Confused with Ouput cliping led behaviour ! Need help please.

Red generally means clipping, yes (or can at least be thought of that way)

I'm wondering if you have any fx active? Also, try pushing and holding the cab button, and see if the "volume" parameter in the offending rig(s) is set to above zero?

Distorting rigs rarely go into red for me, although on some higher gain rigs they can during palm muting.

When you say you lose a lot of volume on recordings - when you play back what you have recorded with the DAW fader at unity (default position), what does the meter of that channel say?

8 dB of headroom is fine. I generally have low-end heavy instruments (kick drums, bass etc.) peaking at around -12 dB, due to them eating up a lot of mixing real estate (relatively more energy in longer waveforms). Guitars peaking at -8 dB gives you plenty of signal to work with in the 24-bit domain. Digital workflow means you don't have to be so worried about signal-to-noise ratios as it was in the old analogue multitrack tape days. Pushing signals up towards 0 dBFS doesnt give you anything other than the risk of digital clipping, which sounds horrible.

Quote from sambrox

8 dB of headroom is fine. I generally have low-end heavy instruments (kick drums, bass etc.) peaking at around -12 dB, due to them eating up a lot of mixing real estate (relatively more energy in longer waveforms). Guitars peaking at -8 dB gives you plenty of signal to work with in the 24-bit domain. Digital workflow means you don't have to be so worried about signal-to-noise ratios as it was in the old analogue multitrack tape days. Pushing signals up towards 0 dBFS doesnt give you anything other than the risk of digital clipping, which sounds horrible.

Sam's the man! I usually go quite a bit lower even, like -18 RMS/-10 peak - somewhere around there.

Me too on everything else, actually. I like to be able to have the fader around unity gain and not have to faff about with trim/gain plugins, @Michael_dk

I don't see any issue if you don't hear any actual audio clipping , I often go to the red zone on a particularity aggressive attack.

So you'd better thrust your ears than you eyes

If you're connected through s/pdif you have a much more precise meter on your interface (or in the DAW) than that output led will provide! If the s/pdif signal doesn't clip, then the Kemper doesn't clip.

That said, as have been mentioned, there's no need to push it to the top

Quote from Trazan

If you're connected through s/pdif you have a much more precise meter on your interface (or in the DAW) than that output led will provide! If the s/pdif signal doesn't clip, then the Kemper doesn't clip.

That said, as have been mentioned, there's no need to push it to the top

Not sure if this is completely true.

Quote from Michael_dk

Not sure if this is completely true.

No? Explain

I seem to remember something about a built-in limiter.

However, thinking about it, it may be that I'm thinking about the power rack/head, and/or input rather than output.

Quote from Michael_dk

I seem to remember something about a built-in limiter.

However, thinking about it, it may be that I'm thinking about the power rack/head, and/or input rather than output.

There might be one on the analog outputs, but not on the s/pdif. Anyway, it would be a good idea to stay away from going into an output limiter. If there is one.

Metering the s/pdif output gives you an exact level/headroom indicator within the Kemper, and what's going to the analog outputs (assuming they're not attenuated). I wish there was such a meter in the Kemper

Sam and Michael are 100% spot-on IMHO.

The thing is that those of us who've been grounded in old-school techniques and equipment have been conditioned to believe that the hotter a signal is without clipping, the better. This mentality has prevailed in spite of the fact that these days 24-bit resolution and much-lower noise floors have rendered this approach obsolete.

There's more to it than meets the eye, too. Take inter-sample clipping (or should that be "intra"?), for instance. Consumer D/A converters such as those on TVs, computers, CD and DVD players are of such poor quality that this phenomenon is more-or-less guaranteed to manifest. The fact that mixes are pushed to very near the 0dB ceiling means that these devices spit out all manner of overshoots of the type described. This is perceived as an overall harshness to the sound. A sort of "spitty" quality, if you will.

In our DAWs, summing multiple tracks which have been recorded close to the ceiling, IMHO, makes us vulnerable to this problem as well as more likely to experience somewhat "choked", "constricted", bandwidth-limited (mainly a loss of bottom end), "small", less "open" results when mixing. The stereo field is narrowed and front-to-back depth lessens significantly. Try mixing a project where all tracks peak around -3dB or less and compare the result with a similar one that was recorded at -16 to -18dB RMS, -8 to -12 dB peak and I promise you the difference will be like night and day. You have to hear it to believe it.

I first became aware of this 10 years ago when I switched from O2R (outboard digital mixing) to ITB (in the box). Even 'though the O2R only converted at 18 to 20 bit for its various I/O, summed channels sounded natural and how you'd expect them to. The minute I made the switch, I thought I'd made the biggest mistake of my life as my worst fears seemed to have come true - I perceived a huge step down in mix quality. The words I used earlier to describe the hot-channel-summing effect were lifted directly from what I thought at that time. That initial impression will remain with me forever and has become my "reference" as to what to look out for and avoid.

