What's the new -9 rms under LUFS?

[dblock]

I know -23LUFS is for broadcast compliance but I'm not concerned whether or not my mix is turned down at that stage or not.
I just want to know what, for example,my target rms of say -8 would translate to what number on my LUFS meter so the perceived loudness is still the same if that makes sense. I just need a new numbers to shoot for under the LUFS system that's equivalent to the rms numbers I used to shoot for if that makes sense.
tia

[Keith99]

Not sure there is an easy correlation. LUFS measures average loudness but RMS measures average signal level

[dblock]

Not sure there is an easy correlation. LUFS measures average loudness but RMS measures average signal level

Thanks for responding Keith. I admit I make be asking an apples & oranges question or the right question in the wrong way but I couldn't come up with a better way to put it. Just thought I'd throw it out there.didn't know how else to put it. Maybe somebody with mental telepathy can figure out what I'm trying to ask.

[Keith99]

I must admit I dont know a lot about it myself, Compyfox is the person to ask. You could perhaps test it. Play some pink noise and set it so on a RMS meter it is -8 and see what a LUFS meter says. Of course that would not be completely accurate as it is just a continuous tone but may give some guide

[dblock]

I must admit I dont know a lot about it myself, Compyfox is the person to ask. You could perhaps test it. Play some pink noise and set it so on a RMS meter it is -8 and see what a LUFS meter says. Of course that would not be completely accurate as it is just a continuous tone but may give some guide

I was thinking about doing something like that. My lazy side was hoping there was a simple calculation like," -9rms = is equal to -12 lufs or something but nooooo. couldn't be that lucky.
Thanks

[dblock]

Not sure there is an easy correlation. LUFS measures average loudness but RMS measures average signal level

Hey Keith, after some googling, I found that Keith99 is exactly right......
"...RMS is only a measurement (or display) of signal voltage, so it doesn't really give us an idea of perceived loudness. Two music tracks measuring the same RMS values may not necessarily have the same perceived loudness because RMS does not take into account the psychoacoustic nature of apparent loudness as heard by the human ear, specifically that low, mid and high frequencies of the same level are not perceived as being the same loudness.
The Integrated loudness measurement specifically takes into account this aspect of human hearing perception of loudness and adjusts accordingly....." Interesting read from Ask Audio.

Guess I was comparing talking it was apples and oranges after all.
Thanks for the input guys!

[woodsdenis]

Not sure there is an easy correlation. LUFS measures average loudness but RMS measures average signal level

Hey Keith, after some googling, I found that Keith99 is exactly right......
"...RMS is only a measurement (or display) of signal voltage, so it doesn't really give us an idea of perceived loudness. Two music tracks measuring the same RMS values may not necessarily have the same perceived loudness because RMS does not take into account the psychoacoustic nature of apparent loudness as heard by the human ear, specifically that low, mid and high frequencies of the same level are not perceived as being the same loudness.
The Integrated loudness measurement specifically takes into account this aspect of human hearing perception of loudness and adjusts accordingly....." Interesting read from Ask Audio.

Guess I was comparing talking it was apples and oranges after all.
Thanks for the input guys!

I think I remember Ian Shepard (mastering) saying that they were roughly equvalent in their basic form, we arent talking about average program loudness but momentary in this case.

[Compyfox]

I must admit I dont know a lot about it myself, Compyfox is the person to ask.

Did somebody say my name?

I think I remember Ian Shepard (mastering) saying that they were roughly equvalent in their basic form, we arent talking about average program loudness but momentary in this case.

I don't think he said that. Because if he did - then I'd personally not trust in his tools anymore.

