Please help me to calibrate meters on Focusrite ISA preamps

You are getting horribly confused now. If you use too little gain, then you have to make it up elsewhere. Your entire signal chain has a hand in this. The usual system for adjusting gain structure means you use the quietest amplifier to do the job. You are also maybe overthinking the +10dBu feature - boosting gain, and using attenuators is a destructive connection system. Sometimes inaudible, sometimes not. Where exactly on the knob is unity gain? On my kit, in truth I have no idea, and all that matters is I have a quiet system. In practice, it means that you have a series of level adjustments in a chain. You want minimum noise, minimum distortion and maximum level end to end. You experiment and find a set of variables that produce this. Seeing full scale/red is bad. It takes a bit of effort to maximise levels from in to out leaving just enough headroom, but I really doubt many manufacturers could give you the position on a rotary gain knob that is unity. Just not how we do things, or need to do things.
 
But the guy from Focusrite said that 2 or 3 o clock is unity gain, so what is 1 o clock? That means the signal is being attenuated! Am I wrong?
Unity with respect to what?!? Does he mean that’s where you have to set the knob so that +4dbu signal hits your DAW at -6dbFS like the back holes? Fuck I don’t think I care where “unity” is.

My point was that it’s almost always variable attenuation into fixed gain. That is, the input is only NOT attenuated when the knob is all the way up. Then there’s the gain state that just does what it does.

Every time you say the ISA is a clean preamp, I wonder why tf you’re even using it. Then I remember it actually has some other cool features like variable input impedance and stuff. In that case, the only impact of lower gain might be more noise.

So yeah I guess if you tell the ISA that 0 is 24, then attenuate by 15, you’ll be at 9, which will be -1, so you’ll never clip the ADC. It’s not ideal gain staging. You might reavaluate what you’re really getting from the ISA and whether it’s worth the added noise and expense and hassle. Maybe it is for you, and that’s fine. If you want to try a different flavor some time, you could put a passive DI in between.
 
The concept of "Unity Gain" is a valid one where there are multiple devices in an audio chain and especially if those devices are likely to be inserted and at times removed from said chaiin.

Thus we get the concept of the 'Operating Level" where every pice of line level equipment runs at an agreed input and output signal voltage.

"Professional" kit such as say a parametric equalizer would be a PITA to use if you had to cluck about making sure it did not clip or introduce extra noise each time you pathced it into a signal path. Same reasoning applies to compressors, reverbs and all other hardware gear, almost all of which is configured around a patch bay.

The most common OL is of course +4dBu (which is, for reasons only to do with confusing newbs!) A voltage of a bit over 1V rms and indicates 0 on a VU meter. That means that almost all pro gear has a headroom of +20dBu (nearly 8V rms) at least and many units 6dB or more above that.

But when we come to 'prosumer' audio kit we are in s*** in a handcart! Virtually no two mnfctrs agree what input will deliver 0dBfs in a DAW from their digital kit..FORK! They are often not even cnsistent over the ranges of their kit! The best you can do really is hook it all up and tune for maximum smoke. I JEST!! Audio kit will NOT catch fire no matter how you connect it. (don't take liberties with powerful amplifiers tho'but)

BTW connecting an attenuator to the output of a hot '+4dBu' preamp will not worsen the noise performace UNLESS you attenuate so much you have to ise gain to restore the signal level. Any noise in the pre amp signal is attenuated by precisely the same amount as the hot signal.

Dave.
 
I guess we generally assume that “unity” means that voltage going in is the same as the voltage coming out, but who can say what these people are talking about. Apparently their website is lying about the max input on the rear panel inputs which then makes you wonder if any of those specs can be trusted. What’s the actual max input on the front panel ones or the mic inputs? What’s the maximum from the line outs. That also says +18dbu, but is actually also only +10?!?

Please reply to this person at Presonus and tell them for me that the should be ashamed of themselves for trying to pass of a unit with +10dbu max “line” input as anything other than a toy and that they should be flat out sued for actual misinformation on their spec sheet.


Just to be that guy, (since [MENTION=89697]ecc83[/MENTION] kinda went there) the idea of a “maximum rms” in this context is meaningless. The limit is the voltage rail and that is all about the actual instantaneous voltage at any given point, and has nothing to do with average level. With real world signals we can’t make any meaningful assumptions about the relationship between rms and p2p, so when talking about this kind of headroom limit, we really should talk in p2p, I think.
 
I guess we generally assume that “unity” means that voltage going in is the same as the voltage coming out, but who can say what these people are talking about. Apparently their website is lying about the max input on the rear panel inputs which then makes you wonder if any of those specs can be trusted. What’s the actual max input on the front panel ones or the mic inputs? What’s the maximum from the line outs. That also says +18dbu, but is actually also only +10?!?

