|
|
Post by bram on Feb 12, 2020 11:49:09 GMT -6
I'm trying to wrap my head around digital and analog reference levels. I have my Apollo x8/x16 system's Output Reference Levels set to +4dBu, with Headroom set to +24 dBu. In the DAW signal generator I generate a 1khz sine wave @ -18dBFS and send it out of x16 Line Out. My CAPI Sumbus's (+4 dBu Inputs) VU meters read this signal as ~0 VU Silverbullet (also +4 dBu) also reads the -18dB signal as ~ 0 VU When I measure the signal with a True RMS voltmeter set to AC Volts , it reads .781V. According to ( Sengpielaudio): .7746 volt = 0 dBu 1.228 volt =+4 dBu or 0 VU Why then is my 0 VU (+4 dBu) signal being measured as .781V (0 dBu) on the multimeter, is this because I'm not using a 600ohm load in the measurement? In order to reach ~1.228 V on the meter, I have to generate a -14dB signal from the DAW (1.241 volts). (I also measured at 100Hz @ -18dBFS = .785 volts, not much difference from the 1khz signal (using a Fluke meter, also double checked on a secondary meter). Sumbus still reads ~0VU although the Silver Bullet shows a slightly hotter signal at 100Hz with the 2dB LED illuminated.) But then it seems this article is saying 0VU = .775 volts (0dBm@600Ω = 0dB = 0VU = 0.775V) "...not only was 0dB the reference, it also became known as 0VU (Volume Units). This handy scale, although directly linked to dB, removes the reference to dBm (and thereby any reference to impedance) and bases the scale purely on voltage with 0VU equal to 0.775V. This new scale, as it was no longer based on power, needed a new designation and became known as dBu. As this scale is based purely on voltage, and a multimeter can read voltage, the dB scale on a multimeter is centered on this voltage referenced scale. Please note: when we talk of VU we are talking true VU, not as an indication on a piece of equipment when running at optimum signal, for example a tape recorder." To further confuse matters for me, Matt Hepworth over at uadforums stated that for the Apollos, "-16dBFS or -20dBFS, depending on the setting, equals +4dBu/0dBV" and not -18dBFS, which is what I was outputting. Any clarification on the subject would be very helpful, thanks! |
|
|
|
Post by matt@IAA on Feb 12, 2020 12:39:24 GMT -6
Ok, so +4 dBm = +4 dBu if and only if the load is 600 ohms. As your article says this is based on the old 1mW = 0 dBm into 600 ohms @ 0.775VRMS sine. In modern gear we ignore power handling (dBm) and stick to a simple voltage gain (modern power transfer is low since input impedance is high - we want voltage transfer, not power). This standard is defined around dBm into 600 ohms, where 0dBu = 0.775 VRMS sine, but is strictly based on the voltage reference. So 0.775VRMS is always 0dBu regardless of load. This is a useful little tool to compare. As you can see with a 10k impedance, +4dBu is -8.2 dBm. Same voltage, much lower power required. www.analog.com/en/design-center/interactive-design-tools/dbconvert.htmlNote in the article about VU meters you linked that he is talking about "true VU". This is a moving target. See how he says "not as indication on a piece of equipment running at optimum signal"? That means "true" VU is out the window and 0VU on that equipment is an arbitrary reference level relating to the headroom of the unit, much like dBFS. The manual for the x16 says the maximum output level when set to +4 dBu reference is +24 dBu. So this implies that +24dBu = 0dBFS. What Matt Hepworth said then should be read as - when in +24dbU max mode, +4 dBu is found at -20 dBFS. So, at -20 dBFS you should read 1.23VRMS. At -18dBFS you should have been outputting +6 dBu, or 1.546 V RMS sine. I assume you're measuring this floating, pin 2 to pin 3 or tip to ring? As for the other gear, the SUMBUS operates at -2 dBu internal right? And a +4 dBu input produces a -2 dBu output? In which case, 0VU = -2dBu = 0.773VRMS - very close to what you measured. |
|
|
|
Post by bram on Feb 12, 2020 12:55:56 GMT -6
Thanks Dogears, I’ll read through this more closely when I’m back in the studio tonight.
According to the SumBus product page, “ The SumBus inputs are at +4dBu nominal and the outputs are -2dBu nominal.”
|
|
|
|
Post by matt@IAA on Feb 12, 2020 13:08:28 GMT -6
Right that means there’s a 6 dB loss, but not necessarily what 0VU is set to. More than likely though 0VU is +4dBu.
|
|
|
|
Post by svart on Feb 12, 2020 15:44:32 GMT -6
A little late to the discussion, but as a practical matter, most DMMs have a much smaller RMS measurement bandwidth than 1khz, or they have high error at that frequency.
My fluke has a 500hz bandwidth for 2% error, and at 1khz it's more like 5%. My other DMM won't measure much above 200hz at all.
|
|
|
|
Post by johneppstein on Feb 12, 2020 16:37:00 GMT -6
A little late to the discussion, but as a practical matter, most DMMs have a much smaller RMS measurement bandwidth than 1khz, or they have high error at that frequency. My fluke has a 500hz bandwidth for 2% error, and at 1khz it's more like 5%. My other DMM won't measure much above 200hz at all. Methinks that a good VTVM would probably have significantly better bandwidth? |
|
|
|
Post by BenjaminAshlin on Feb 12, 2020 18:52:30 GMT -6
A little late to the discussion, but as a practical matter, most DMMs have a much smaller RMS measurement bandwidth than 1khz, or they have high error at that frequency. My fluke has a 500hz bandwidth for 2% error, and at 1khz it's more like 5%. My other DMM won't measure much above 200hz at all. Methinks that a good VTVM would probably have significantly better bandwidth? Most VTVM are not accurate at such low AC voltages.The HP-400D VTVM would be one such exception. My Fluke 287 is suppose to have 100kHz bandwidth with accuracy ~0.3% 0-20k and ~3.5% 20k-100k. I have no idea if this is true in reality. Make sure you terminate with a 600ohm when you measure. RME does my head in with their reference levels. My old Apollo had no issues. |
|
|
|
Post by johneppstein on Feb 13, 2020 13:19:39 GMT -6
Methinks that a good VTVM would probably have significantly better bandwidth? Most VTVM are not accurate at such low AC voltages.The HP-400D VTVM would be one such exception. My Fluke 287 is suppose to have 100kHz bandwidth with accuracy ~0.3% 0-20k and ~3.5% 20k-100k. I have no idea if this is true in reality. Make sure you terminate with a 600ohm when you measure. RME does my head in with their reference levels. My old Apollo had no issues. The HP400D is the one I have. Also have the matching signal generator.