Now, I'm aware of the ridiculous theoretical dynamic range at our fingertips - my DAW, Digital Performer, boasts a figure of around 1500dB. I'm also aware that the other old-school stalwart, that of maintaining correct gain structure throughout our signal paths, is now obsolete in our virtual-mixing environments due to said theoretical dynamic range. In light of this, one might well ask why the recorded levels of tracks, if they were indeed very hot, would matter at all. All I know is that, to my ears, the "sweet spot" isn't in the top 6dB; it's in the 12-dB range below that. This is why I agree with Michael and Sam's propositions. IMHO, peaking at -10 isn't a bad place to start. RMS values aren't much help to many of us now that most rely on peak readouts only, but if the signal hovers around -12 to -18 and peaks at -8 to -12, you can't go too far wrong IMHO. As a bonus, those inter-sample peaks won't be able to breach the ceiling as they'll not have enough energy to do so from this much-lower base.

I wish I could remember an explanation I read years ago as to why, as a digital signal nears the 0dB ceiling, it loses definition, becoming more "grainy". It's counter-intuitive, I know; we know that low-level signals fall into ever-larger stair-stepped regions, and therefore expect the opposite to be true at the other end of the scale. It is, but only to a point, and that, my dear, fellow Kemperites, is the point. Accentuating this shortfall by summing multiple tracks which were recorded thusly, IMHO, explains the constricted, tiny, lacking-in-openness mixes that result and which we'd all like to be able to avoid. Well, hopefully, assuming I'm right, those who've bothered to read this rant will find themselves a huge step closer to that Holy Grail.

Quote from Monkey_Man

Sam and Michael are 100% spot-on IMHO.

Thanks!

Quote

Now, I'm aware of the ridiculous theoretical dynamic range at our fingertips - my DAW, Digital Performer, boasts a figure of around 1500dB. I'm also aware that the other old-school stalwart, that of maintaining correct gain structure throughout our signal paths, is now obsolete in our virtual-mixing environments due to said theoretical dynamic range. In light of this, one might well ask why the recorded levels of tracks, if they were indeed very hot, would matter at all. All I know is that, to my ears, the "sweet spot" isn't in the top 6dB; it's in the 12-dB range below that. This is why I agree with Michael and Sam's propositions. IMHO, peaking at -10 isn't a bad place to start. RMS values aren't much help to many of us now that most rely on peak readouts only, but if the signal hovers around -12 to -18 and peaks at -8 to -12, you can't go too far wrong IMHO. As a bonus, those inter-sample peaks won't be able to breach the ceiling as they'll not have enough energy to do so from this much-lower base.

That's not true, actually (it being obsolete). Many plugins - PARTICULARLY (but by no means only) the ones that are "recreations of old analogue gear goodness" that seem so popular these days - are in fact pretty sensitive to gain staging. If all your tracks are peaking at -3 dB and you're slapping analogue emulations on all the tracks than you're really piling on the saturation (and potentially some kinds of unwanted artefacts, depending on how the plugins are coded, for non-emulation stuff too).

What used to be zero (unity gain) in the olden days of analogue gear is actually more around -18 dBfs in the digital realm, if I recall correctly (or something around that - let's assume that's it). Consequently, peaking at -3 dBfs in your DAW would equate peaking at +15 dB in the analogue realm. That's... um... significant, let's say.

That's why I use clip gain/trim plugs/whatever to bring the tracks down to around 18 RMS (or 10 peak for more transient material like drums - whatever limit is reached first) before starting to mix (not that I do much of mixing these days). I go purely by eyes at this stage, with the master muted. I want to start off the right way, not wanting to fix shit as the first to-do in the mixing process. It's more PREPARATION than mixing.

This still gives you much better headroom than the ole analogue circuitry (between noise floor and unpleasant clipping).

BTW, there are many cheap/free RMS meters out there. Some are built into plugins that many people already have. So you don't need to rely on peak meters (though they are fine for the purpose, I guess - whether you're hitting -15, -18 or -21 dB RMS shouldn't matter much).

Quote

I wish I could remember an explanation I read years ago as to why, as a digital signal nears the 0dB ceiling, it loses definition, becoming more "grainy". It's counter-intuitive, I know; we know that low-level signals fall into ever-larger stair-stepped regions, and therefore expect the opposite to be true at the other end of the scale. It is, but only to a point, and that, my dear, fellow Kemperites, is the point. Accentuating this shortfall by summing multiple tracks which were recorded thusly, IMHO, explains the constricted, tiny, lacking-in-openness mixes that result and which we'd all like to be able to avoid. Well, hopefully, assuming I'm right, those who've bothered to read this rant will find themselves a huge step closer to that Holy Grail.