You could perhaps test it. Play some pink noise and set it so on a RMS meter it is -8 and see what a LUFS meter says. Of course that would not be completely accurate as it is just a continuous tone but may give some guide

Now here is the thing:

LUFS means "Loudness Unit Full Scale"
RMS means "root mean square"

Both measure the so called "average signal strength", that some call "loudness", but both work fundamentally different. At least if we talk "pure measurement". In usability, they're indeed "similar" (if we ignore the age of the metering tools and their general tasks - they can always be fine tuned, and used for "more" than just their basic function - if you know how)

• RMS is measured "unweighted", meaning: flat frequency response.
• LUFS uses a "weighting filter", which filters out the extreme low frequencies and puts an emphasis (boost) on everything above 1kHz. Therefore if you send in a Pink Noise signal, it should be offset by a couple of dB. Since RMS meters respond stronger to bass heavy material, and the LUFS meter does not (compensated!), the readout is a bit lower on a static signal.

• RMS uses "common known" time frames (it's a really complicated topic that even I don't fully understand) of 300 or even 600ms (the Dorrough meters first coined the term "RMS meter")
• LUFS uses 400ms (called Momentary) and 3s (called Short Term) for the realtime meters, and an average value (declared every 2-5s) from the whole measured stream so far (really complicated process).

Another important thing:

• RMS measurement exists in two "flavors": RMS, and RMS+3 (AES-17 compensation). These days, the AES-17 compensation recommendation is in use. So a Sine Wave at - xyz dBFS reads out both the same vale on max digital peak, and RMS. Else the RMS value would be 3dB lower. Also, each channel responds independent.
• LUFS "sums" the signals (which automatically covers the +3dB AES-17 compensation), which on the long run is easier for loudness measurement for multi-channel program streams (read: surround)

Bonus - the "+9dB range on top of the average LUFS value" you might have probably heard about, is a safety mechanism for the MLk meter for noise bursts and the likes (forte fortissimo passages). Important for broadcast is the ILk value (Integrated Loudness), for CD releases I do recommend to take a look at the SLk value (see below). But you want to compare an RMS meter with an LUFS meter, so we look at the MLk meter (Momentary Loudness) for the time being.

Okay, now that we got this out of the door, we can look at the meters themselves, and what is roughly what equivalent.




The following is now directly aimed at dblock:

Since I don't know if the "magical" -8dB RMS is maximum or average, Let us assume that you shoot for your desired -8dB RMS avg. Meaning about -5dB RMS absolute max, which tends to be the limit for MOST modern productions these days - and also results in a "squarewave" or "head cheese" (however you want to call it).

Now let us also assume that you used the Dorrough 40A meter to measure that value (which was the first "known"/declared RMS meter), we also know that the response time of the meter was 600ms (can be found in the patent of this meter)

LUFS only has 400ms or 3s. I assume you want to get as close to the RMS meter as possible, so... 400ms it is. Keeping in mind that there is an offset of the signal readout due to the weighting filter, in order to shoot for about -8dB RMS avg, I'd shoot for -12LUFS to -10LUFS MLk (Momentary Loudness). (offset is about 3-4dB!)



HOWEVER... if you're not into the Loudness War (anymore), then use the SLk meter instead (which is way more relaxed in terms of the analysis) and shoot for -12LUFS avg, with -8LUFS max. This is way more musical oriented and currently shoots at "Loudness Normalization" on Streaming Services like Youtube (which hovers between -16LUFS and -12LUFS SLk), while it's still loud enough for CD releases without loosing too much of the desired transients.


As usual, my KVRmarks are your friend. But I hope this answers your initial question.

[dblock]

I must admit I dont know a lot about it myself, Compyfox is the person to ask.

Did somebody say my name?

I think I remember Ian Shepard (mastering) saying that they were roughly equvalent in their basic form, we arent talking about average program loudness but momentary in this case.

I don't think he said that. Because if he did - then I'd personally not trust in his tools anymore.

You could perhaps test it. Play some pink noise and set it so on a RMS meter it is -8 and see what a LUFS meter says. Of course that would not be completely accurate as it is just a continuous tone but may give some guide

Now here is the thing:

LUFS means "Loudness Unit Full Scale
RMS means "root mean square"

Both measure the so called "average signal strength", that some call "loudness", but both work fundamentally different. At least if we talk "pure measurement". In usability, they're indeed "similar" (if we ignore the age of the metering tools and their general tasks - they can always be fine tuned, and used for "more" than just their basic function - if you know how)

• RMS is measured "unweighted", meaning: flat frequency response.
• LUFS uses a "weighting filter", which filters out the extreme low frequencies and puts an emphasis (boost) on everything above 1kHz. Therefore if you send in a Pink Noise signal, it should be offset by a couple of dB. Since RMS meters respond lower to bass heavy material, and the LUFS meter does not (compensated!), the readout is a bit lower on a static signal.

