressive.
|
|
|
|
Post by Ward on Sept 3, 2024 10:12:53 GMT -6
Hi all, I know this a question that is posted all over the internet (so I do apologise) but lately the information I’ve been reading online has just wasted a lot of time as whether it’s right or wrong information all I can say is that it hasn’t seemed applicable. (To be fair the more I progress the more I feel like the majority of what is spouted online doesn’t make much sense). I’m going to try and explain my thought process in my own words: I’m working on an acoustic style vocal that I don’t really want to mess too much with the dynamics of as volume in certain phrases is used for artistic expression. The dynamic range is intended. But there are peaks (very short spikes sometimes) which most of the time seem to do nothing but hurt the listeners ears. I’ve always read how I shouldn’t really use a limiter in this case and to that end I’ve been trying different compressor plugins but it invariably is always doing more than I want or not enough. And even the compression is gentle it doesn’t catch very harsh peaks. I tried a limiter (fabfilter) to just take down the peaks around a max of 2-3db and then followed it with gentle compression just to bring up some detail. To my surprise it’s the best result I’ve had all week. What are the arguments against using a limiter in this scenario? Am I going about this wrong way? Im not looking to colour the vocal at this stage especially. Thanks Here is the most basic operating procedure: A compressor rides the waves and delivers a more consistent result A limiter catches all the peaks and crushes the program material. Get yourself a good optical like an LA2 or LA3 clone, and a good 1176 clone - Rev D or F. Experiment. Keep both and spend time with them over a period of years and many projects. There is magic in both, just take your time and listen critically. |
|
karlo
Full Member
Posts: 42
|
Post by karlo on Sept 3, 2024 10:24:57 GMT -6
I use a clipper to cut those spikes down, then I limit, but barely... after clipping I increase the gain into the limiter to just periodically kiss it.
Thank you for taking the time to listen to this advice from a stranger on the internet.
|
|
|
|
Post by tonycamphd on Sept 3, 2024 10:47:59 GMT -6
just to clarify, "clip gain" has nothing to do with peak clipping
|
|
|
|
Post by Ward on Sept 3, 2024 10:50:44 GMT -6
just to clarify, "clip gain" has nothing to do with peak clipping But an incredibly useful tool/technique! |
|