I'm not sure this is true, my monkey man. I'd love to read what you refer to here. I'm wondering if it's outdated information. It's entirely possible that I'm wrong, but I'd like to know the theory behind this regardless.
It would, however, fit with what I said earlier about plugins and saturation/clipping artefacts.

Thanks (I think... it's 8AM and way past bedtime so I should've known better than to check the thread!) for responding, Michael.

Gain-staging comments
My gain-staging comments were specifically based upon the practise of maximising S/N ratios back in the day using analogue-only equipment, and I obviously didn't have the nuances of how one might hit compressors, limiters and so on in mind. Those processors, after all, colour and shape tone and dynamic range along with amplitude envelopes, based in part on how hard they're hit. I was obviously more concerned with what happens to signals at the various stages before they get there (to said plugs, which are after all near the end of the signal's journey).

Besides all this, the context of that statement was to set the stage for the sentence which followed, and I think that one places the intent of it beyond doubt. In fact, the 1500dB comment, along with the sentence you highlighted, were intended only to illustrate that I'm aware of the arguments against what I was about to propose. I said, "In light of this, one might well ask why the recorded levels of tracks, if they were indeed very hot, would matter at all."

RMS-meter statement
Yes, I'm aware of the fact that there are many RMS options out there, Michael, but it seems few "amateurs" bother to use 'em. I aimed that "rant" at amateurs; I had typical home-recording guitarists seeking "clearer, bigger mixes" in mind. Pros use phase-corellation meters, RMS, VU or Virtual-VU and so on, I know, but I wouldn't dare attempt to advise the pros; I can only learn from them.

Grainy nearing the ceiling?
I remember who it was who wrote this now, Michael. Well, sort of. It was either Michael Stavrou or Roey Izhaki. I wouldn't question either's experience, or in the case of Michael, sheer genius. I will attempt to find the explanation later after I wake up and... have my day (LOL), so I'll try to get back to you in 12 hours or so.

I may well have got the detail wrong (grainy?), but the gist of it, the concept, if you like, dovetailed beautifully with my passion for recording at levels way below what most (not you!) would consider reasonable. This is due to the fact I pointed out earlier that summing hot-recorded channels accentuates said inaccuracies.

Im off. Will research this later; it may have been in Roey's "Mixing Audio"... or Michael's "Mixing with Your Mind" book, both of which I have. Frankly, I'm dreading having to do his, but hey, I did put it out there and one has to sleep in the bed one makes... or the couch, in my case.

Quote from J-War

Thanks all for your help. I am not sure to understand correctly what you're all saying tho ! so please correct me if i'm wrong :

- My daw being full 24bit processing, It's better to record my kemper at -8db than to try to be as close as possible to 0db ?
- In this case, how to compensate for the lost volume on the track where the kemper was recorded ? By raising the track gain or maybe track volume is better ?
- Is it the same thing for let's say... Virtual instruments, like easy drummer ? We should lower the volume on those tracks too ? I don't clearly get the point.

J.

Point 1 - Yes, -8dB is plenty hot enough, as you'll see if you read the post I'm about to quote.

Point 2 - No need to compensate; there is no lost level as you're already running hot. As you'll see in the article, our digital "0" is the point of meltdown. We all know that, but what many don't realise is that in the analogue world, such as in the case of the SSL desk used as an example in the post I'll quote, this point of meltdown is at +24dB. That's where you're frying components and everything's gone to crap. So, in analogue land this particular desk, and this is typical, is set up so that its 0VU (Volume Units) reference (like the one on our DAW faders that's marked "0" and usually shows markings to about 6dB above it by default and perhaps 40 or 50 dB below it), is referenced at +4 or 1.23 volts. "Referenced" in this case means it's calibrated to read "0" when 1.23v is pumping through it. The +4 bit means that there's 20dB to play with before things go to poo (24-20=4). The reasons this matters are many, but I think it's important to get the gist of what's going on here. The +4, 1.23 volt reference essentially means that anything you plug in, provided it's pro gear, will be happy to play nice with it. Surely you've heard of the fact that consumer gear (hi-fi equipment such as tape decks, CD players and so on) runs at -10dB? So, in the real (pro) world, you want your stuff (in this case, I/O signals) to run at the "pro level" which is a higher-voltage standard (+4dB, 1.23v) that has 20dB of headroom built in. That's the kicker, right there.