Neither of my Flukes is that good, but I don't have a 287.
|
|
|
|
Post by bram on Feb 13, 2020 16:10:26 GMT -6
The manual for the x16 says the maximum output level when set to +4 dBu reference is +24 dBu. So this implies that +24dBu = 0dBFS. What Matt Hepworth said then should be read as - when in +24dbU max mode, +4 dBu is found at -20 dBFS. So, at -20 dBFS you should read 1.23VRMS. At -18dBFS you should have been outputting +6 dBu, or 1.546 V RMS sine. I assume you're measuring this floating, pin 2 to pin 3 or tip to ring? As for the other gear, the SUMBUS operates at -2 dBu internal right? And a +4 dBu input produces a -2 dBu output? In which case, 0VU = -2dBu = 0.773VRMS - very close to what you measured. Thanks for the replies everyone. So first things first, total face palm moment, I was measuring between hot and ground, which was accounting for the incorrect measurement. When measuring from pins 2 to 3, -20dBFS = ~1.204. The Fluke 87 measurement stays consistent from 100-1000Hz. Also measured with a Digilent Analog Discovery 2 oscilloscope. I wasn't terminating with a 600ohm load for this measurement. That being said, it would appear that the Apollo X series in +24 mode indeed finds -20dBFS at +4 dBu. The Sumbus input VU LEDs are showing 0 VU (not measuring off of the summed output), but the Silver Bullet's VU LEDs are only illuminated up to -2 VU, so I would imagine the resolution/accuracy of the LED meters aren't really something I should be basing any sort conclusions off of. |
|
|
|
Post by matt@IAA on Feb 13, 2020 16:51:56 GMT -6
It would be kind of interesting to see where the actual light turns on for the 0VU for both based on the dBFS level. You might find a few tenths of a dB difference between them or something. I wouldn't really worry much, but I know some people are a lot more OCD than me.
|
|
|
|
Post by dror520 on Feb 14, 2020 3:07:10 GMT -6
I'm also very interested in this topic. I tried to check the levels with my Motu 828es interface which is supposed to be +20dBu max output level. So that means +4dBu should be -16dBFS, but for some reason this isn't the case. In order for my multimeter to show 1.23v I need to raise the level to -14.8dBFS. Is this normal? Is this an indication that the specs are "wrong"?
Thanks!
|
|
|
|
Post by bram on Feb 14, 2020 11:02:04 GMT -6
It would be kind of interesting to see where the actual light turns on for the 0VU for both based on the dBFS level. You might find a few tenths of a dB difference between them or something. I wouldn't really worry much, but I know some people are a lot more OCD than me. For Apollo x8/x16 at +4 dBu / +24 dB Headroom, the following is true: Silver Bullet Mix (Mojo Off): -18dBFS = 0 VU Silver Bullet TRK (Bypass) : -20dBFS = 0 VU CAPI Sumbus Inputs: : -20dBFS = 0 VU |
|
|
|
Post by bram on Feb 14, 2020 11:05:32 GMT -6
I'm also very interested in this topic. I tried to check the levels with my Motu 828es interface which is supposed to be +20dBu max output level. So that means +4dBu should be -16dBFS, but for some reason this isn't the case. In order for my multimeter to show 1.23v I need to raise the level to -14.8dBFS. Is this normal? Is this an indication that the specs are "wrong"? Thanks! Are you measuring tip to ring? What waveform and frequency are you sending the multimeter? Ex. If you're sending a 1khz sine, what happens to the measurement if you send it a lower frequency, say 400hz? |
|
|
|
Post by dror520 on Feb 14, 2020 15:45:24 GMT -6
I'm also very interested in this topic. I tried to check the levels with my Motu 828es interface which is supposed to be +20dBu max output level. So that means +4dBu should be -16dBFS, but for some reason this isn't the case. In order for my multimeter to show 1.23v I need to raise the level to -14.8dBFS. Is this normal? Is this an indication that the specs are "wrong"? Thanks! Are you measuring tip to ring? What waveform and frequency are you sending the multimeter? Ex. If you're sending a 1khz sine, what happens to the measurement if you send it a lower frequency, say 400hz? I am using a 1khz sine tone and checking from tip to ring. There's no difference between 400hz to 1khz on the multimeter. |
|
|
|
Post by christopher on Feb 21, 2020 11:44:06 GMT -6
Something else to consider when using signal generators inside the DAW is whether the generator's output is RMS or PEAK. For our uses we assume its RMS, but most things in the DAW are Peak. So that can mess with the measurements as well, something I always forget to check. So you may measure 1.23v RMS but.. is the plugin setting at -20 dBFS Peak or -20 dBFS RMS? I honestly don't know the answer. Might be less trouble to get a -18dB RMS reference track and monitor in RMS in the DAW while playing with the fader? I think I'll try that next time. Might as well try out a VU Meter plugin while I'm at it.
|
|
