Home Recording with Kemper - Setting Gain Structure

Welcome, Michael!

Boycotted digital interfacing as of 15 years ago, mate, but your question did have me scratching my head for a moment. Probably best I leave this one to the many Kemperites who use SPDIF.

You're right , master vol. has no influence in this case.

I set up my volume to to be around -6 dB in my DAW , using the profile's volume and sometimes the AMP volume (AMP menu) and all my own profiles are setup this way, it's very convenient.

You can dial the volume up to a point where the master led is the red on very hard attacks , as long as you don't hear any clipping.

In the output / master section you might want to configure SPDIF:

"Output Volume and Output Volume Link are also available for S/PDIF OUT."

I am not aware that S/PDIF needed additional leveling. Set the proper gain stage at the Kemper and just send it. The Audio interface should take the bit stream as is.

Quote from mikegrahamguitar

I can adjust the level using the volume knob for the individual profile but the master volume knob has no effect.

I think the volume knob for the individual profile is best used for balancing among different profiles (so you don't get any volume jumps when switching between them), not for setting the overall output volume. On average, aim to keep it somewhere around the middle to avoid internal clipping (output led flashing red) and to have room on both sides for leveling overly loud or soft profiles.

Like @SpinnerDeluxe says, there's a separate S/PDIF output volume parameter in the output section (press the Output button on the Kemper, then find the relevant page using left/right). If you activate the Link checkbox above it, S/PDIF output volume will also be affected by the Master volume knob. But since it's all digital anyway, I'm not sure if it's necessary to adjust it on the Kemper -- maybe someone else can help you there.

as a rule of thumb:

_Avoid digital leveling in your recording chain if you can. Start with everything at zero (doing nothing, so not all the way down) on your digital devices. (No clue where zero is with the kempers volume knob, would be interesting to know)
_Utilise all the bits of your daw...record as loud as you can.
_if you have to boost, do it analog if you can, before going in the digital domain.

Reason for this is, as an example with 16 bit devices:

If you cut the level of a 0db sound to half, only 8 of the 16 bits are used, if you boost it again in a second device, those 8 bits will be devided over 16 again, if you cut it in a 3rd application in, those 8 bits will be degraded to a 4 bit soundquality ..in a 16 bit system.

Most systems are 24 bits and up nowdays...so the effects of leveling are much less..but still there

Yeah, in the output/master section, there's a control for the spdif volume - use that

I myself shoot for hitting a bit hotter (-18 on a VU meter, or -10 dbfs on peaks), but the priciple is the same

You shouldn't be able to clip your interface using spdif, unless the kemper is already clipping, but it's good practice to ensure you don't get unwanted clipping/distortion from plugins later on.

Quote from Job Posthuma

If you cut the level of a 0db sound to half, only 8 of the 16 bits are used, if you boost it again in a second device, those 8 bits will be devided over 16 again, if you cut it in a 3rd application in, those 8 bits will be degraded to a 4 bit soundquality ..in a 16 bit system.

Cutting the volume to half is exactly one bit. No matter if from 24 bit to 23 or 16 bit to 15.

"_Utilise all the bits of your daw...record as loud as you can."

The dynamics of 24 bit is higher than any sound in this universe. Especially higher than any analog gear we currently have. The best tip for 24 bit conversion hence is: record well above the analog noise floor but by all means well below clipping (0dB).

as i remember it, every bit is equevilent with 6db.
Anyway....google maybe ts best friend on this...where very differnt points of view can be found on this one.

Quote from Job Posthuma

as i remember it, every bit is equevilent with 6db.

Correct.

-3dB is cutting the power by half.
-6dB is cutting the voltage by half
-10 dB is cutting the perceived volume by half.

All of that is far away from "losing 8bit"

Quote from mikegrahamguitar

Also, what are your thoughts on recording using the SPDIF output versus recording using the analog output

As you got the S/P-DIF input: go ahead.You cant get the Kemper sound cleaner than that. Its just the original. No matter how good the converters of the Kemper or your interface may be.

Quote from mikegrahamguitar

Also, what are your thoughts on recording using the SPDIF output versus recording using the analog output?

If you're going to an interface on a computer to record (so I guess that's most of us) then it will have to do an Analogue to digital conversion at the point you squirt it into the computer. If you're using the Kemper analogue out to go to a computer when you could use SPDIF, I personally think it's a bit silly as you're converting digital to analogue when it leaves the Kemper then just converting it straight back again. It's probably not going to lose right a lot of anything in particular but why bother? The Kemper puts out 1's and 0's and that's what the computer wants. Why not cut out the middle man?

As to recording 'hot'. I've been reading up on this lately (a little knowledge is dangerous I know ) and the theory seems to be to keep the level to -18 or so on each channel in the DAW as far as your signal level goes. Some of the plugins don't like it up 'em if you're too hot. Not my words, just what I've read of late. I'm still a relative beginner but have to say that since I've used a trimmer on the DAW to gain stage and turn everything down, things have come out better. When everything is summed together to the master, there is plenty of oomph still.