• RMS uses "common known" time frames (it's a really complicated topic that even I don't fully understand) of 300 or even 600ms (the Dorrough meters first coined the term "RMS meter")
• LUFS uses 400ms (called Momentary) and 3s (called Short Term) for the realtime meters, and an average value (declared every 2-5s) from the whole measured stream so far (really complicated process).

Another important thing:

• RMS measurement exists in two "flavors": RMS, and RMS+3 (AES-17 compensation). These days, the AES-17 compensation recommendation is in use. So a Sine Wave at - xyz dBFS reads out both the same vale on max digital peak, and RMS. Else the RMS value would be 3dB lower. Also, each channel responds independent.
• LUFS "sums" the signals (which automatically covers the +3 compensation)

Bonus - the "+9dB range on top of the average LUFS value" you might have probably heard about, is a safety mechanism for the MLk meter for noise bursts and the likes. Important for broadcast is the ILk value (Integrated Loudness), for CD releases I do recommend to take a look at the SLk value (see below). But you want to compare an RMS meter with an LUFS meter, so we look at the MLk meter (Momentary Loudness) for the time being.

Okay, now that we got this out of the door, we can look at the meters themselves, and what is roughly what equivalent.




The following is now directly aimed at dblock:

Since I don't know if the "magical" -8dB RMS is maximum or average, Let us assume that you shoot for your desired -8dB RMS avg. Meaning about -5dB RMS absolute max, which tends to be the limit for MOST modern productions these days - and also results in a "squarewave" or "head cheese" (however you want to call it).

Now let us also assume that you used the Dorrough 40A meter to measure that value (which was the first "known"/declared RMS meter), we also know that the response time of the meter was 600ms

LUFS only has 400ms or 3s. I assume you want to get as close to the RMS meter as possible, so... 400ms it is. Keeping in mind that there is an offset of the signal readout due to the weighting filter, in order to shoot for about -8dB RMS avg, I'd shoot for -12LUFS to -10LUFS MLk (Momentary Loudness). (offset is about 3-4dB!)



HOWEVER... if you're not into the Loudness War (anymore), then use the SLk meter instead (which is way more relaxed in terms of the analysis) and shoot for -12LUFS avg, with -8LUFS max. This is way more musical oriented and currently shoots at "Loudness Normalization" on Streaming Services like Youtube (which hovers between -16LUFS and -12LUFS SLk), while it's still loud enough for CD releases without loosing too much of the desired transients.


As usual, my KVRmarks are your friend. But I hope this answers your initial question.

Geez! That was some heavy but extremely useful stuff you laid out there.
I'll have to bookmark it and read it about 10-15 times to let it soak in
but you definitely understood my question and you answered it.
Much appreciate it Compyfox!
and thanks woodsdenis for rubbing the genie bottle to get Compyfox to come out.

[Compyfox]

Glad I could help. Tried to simplify it as much as possible.

I might sometimes shake my head on threads and move on, say "let others have it", I might respond super late on a topic. But I do have an eye on threads that mention metering tools...


EDIT:
Updated my last post - had a slight mistake there

[dblock]

Glad I could help. Tried to simplify it as much as possible.

I might sometimes shake my head on threads and move on, say "let others have it", I might respond super late on a topic. But I do have an eye on threads that mention metering tools...


EDIT:
Updated my last post - had a slight mistake there

Just bought the Klangfreund's LUFS Meter. Love this thing.Very unique
and useful.

[Compyfox]

General warning:
ITU-R BS.1770-x meters (or it's "presets" like EBU R-128) are not made for "initial gain staging" (project setups). This is a bonus feature this thing has - I wouldn't use it.

For (pre)mastering, any decent ITU-R spec meter is fine these days.