-18dBFS = 0dBV, Always?

[hduncan]

Is this a solid mathematical relationship or does this figure vary between systems?

[UnDeFiNeD]

It vary's, but it seems -18dBFS has become something of a standard.
Most decent converters allow you to chose.

pzz

[MarkRB]

No, it is whatever you want it to be.
Digital full scale and analogue metering are different.
Analogue hardware can handle peak levels of +6. =7dB or even 12db without breaking a sweat, whereas digital just STOPS dead at 0dB.

So you have to scale analogue down to allow yourself the headroom in a digital world.
It is entirely up to you where you decide to put analogue 0 in comparison to digital 0. You just have to set your gear up to whatever you need or want.

Does that make sense?

I'm sure someone with way better skills will come in and elaborate.

[Farview]

It's what ever the converters are calibrated to. Motu is -15dbfs = 0dbVU for example. It should say in the specs of your converter or interface how it is calibrated.

My old DAT machine was calibrated to -12dbfs = 0dbVU.

Some converters allow you to calibrate them.

If you don't know where your calibration is, -18dbfs is a safe target.

[peat]

i usually go for -12dBFS=odBVU

try this in PT
put a signal generator with a 1k sine wave
after that put the bomb factory VU meter

i always get -12 on the signal generator giving 0dbVU on the meter

[narcoman]

Originally Posted by MarkRB

No, it is whatever you want it to be.
Digital full scale and analogue metering are different.
Analogue hardware can handle peak levels of +6. =7dB or even 12db without breaking a sweat, whereas digital just STOPS dead at 0dB.

So you have to scale analogue down to allow yourself the headroom in a digital world.
It is entirely up to you where you decide to put analogue 0 in comparison to digital 0. You just have to set your gear up to whatever you need or want.

Does that make sense?

I'm sure someone with way better skills will come in and elaborate.

its whatever you want it to be if you dont give a monkeys. The -18Dbfs is a standard suggested by the AES as representing 0. This allows you to go +12 into the red and process accordingly. If you line up the tape emulation plugins according to this, you'll find tht they behave a LOT more like their real world counter part rather than the mess that most seem to make o using them...

Set up according the AES -18Dbfs and you'll liberate yourself - no more overloading plugins is one benefit.... see the depth come back (halfway) to analogue recordings!

[MarkRB]

Originally Posted by narcoman

its whatever you want it to be if you dont give a monkeys. The -18Dbfs is a standard suggested by the AES as representing 0. This allows you to go +12 into the red and process accordingly. If you line up the tape emulation plugins according to this, you'll find tht they behave a LOT more like their real world counter part rather than the mess that most seem to make o using them...

Set up according the AES -18Dbfs and you'll liberate yourself - no more overloading plugins is one benefit.... see the depth come back (halfway) to analogue recordings!

Well, I was trying explain it in a simple theoretical way. As you and I know you need to calibrate your 0vu from an analogue system to 0.775 volts at 1 milliwatt and then set your ADs accordingly but I thought that would be confusing the issue.

[Farview]

Originally Posted by peat

i usually go for -12dBFS=odBVU

try this in PT
put a signal generator with a 1k sine wave
after that put the bomb factory VU meter

i always get -12 on the signal generator giving 0dbVU on the meter

Wouldn't that be testing the calibration of the VU meter plugin? If you don't have an analog VU meter on the signal generator, you don't know what you are feeding the converters. (at least without taking a voltage measurment.

[narcoman]

Originally Posted by MarkRB

Well, I was trying explain it in a simple theoretical way. As you and I know you need to calibrate your 0vu from an analogue system to 0.775 volts at 1 milliwatt and then set your ADs accordingly but I thought that would be confusing the issue.

sure !!

[MarkRB]

fuuck

[peat]

Originally Posted by Farview

Wouldn't that be testing the calibration of the VU meter plugin? If you don't have an analog VU meter on the signal generator, you don't know what you are feeding the converters. (at least without taking a voltage measurment.

i run this -12dbFS out into the console
with the channel line trims at 0
the channel VU reads 0dbVU

i am coming out of a 002 but

i also have an alesis adat bridge set to +4dBu
when i come out of one of its inputs
the channel line trim has to be a bit lower to get the channel VU to read 0

[Farview]

Originally Posted by peat

i run this -12dbFS out into the console
with the channel line trims at 0
the channel VU reads 0dbVU

i am coming out of a 002 but

i also have an alesis adat bridge set to +4dBu
when i come out of one of its inputs
the channel line trim has to be a bit lower to get the channel VU to read 0

The ADAT bridge must be calibrated differently than the 002.