Quote from Gary_W

As to recording 'hot'. I've been reading up on this lately (a little knowledge is dangerous I know ) and the theory seems to be to keep the level to -18 or so on each channel in the DAW as far as your signal level goes. Some of the plugins don't like it up 'em if you're too hot. Not my words, just what I've read of late. I'm still a relative beginner but have to say that since I've used a trimmer on the DAW to gain stage and turn everything down, things have come out better. When everything is summed together to the master, there is plenty of oomph still.

Yeah, all true especially now that many more people (including us amateurs) use plugins which model analogue gear.

-18 dbfs is more or less what the old consoles had as "zero". Keep in mind that we're talking RMS/VU values here, not peak.

Some things - for some genres/applications can stand a bit of a hotter signal; unless there are plugins that don't react well to this.

Quote from Michael_dk

-18 dbfs is more or less what the old consoles had as "zero". Keep in mind that we're talking RMS/VU values here, not peak.

Many DAWs will show us peak though, at least more prominent. And still its good beginners advice. -18 dB peak means you are recording up to 21 bit. And this means you will get about 120+ dB of dynamics. More than any analog amplifier I am aware of. More than a full orchestra in tutti. And 18 dB of headroom is good enough not to wreck the tracks of even the worst drummers or bass players by clipping.

And no, we will *not* lose 3 bit of resolution. Unfortunately we audio people are using resolution the wrong way. The physical term resolution in audio is the sample rate (how fine or precise can the digital wave follow the original). The bit depth is all about dynamics. The largest difference between loud and soft. (Kindly ignore the quantization error, please. In 24 bit its OK for any practical use)

Funny enough: With our computer screens and fancy cinema beamers we all use the terms physically correct: screen resolution is about fine details is the number of pixels. 1024 pixels can resolve 512 lines. Times 2 = Nyquist. And the bit depth is all about colour or greyscales, hence dynamics.

EDIT: Overlooked that:

"Also, when you say, 'Set the proper gain stage at the Kemper', how does one go about that?

Is it just a case of having the selected profile at a satisfactory volume while ensuring the output LED doesn't flash red? Thanks!"

Yep. When you see red LEDs in input / output the Kemper tells you its reaching its limits. The factory profiles and good commercial profiles will make your life easy. Everything sounds fine with volumes about 0dB. Then adjust your master to what your (analog) recording interface likes. With S/P-DIF you should be fine anyway. But wtach the levels in your DAW if your keen on those -18dB. With high gain guitar you will not need so much headroom. With undistorted it depends. With (uncompressed) bass its a good idea.

The input is more interesting, though. Adapt the clean sense and distortion sense to each of your guitars and you will have the maximum fun browsing through the various profiles. Nothing should clip red and crunch should sound like crunch, not heavy-gain or no gain. Its all easy, the Kemper is kindof tolerant.

Quote from SpinnerDeluxe

The dynamics of 24 bit is higher than any sound in this universe. Especially higher than any analog gear we currently have. The best tip for 24 bit conversion hence is: record well above the analog noise floor but by all means well below clipping (0dB).

This.

Most DAWs using floating point data 32bit or even 64bit), set with the Mantissa quite high (normally you can assume that audio is primarily in the zero to one volume range). With floating point precision tends to bias towards lower values due to the exponent, so recording quietly would typically be just fine, but it's not the whole picture. While 32 and 64bit ADCs do exist it's still quite common to encounter 24bit ones, and these use integer values, so yes you drop a whole bit for a halving in volume, the issue though isn't really the loss of this one bit, but the increase in percentage of the signal that's now your noise floor.

So the take away from it all is try to strike a balance. Digital hardware will clip rather than distort nicely if the signal is to hot coming in from an analog source but you will be fighting your noise floor if you're too quiet. In both cases though modern hardware is very good, you can really afford to be quite lazy once you check that you're not clipping or barely moving the meter. The harder thing is turning down all the input pots and faders once you've got the data in there so you don't overload the desk, that's something I'm still working on.

Quote from Per

Most DAWs using floating point data 32bit or even 64bit), set with the Mantissa quite high (normally you can assume that audio is primarily in the zero to one volume range). With floating point precision tends to bias towards lower values due to the exponent, so recording quietly would typically be just fine, but it's not the whole picture. While 32 and 64bit ADCs do exist it's still quite common to encounter 24bit ones, and these use integer values, so yes you drop a whole bit for a halving in volume, the issue though isn't really the loss of this one bit, but the increase in percentage of the signal that's now your noise floor.

So the take away from it all is try to strike a balance. Digital hardware will clip rather than distort nicely if the signal is to hot coming in from an analog source but you will be fighting your noise floor if you're too quiet. In both cases though modern hardware is very good, you can really afford to be quite lazy once you check that you're not clipping or barely moving the meter. The harder thing is turning down all the input pots and faders once you've got the data in there so you don't overload the desk, that's something I'm still working on.

YES! Exactly! The bit depth has to do with noise floor, nothing else really.

I find this video awesome for clearing up misconceptions about bit depths and sample rates