Please reply to this person at Presonus and tell them for me that the should be ashamed of themselves for trying to pass of a unit with +10dbu max “line” input as anything other than a toy and that they should be flat out sued for actual misinformation on their spec sheet.


Just to be that guy, (since [MENTION=89697]ecc83[/MENTION] kinda went there) the idea of a “maximum rms” in this context is meaningless. The limit is the voltage rail and that is all about the actual instantaneous voltage at any given point, and has nothing to do with average level. With real world signals we can’t make any meaningful assumptions about the relationship between rms and p2p, so when talking about this kind of headroom limit, we really should talk in p2p, I think.

Err? No Ash'. I quoted voltages in rms because that is the way signal and other voltages are 'usually' specified. 'RMS' in in fact usually assumed and not given.

There has grown up a tendency in audio circles to quote "peak" signal levels and this is rather meaningless. The terms "peak" "average" "rms" "peak to peak" have precise meanings in electrical and electronic engineering but are bandied about willy-nilly in audio writings.

The term "rms" simply means that voltage that would produce the same heating effect in a resistor as DC. Thus our mains supply is nominally 230V rms but if you put 230V DC from batteries into your kettle it would make your cuppa just as quickly. The relationship between 'peak' and rms depends upon the waveform. For pure sine divide peak (or 1/2 pk-pk) by 1.414 but for a square wave the peak and rms value is identical. Other waveforms will have other relationships.

I will agree this is all a mess. For the representation of signals in a DAW we are only interested in the peak value, how close it comes to 0dBfs. Often it is said we should be recording at -18dBfs as an "AVERAGE" level but average in this context simply means "as our eye/brain" interprets the visual display over a short time period. The meters are still reading the peak of the signal (as the software computes it).

I also agree that signal levels are, in general limited by supply rails but not always! A 'pro' balanced output will use two amplifiers of opposite polarity and deliver twice the supply limited voltage (rms!) If the output comes from a transformer you can have just about any signal voltage you fancy!

A word is perhaps moot here about "true rms" digital test meters? They only read rms for power frequencies, up to 400Hz, 1kHz if you are lucky! Do not trust them to give an accurate reading for audio signals or noise.
They also have a VERY limited HF response. Most are 6dB down by 2kHz and some by 1kHz. Digital meters are superbly accurate, even the £20 ones and have a high input impedance and are in many ways the best things to come into electronics testing since bread-come-cut but, if you want to calibrate your Revox B77...FIRST calibrate your DMM!

(I see Mary Whitehouse lives on in the forum censor!)

Dave.
 
You are getting horribly confused now. If you use too little gain, then you have to make it up elsewhere. Your entire signal chain has a hand in this. The usual system for adjusting gain structure means you use the quietest amplifier to do the job. You are also maybe overthinking the +10dBu feature - boosting gain, and using attenuators is a destructive connection system. Sometimes inaudible, sometimes not. Where exactly on the knob is unity gain? On my kit, in truth I have no idea, and all that matters is I have a quiet system. In practice, it means that you have a series of level adjustments in a chain. You want minimum noise, minimum distortion and maximum level end to end. You experiment and find a set of variables that produce this. Seeing full scale/red is bad. It takes a bit of effort to maximise levels from in to out leaving just enough headroom, but I really doubt many manufacturers could give you the position on a rotary gain knob that is unity. Just not how we do things, or need to do things.

I am not making this up. FOCUSRITE told me that unity gain is 2 or 3 o clock. They told me this to give me an idea of how I could use outboard preamps without further gain or attenuation from the front knobs on the Clarett unit. You are right, I am confused, just not about that.
 
BTW connecting an attenuator to the output of a hot '+4dBu' preamp will not worsen the noise performace UNLESS you attenuate so much you have to ise gain to restore the signal level. Any noise in the pre amp signal is attenuated by precisely the same amount as the hot signal.

Dave.

That makes sense!
 
BTW connecting an attenuator to the output of a hot '+4dBu' preamp will not worsen the noise performace UNLESS you attenuate so much you have to ise gain to restore the signal level. Any noise in the pre amp signal is attenuated by precisely the same amount as the hot signal.

Dave.

That makes sense!
 
I think you still seem to be confused? Dave pointed out that it's not destructive attenuating, until you have to add gain somewhere else, and your structure seems to be cut, then boost, then cut a bit more and boost again. Why not settle your mind with some experiments. Record some stuff and analyse which sounds better, noise wise? Have you considered that the reason Focusrite told you 2 OR 3 o'clock is simply because they don't know the exact answer because it's somewhere around there, and for them, that's close enough? You seem to want absolutes, and that rarely happens. We're not criticising or implying anything, just that you're perhaps chasing rainbows for exactness.