[NathanEldred]

I find -16dbFS for individual tracks, while a little hotter @ -12dbFS for master bus (off a real console) works well.

[doug_hti]

as others are saying, it's what you want it (as long as the converter can calibrate). However in tv, it's still -20. I think it's crazy, as some (even $20k+) video decks do NOT sound good staying down that low if your console doesn't do well into the red...As some have mentioned and depending on how well or not your analog gear goes into the red...you may want to consider closer to -12. I like -14...you get plenty of signal into the converter and are less tempted to drive your analog gear too hard.

In general,
if you like to track hot into the converter, then calibrate closer to -12
if your gear doesn't have metering calibrate closer to -12
if your gear doesn't distort or get "hot" in a nice way go closer to -12

learn your analog front ends sweet spot and calibrate accordingly.

if you are on a large format console and tracking many channels of a band simultaneously, then I would go closer to a -16 or -18 so while going into your daw you can have the "faders" all up and not be completely hammering the mix bus....

[postprosound]

Not that it relates to the music side of sound directly, but for those of us in post production for film and tv, we use -20 dbfs = 0vu.

[slipstream]

Originally Posted by postprosound

Not that it relates to the music side of sound directly, but for those of us in post production for film and tv, we use -20 dbfs = 0vu.

+1thumbsup

[DAWgEAR]

See this article ...

THE INS AND OUTS OF

[YZ!]

Originally Posted by MarkRB

No, it is whatever you want it to be.
Digital full scale and analogue metering are different.
Analogue hardware can handle peak levels of +6. =7dB or even 12db without breaking a sweat, whereas digital just STOPS dead at 0dB.

Professional analog gear rated for +4dBm operating level usually have a a maximum I/O level of +22 to +28 dBm. (*)

If you calibrate your system so that -18dBFS = 0VU on a set of VU meters calibrated to read zero when presented with a +4dBm signal, you have full VU and peak metering and can use the full available dynamic range of both digital and analog gear.

Do not confuse measurement units...

dBm: measured relative to the power of one milliwatt in a 600 ohm load;
dBv: measured relative to a zero set at 0.775v (which is 0dBm if you have a 600 ohm load, btw);
dBV: measured relative to a zero set at one volt;
dB FS: means "decibels full-scale", used in digital systems and it is always equal or less than zero;
VU: Volume Units, can only be compared to dB on sine waves and when the "zero" reference is given; if no reference is given it is assumed to be zero VU= +4dBm; there's no "dB VU".


Personally, I find the old VU meter very handy for setting the overall levels of a mix while the digital peak meters tell me very little about loudness and density; OTOH, digital meters are better when tracking to check for actual headroom.


(*) check the specs of your analog gear for its maximum I/O level when defining your reference.

[hduncan]

I am using the apogee duet, so the specs for that should be what governs my relationship between dbfs and 0dbvu right?

I get the point that since analogue and digital react differently that the whole formula exists on the first place.. I still don't really get why we have this concept in the first place...

What is optimal analogue level and What happens at 0dBVU?

The saw can handle anything up to 0dbfs no problem, but its my converter that's holding me back, right?

I do everything ITB with softsynths and no recorded audio yet, do I even need to care about this?

[kooz]

Originally Posted by YZ!

Professional analog gear rated for +4dBm operating level usually have a a maximum I/O level of +22 to +28 dBm. (*)

If you calibrate your system so that -18dBFS = 0VU on a set of VU meters calibrated to read zero when presented with a +4dBm signal, you have full VU and peak metering and can use the full available dynamic range of both digital and analog gear.

Do not confuse measurement units...

dBm: measured relative to the power of one milliwatt in a 600 ohm load;
dBv: measured relative to a zero set at 0.775v (which is 0dBm if you have a 600 ohm load, btw);
dBV: measured relative to a zero set at one volt;
dB FS: means "decibels full-scale", used in digital systems and it is always equal or less than zero;
VU: Volume Units, can only be compared to dB on sine waves and when the "zero" reference is given; if no reference is given it is assumed to be zero VU= +4dBm; there's no "dB VU".


Personally, I find the old VU meter very handy for setting the overall levels of a mix while the digital peak meters tell me very little about loudness and density; OTOH, digital meters are better when tracking to check for actual headroom.


