Loueasy wrote:It is very confusing to me.
You're not alone in that! And part of the problem is understanding what is meant by the term 'headroom'...
Headroom is the space between the defined maximum working level of the system and the level to which normal programme peaks are allow to reach -- something which you or the manufacturer decides.
In the case of analogue equipment, the nominal 'alignment' or 'reference' level -- which is usually where the average level of normal signals sits -- is well defined. In European broadcast markets it is 0dBu while in professional recording studios it is +4dBu (in most cases). This alignment level is also shown on as a reference mark on metering systems (eg. PPM4, 0VU, Test, etc).
In an analogue system, normal programme peaks are allowed to rise 8dB (in the UK) or 9dB (in Europe) above that reference or alignment level -- so typically, +8dBu for UK broadcast, or +13dBu for a recording studio (and on meters it would be PPM6 or off the red end of the scale on a VU! :D ).
So normal programme peaks would be expected to fall somewhere in the yellow region of that virtual meter above. The average level would be expected to sit around the transition between green and yellow, but obviously the quieter bits will fall down into the green area. Very loud uncontrolled transients might reach up into the read -- and the headroom is there to accommodate them without overload distortion.
However, the 'maximum working level' in analogue equipment is a bit vague as the amount of distortion increases rapidly as the level increases, long before it actually clips. In the days of analogue tape machines the nominal peak level was generally chosen to keep distortion at or below 3% THD.
In practice, though, most analogue audio electronics will cope with much higher signal levels, albeit with rising distortion before the system finally clips. High-end analogue equipment like mixing consoles and outboard gear will usually cope with +24dBu before clipping (and some may go even higher), so +24dBu is generally taken as the maximum practical pro-audio level.
So with analogue systems, the manufacturers specify a nominal alignment level, and the user is expected to allow signal peaks no higher than 8 or 9dB above that... knowing that larger transient peaks can still be accommodated but possibly with some distortion. Some users will want to avoid that increasing distortion, while others might actively make use of it for musical colour...
In digital equipment, the situation is very
different. As the signal level rises the amount of distortion remains extremely low right up to the point where the converter runs out of quantising levels, and then it clips horribly!
So the maximum allowable level is clearly defined here (0dBFS)... but sadly there is no built-in 'alignment level' like there is in the analogue world -- that's (unhelpfully) left entirely to the user's own preferences.
Having said that, the SMPTE (in America) recommends an equivalent alignment level of -20dBFS, while the EBU (Europe) requires -18dBFS... and mastering engineers working on pop music might choose -15, or -12dBFS, or something else...
So that's 'headroom'. Where this becomes important (and confusing) is in the analogue-digital calibration -- how we translate between the analogue and digital domains.
In general, it makes sense to choose a calibration that provides similar headroom margins for both the analogue and digital domains, so that neither one runs out of space prematurely and thus becomes the limiting factor when moving signals between the two domains.
However, that's not always practical, depending on the equipment you're working with. For example, semi-pro devices may well not have the same peak-level capacity as high-end professional equipment, and some European equipment designed to the EBU spec will not be fully compatible with American equipment designed to the SMPTE spec.
I have a new Rme Fireface UFX II and it is really my understanding that in low/high gain settings have 15db of headroom not 19.
RME products usually offer several selectable calibration options, and on the UFX2 they are labelled as 'LoGain' and '+4dBu' on the line inputs, with the corresponding (if slightly confusing) 'HiGain' and '+4dBu' options on the line outputs. A third output alignment is labelled '-10dBV'.
Ignoring the -10dBV option for the moment, both
of the calibrations (lo/hi and +4) assume a nominal alignment or reference level on the analogue
side of +4dBu. The difference between them is that the lo/hi option is configured to reach maximum digital
level (0dBFS) at +19dBu, while the +4 options does so at +13dBu.
So, relative to the analogue alignment level of +4dBu, the lo/hi mode has 15dB of digital headroom (19-4 dB), while the +4 mode has 9dB of digital headroom (13-4 dB). The manual is correct in that!
Just for completeness, the -10dBV mode provides 12dB of digital headroom, but its analogue calibration is designed to match the semi-pro standard alignment level of -10dBV rather than +4dBu.
It might make more sense, and help to put all this in context, to look at a graph:
So here I've shown the professional analogue world on the left, with the 'alignment level' at +4dBu on the border between the green/yellow sections of a virtual meter. I've shown the the nominal working range (normal signal peaks) as a 9dB window in yellow (to match RME's specifications), and the remaining headroom in red above that.
As you can see, the SMPTE calibration exactly matches the nominal professional analogue world.
For comparison, I've also shown the EBU calibration which uses an alignment level of 0dBu and a peak level of +18dBu. It obviously has 6dB less headroom than the SMPTE standard, and that's because it is expecting 'controlled' broadcast-ready audio where someone has already mixed and managed the sound levels, so there should be no uncontrolled transients, and less need for a large headroom margin.
RME's lo/hi gain calibration sets a maximum peak level slightly (1dB) above the EBU requirement but it can't go as high as the SMPTE standard. In practice, this is rarely a problem as only the most enormous uncontrolled transients would normally be affected... but it does mean that you may need to be a little cautious when setting levels in high-end analogue equipment.
The RME +4 option reduces the analogue peak level even further, and is really intended for use with highly-compressed and commercially mastered audio where there is no need for a large headroom margin at all.
I would like to add a Grace Designs M108 via ADAT and possibly a Dangerous Music Convert 8 via ADAT. Considering the grace is +24 and alignments can be set to either -14 or -20 dbfs and the Convert can be aligned to with -14, -16 or -18 it gets confusing to me how all this would be calibrated with the Fireface.
So you're connecting some outboard preamps/converters via a digital interface, which means the RME's own internal converter calibrations are irrelevant. They only matter if you're passing analogue signals into or out of the RME directly.
So if you were taking the analogue line output of the Grace into the RME you'd need to think about the RME's calibration... but you're not as the conversion is all happening inside the Grace or Dangerous boxes.
However, if you might be connecting an analogue line source to either the Grace/Dangerous, or the RME's analogue inputs, AND you want the meters on those devices to give broadly the same kind of level indications, then you should choose similar calibrations.
In Hi/Lo mode the RME has a nominal 15dB range between alignment level and the maximum (0dBFS) level. So the closest option to that in the Grace is the -14 meter mode, and I'd go with the same setting in the Dangerous box.
Hope that helps!