I'm actually quite convinced some manufacturers deliberate have a little more headroom than their meters and alignment suggest. Zoom recorders are good examples. Accidental over-level, as in seeing the 0dB maximum hit, often doesn't cause any distortion and the waveform doesn't have a flat top, because a 0dB indication is maximum, and the chances of the signal 'just' tickling 0 are slim - even .5dB over would only read 0, and have a flat top, but you have to go over by quite a bit to produce distortion. I have a JVC video camera that does the same thing, but I have a Lexicon interface that going above -3dB is quite dangerous as the sound definitely in danger going higher than that. My current Presonus interface seems very tolerant of over level as long as you don't push it, while my old Tascam had a red LED that really did mean trouble. Gain staging now for me just means maximising dynamic range really. Transformer DIs with no electronics work fine for me, but they do change the sound, which I can live with, but never mix two different brands for left and right because then you really get aware of what they're doing to the signal.

All I can suggest is doing some tests so you can see which bits of your kit are noisy and which aren't. Which can boost invisibly and which can't. Who cares at which knob position unity is achieved. That might not be the best for your end result if a noisier device follows in the chain. If you can get cleanly hot from one device, the noise ones after are less destructive. As an example, the current temporary system in my office here works better when the control room volume in Cubase is turned up higher than normal and the monitors are turned down. Never used these monitors before but using the control room level recorded at it's normal setting just sounds less good with these speakers. No idea why, but turning it up seems to restore what I get back home.
 
Err? Yes Rob!

There are situations where absolute levels need to be preserved of at least 'watched' to prevent overs. Whilst tape recording is tolerant of a db or so over, disc cutters are not! (ok, yes, they have very sophisticated limiters but you don't want to trip those either)

Radio stations strive to keep transmitted levels consistent (well, if you are the BBC and you GAF you do!) and with the myriad signals that can feed in to make up a programme, levels must be 'evened up'. In the days of Dolby A it was vital that levels were adhered to with in a dB or less else frequency response went to the dogs.

Not so with home recording or even 'semi pro' work. Certainly level exchanges between modern gear has a vast dynamic 'space' in which to work and as Rob says "give it a go and give it a look" !

What baffles me, as I have said before, is WHY have a meter scaled in dB fs for a piece of soley analogue gear? And IF you are going to do that give some very clear information as to how to internally calibrate the kit AND include a line up oscillator. Plus, give clear instruction as to how it all relates to external equipment. I sense rhe dead hand of the Ad Puff dept there.

Dave.
 
What baffles me, as I have said before, is WHY have a meter scaled in dB fs for a piece of soley analogue gear?
I kind of see the logic.

It’s pretty obviously meant to be a replacement for onboard interface preamps or more likely just plug straight into an interface without preamps of its own. In those situations the actual analog level doesn’t really much matter. In some situations it might be awkward or impossible to see the actual computer while setting gain on the preamp, so it could be convenient to have that mirrored there.

And it’s really pretty straightforward to set up as long as you can trust the specs your interface gives you and of course that spec falls within the range of adjustment for the preamp. The problem we have here is that neither of those conditions are true. The Presonus’s published spec lies AND is pitiful and falls way outside the range of calibration!

You mentioned a test tone, and yeah if that calibration knob is actually just a pot, you’d pretty much have to use a tone to get exact settings anywhere but the two extremes. Honestly, though, that could be almost anything. I don’t really fault Focusrite for expecting some basic competency on the part of its customers. It has some advanced features that put it out of the class of “My First DAW”.
 
hi,

I'm just wondering how to do factory reset on ISA Two, user guide mentions it in peak meter calibration section, but there is no explanation how to perform it?
 
hi,

I'm just wondering how to do factory reset on ISA Two, user guide mentions it in peak meter calibration section, but there is no explanation how to perform it?
""Calibration The LED meters can be calibrated using the PEAK METER CALIBRATIONcontrol on the rear panel (see the Rear Panel diagram for the exact location) enabling the 0 dBFS point on the LED meters to correspond with that on an external A/D converter. The meters are calibrated in the following way: LED Meters To calibrate the LED meters, use the rear panel PEAK METER CALIBRATIONcontrol. With the control in the default central (‘detented’) position, 0 dBFS is equivalent to an analogue audio level of +22 dBu. Rotating the control in either direction sets a new value for 0 dBFS from +16 dBu (fully anticlockwise) to +24 dBu (fully clockwise). When performing a factory reset, the peak meter calibration control must be in the central (‘detented’) position ."

^ from the user manual.

???????

Dave,
 
AFAIK the ISA 2 has no digital hardware in it? (is an option) therefore "factory reset" is something of a misnomer. All you really seem to need to do is get the pre amp's meters to correspond with "0dBfs" in your DAW.
Note, this is really just something of a convenience and should optimize gain staging. There is no need these days for precise levels as there was in the days of Dolby. Always good practice to maintain fairly consistent levels tho'but.

Dave.
 
Back
Top