(*) check the specs of your analog gear for its maximum I/O level when defining your reference.

Indeed, the digital scale (dBFS) is in reference to headroom. If you're sending a signal into your A>D converters from a piece of equipment like a preamp that has a clip point of +28dBm, it makes sense to calibrate your converter's input to clip just below (or at) that same voltage level.

Take the digi oo2 as an example: page 46 of the Getting Started guide that came with it states that "All Audio Inputs and Outputs are set for 14 dB of headroom below 0dB, or full code...nominal reference input level is +4dBu or -10dBV" It goes further to note that "maximum input and output...is +18dBu (6.15 VRMS)." The math: +4 nominal plus 14 dB of headroom equals +18dBu. But if an external preamp feeding into a oo2 analog input clips at +24dBu, you likely have (in general) about 6dB too much gain at the preamp for the converter to deal with. The preamp is fine, but the input converter chokes (clips) sooner...and that's the clip that really matters, because the math the software (and the hardware prior to it) runs on is based around that spec.

As noted above, when calibrating your converters you must be mindful of the different voltage/impedence references so that you're comparing apples to apples: gain staging your system properly so that everything clips at the exact same voltage level- the preamps, busses, direct outs, etc. It's a bit of a daunting task for the uninitiated, but once it's been addressed (especially if you have brilliant mics, preamps and conversion, along with proper technique using those tools and a space with a good acoustic to record in), digital is a wonderful sounding place.

Here, so you don't have to do too much math: dB dBu dBV dBm to volts conversion calculator volt - volts to dBu and dBV dB - convert dB volt calculation online attenuation loss gain ratio reference - sengpielaudio

[Tim Farrant]

This is a very interesting topic, and one I have a very strong interest in. 0dbVU (on a standard VU meter) = +4dbu (or formerly known as +4dbm). This relates to a signal voltage level of 1.23V RMS. To obtain peak level of 0dbFS on the DAW requires that the analog driving stage will need to deliver +22dbu (as mentioned above). Here is where it gets important because in many caes, this +22dbu level will be at or close to clipping in many analog equipments.

In the old days of analog tape, you simply could not record these levels without undue distortion, with digital however, this is not the case. So analog electronics has to deliver more than ever greater levels, and quite simply, the analog electronics needs to be able to deliver this higher level and ultimately more power. Suddendly, op amps with +/- 15 volt rails seem inadequate for this task, because they are operating so close to maximum output and this is going to sound bad, believe me. This is why perhaps much low end gear still using +/-15V rails fall short and get a bad name, especially on percussive sounds like drums.

My advice, if you have a low cost mixer or outboard that runs on low power supply rails, use a higher dbFS setting like -12 or -14. This will put less stress on your analog, and therefore better overall sound.

Tim.

[nucelar]

Originally Posted by postprosound

Not that it relates to the music side of sound directly, but for those of us in post production for film and tv, we use -20 dbfs = 0vu.

when speaking about postproduction, -20 dbfs is a mainly US standard, here in europe we use -18dbfs as standard. I believe -18 is an AES standard or reccomended practice.

[jimmyjimjim]

Originally Posted by Tim Farrant

T
My advice, if you have a low cost mixer or outboard that runs on low power supply rails, use a higher dbFS setting like -12 or -14. This will put less stress on your analog, and therefore better overall sound.

Tim.

Brilliant post, Tim.
Regarding the stuff quoted above.
I always advice folks to stay at the usual calibration (-16 to -20 or wathever) and simply just record lower into digital.
Hovering around -20 to -10dBfs, is a great place to be, and gives you headroom for processing in the digital domain without the need for trimming the signals.

[psycho_monkey]

Where I work we use a lineup of -18 with the SSLs and -16 for the Neves.

One issue I have noticed is that when printing back to the DAW, -18 can be quite quiet for the SSL - people find they have to use makeup gain to get a reasonable level (the compressor isn't modded a la clearmountain, so it kicks in quite early).

To get round this, on the B-trims of the 192s, we have pairs set to -16, -14, and -12. I find -14 makes sense for 2-track printing - that's what I'm used to using when printing to DAT.

[davenutz]

Originally Posted by nucelar

when speaking about postproduction, -20 dbfs is a mainly US standard, here in europe we use -18dbfs as standard. I believe -18 is an AES standard or reccomended practice.

When dealing with stuff from the UK, we often get notes to indicate that their gear is calibrated for -14.