By "obeying" this standard in your DAW, you'll ensure that you too have 20dB to play with, not to aim for, but to play with. It's there to accommodate transients, hiccups, elephant farts... whatever might occur well above the "normal" operating levels around, say, -20dB, the rough equivalent to reading 0VU on your analogue-desk meters, that you might choose as "your" standard. Some might choose -18, -22 or whatever, but that's the general range that will reflect the amount of headroom that decent consoles offer, and that will ensure your signals play nicely with outboard gear should you route them out to it.

I've forgotten most of what I learned at AE school... twice; two courses proved I could study up, do well and then forget everything I was taught, if nothing else. LOL

I'll therefore defer to the GearSlutz post I'll link to, as the fella concerned is about as knowledgeable on this and related subjects as one could hope to be.

Point 3 - As for Easy Drummer and so on, now that you're aware of the context when we talk about levels, you might just want its levels to sound right without pushing the overall mix level too high. Michael would have a better idea of RMS levels than I, and he's recommended -18dB RMS and -10dB peak. Personally I'll aim for peaking at -10 or lower, even 'though I've really no idea what my RMS figures have been. Yes, I'm a naughty boy.

I'll tell you what, the mastering guys will love you if you hand them mixes that peak below -6dB. You might even be given a discount or encouraged to bring more work to them, 'cause the poor folk have to fight all day long with levels hotter than George Clooney's trousers which, I can only imagine, must be very fatiguing. Actually, not sure I want to imagine that...

Anyway, here's the thread. The first post is all we need focus on, I think. If you fancy a throbbing headache, feel free to read all 255 pages! LOL

The Reason Most ITB mixes don’t Sound as good as Analog mixes (restored) - Gearslutz.com

Sorry if there are still anomalous errors in this post; I copped the <p>; thing, whatever it's called, and had to edit every sentence as well as yours. Unfortunately I lost your smilies in the process as they were represented by hieroglyphics also, J. Take care mate.

As a side comment, most presets and patches that come with software instruments are set WAY too high in volume. The first thing I have to do when using SI/VIs is to insert a gain/trim plugin set to anywhere between -12 and -18 dB, so as to not run out of real estate at mix down.

EDIT: Just saw your interim post, SamBro'. Timely point, Brother.

Oh you're kidding! I just dragged the third link to this window for you, Michael, and the bloody board tried to load the page. Of course, I lost my long love-letter to you.

OK, to sum up:

Stavrou's skyscraper analogy stated that the analogue-gear sweet spot was, let's say for argument's sake, 60-80% of the height of the building, whereas digital's most accurate point is immediately below the 0, full-scale "height". I apparently had a vague image in my mind based upon the analogue analogy, which must've morphed over the years (Chinese whispers of my neurones?) into somehow representing the digital side of things. I'm very glad I looked into this, and I must thank you wholeheartedly for pointing me towards this misconception.

You see, I'd figured that even 'though the most bits are being employed at the highest levels, some real-world circuit limitations / component-quality considerations must've been preventing our getting the most out of the theoretical potential of all 24 bits, and that the sweet spot where the two factors converged must've been 6 to 12 dB down. What a goose I am!

Anyway, here are some threads discussing Stavrou's skyscraper analogy, which probably holds far less relevance these days as it was penned in 2003:

Ok here is why we should record to digital as hot as we possibly can!! | Cakewalk Forums

volume, bit depth, and quality - for the boffins | Apple Support Communities

setting recording levels

Something Bob Katz was quoted as saying (from the Fender Forum's "setting recording levels" post linked to above) at least supports something I said early in my rant about inter-sample peaks, which as I suggested, one doesn't hear an awful lot about:

Levels: Peak levels ideally should not exceed -3 dBFS on your meters. Sure you could go higher, but standard digital meters do not reflect intersample peaks which can be OVER 0 dBFS even if not shown on your meter. At 24-bit, you do not have to worry about signal to noise ratio and you will get a better result with a lower level and leaving some headroom for the mastering house.

Anyway, I'm about done, Michael. I promised myself a couple o' months back that I wouldn't get too deeply into these sorts (complex) of discussions as the energy required to type, research and edit each post kinda tips me over the edge and I can get quite ill. Let's blame chronic fatigue for that, or should we go fibromyalgia? Either way, I'm sure you understand my "predicament" - it's passion vs pain, brother!

Bottom line here is that I'm sure Sam, you and I all hoped to help our fellow Kemperites traverse the thorny wasteland that is the Realm of the Levels™, so to this end, I sure hope we succeeded. There'll be more to come, surely, and questions will follow, but I'd love to take the opportunity to (try to) bow out gracefully for now in the hope that my esteemed, more-talented and oh-so-much-more-knowledgeable colleagues are willing to uphold the service.

Er, in English (sorry!), that's, "Thanks for everything, fellas; I'll be off like a bucket o' prawns now". D'Oh! I think that was in fact Aussie.