We keep our 192's calibrated to -20 on the "A" trim, and -14 on the "B" trim

[hduncan]

if daws gave an absolute ceiling of 0dBFs, what is the maximum possible for analogue before clipping?

What does 0dBFS = as far as real world decibels? That shouldn't vary right?

[Farview]

Originally Posted by DarkEcho

I am using the apogee duet, so the specs for that should be what governs my relationship between dbfs and 0dbvu right?

Yes.

Originally Posted by DarkEcho

I get the point that since analogue and digital react differently that the whole formula exists on the first place.. I still don't really get why we have this concept in the first place...

analog and digital don't react differently. The meters we use react differently.

A VU meter measures the average level of the signal. This corresponds to how we hear the loudness of the sound.

A digital meter measures the peak level of the signal. This is used because there is a hard ceiling on how much signal can be represented digitally.

Originally Posted by DarkEcho

What is optimal analogue level and What happens at 0dBVU?

0VU or 'line level' is the optimal level. That is the level that the equipment was designed to run. That is where there is the best signal to noise and the lowest distortion.

Originally Posted by DarkEcho

The saw can handle anything up to 0dbfs no problem, but its my converter that's holding me back, right?

The analog part of the converters and anything they are hooked to.

Originally Posted by DarkEcho

I do everything ITB with softsynths and no recorded audio yet, do I even need to care about this?

The only thing you have to worry about is overloading the plugins. Any plugin that emulates hardware will distort (just like the hardware) when hit with a signal too hard. There is also the issue of all the tracks adding up and clipping the master buss.

[Farview]

Originally Posted by DarkEcho

if daws gave an absolute ceiling of 0dBFs, what is the maximum possible for analogue before clipping?

That would depend on the specific equipment you are using. Sometimes it's +18dbu, +24dbu, +28dbu. It is not standard, but normally better equipment has more headroom.

Originally Posted by DarkEcho

What does 0dBFS = as far as real world decibels? That shouldn't vary right?

That depends on the calibration of the converters. If the converters are calibrated to -18dbfs =0dbVU, then 0dbFS = +18dbVU. So if 0dbVU = +4dbu then +18dbVU = +22dbu.

0dbFS = +18dbVU = +22dbu.


Note that we are using three different db scales. Each scale references a different voltage, that is why it gets confusing.

[hduncan]

Ok, so in a purely digital realm, there is no optimal level (well, technically as close to 0dbfs without overs, but the difference between that and say, -20dbfs is so negligible that its just easier to say there isn't a functional optimal). So the full scale metering makes sense as showing the total possible range of level from -inf. To 0dBFS.

In analogue, there IS a functional optimal where the signal is best placed for both s/n ratio and distortion reasons, which is 0dB(vu?) even though that's not the ceiling (which is dependent on the weakest link in the analogue chain) so comparing the two meterings isnt very reasonable.. Apples to oranges.

However! There is a need to recognize an equality because at some point your signal will pass through analogue equipment (converter) and you don't want to exceed its optimal level.

Is this logic acceptable? I think I'm grasping this better.

Now that said, exept for going to my monitors, I don't need to worry about the sweetsoot of the analogue side converter IF I am working purely on digital. This is because different converters have different operating levels and levels that are fine on digital may distort after a crappy DA might not after a better DA. Is this correct?

[Farview]

Originally Posted by DarkEcho

Ok, so in a purely digital realm, there is no optimal level (well, technically as close to 0dbfs without overs, but the difference between that and say, -20dbfs is so negligible that its just easier to say there isn't a functional optimal). So the full scale metering makes sense as showing the total possible range of level from -inf. To 0dBFS.

In analogue, there IS a functional optimal where the signal is best placed for both s/n ratio and distortion reasons, which is 0dB(vu?) even though that's not the ceiling (which is dependent on the weakest link in the analogue chain) so comparing the two meterings isnt very reasonable.. Apples to oranges.

However! There is a need to recognize an equality because at some point your signal will pass through analogue equipment (converter) and you don't want to exceed its optimal level.

Is this logic acceptable? I think I'm grasping this better.

You seem to have it.

Originally Posted by DarkEcho

Now that said, exept for going to my monitors, I don't need to worry about the sweetsoot of the analogue side converter IF I am working purely on digital. This is because different converters have different operating levels and levels that are fine on digital may distort after a crappy DA might not after a better DA. Is this correct?

You are correct.