K-Meters are now "Dynamic Range Meters"?

[tonAP]

148. Wave Candy: added new metering modes for loudness visualization.

Usually i don't go jumping from thread to thread asking the same question:
i've seen this news on Hosts forum so I've already asked there.
BUT, since it seems to me that's related to the general loudness-metering subject discussed here, i'd really like to know from the "Guru" about it.

Seriously.

[Compyfox]

Since I don't use or have FLStudio, I can't tell you much about FLStudio Wave Candy

But seeing from the specs of it, it can only measure peaks (volume) in dBFS, and not loudness (RMS). At least from this page. Sorry about that.

Else if you want to know how to setup a meter with RMS metering so that it responds to the K-System, there is a thread in Production Techniques where I gave a short overview (in PSP VU it is for example: 600ms Integr. Time if I'm not mistaken, 0VU Reference Level is the K-System aka -20, -14, -12, the rest of the settings are pretty much untouched).


EDIT 20mar2009 - 11pm:
Just had a longer email conversation with Mr Tischmeyer, CEO of the Pleasurize Music Foundation.

He listened to my ideas and said, that nothing is really set yet and there's still room for discussion, changes and the likes.

The team does neither believe that there is no room for improvement suggestions, nor are they locked to constructive critism and suggestions.

*hope I could reflect that properly in english*

I'd say this is a basis that somebody can work with and I definitely look forward to talk shop at Musikmesse 2009.


I guess the best support that we can give now is spreading the word while keeping an eye on this. Personally I'm not 100% convinced yet, but as somebody who adapted the K-System as way against the Loudness Race, I definitely support the initial idea.

[justified]

There's at least one (or more?) mp3 tags ('Relative volume adjustment'?) to apply a gain. Imagine most players out there supporting this, they would by default output music at, say -6dB or -12dB, unless such a tag is present.

So on your side, you'd produce mp3's with a volume adjustment of +12dB, the player would play it this time with a 0dB gain (-12+12), you'd get yourself a 12dB headroom compared to other, compressed tracks, you will then need less compression.

While this is still unrealistic (no one will probably ever agree on even little things like this), it's still more realistic than expecting to educate listeners.
Another problem is that engineers won't behave, they will add the +12dB tag, and still compress, but this time it won't just sound louder, it will also damage the listener's ears who will get a sudden +12dB compressed noise in his ears - don't forget that the ceiling also acts as a protection.
So to be safe it could be a lower +6dB headroom, with soft-saturation involved maybe.

There is "replaygain" mp3/ogg/flac/etc tag and, yes, it can be faked like this.

Another much more realistic solution: all players coming with a limiter/expander (very long release, simple auto-gain), but instead of being based on peak or RMS as usual, it'd need to be based on a good, reliable perceived loudness metering (if one even exists - this would be the 'invention').
Autogains already help a lot, especially for movies, but they really need to be based on better metering.

Probably, the best, though unrealistic solution would be to have an option on players to scan album or track and calculate replaygain info before playing it for the first time. Thus, fake loudness info avoided, and all music sounds with same percieved level.

[fanfarecircuit]

interesting. rem's orange crush from original green cd measures a 13, from the best of remaster measures a 7. but why? just because they want to suck? i just don't get this. i'm not really hearing any transient problems. am i deaf?

btw... the remaster rms registers 0 on a k12 scale. i used barricade pro, but i don't really understand the weightings yet... though i just read a short document from bob katz, so i'm beginning to understand.

basically, k12 sucks?

or since k12 is meant for broadcasting, they mastered to broadcasting levels so that no additional dynamic processing would occur during the broadcast?

[eduardo_b]

interesting. rem's orange crush from original green cd measures a 13, from the best of remaster measures a 7. but why? just because they want to suck? i just don't get this. i'm not really hearing any transient problems. am i deaf?

I think you're likely to notice more volume, less dynamic range before any transient issues. The band my have also set a limit on limiting, and the record company said okay because REM makes them major amounts of money.

[fanfarecircuit]

interesting. rem's orange crush from original green cd measures a 13, from the best of remaster measures a 7. but why? just because they want to suck? i just don't get this. i'm not really hearing any transient problems. am i deaf?

I think you're likely to notice more volume, less dynamic range before any transient issues. The band my have also set a limit on limiting, and the record company said okay because REM makes them major amounts of money.

i edited my comment and included a question on mastering to broadcast levels. is this to eliminate or at least limit the possibility of additional dynamic processing at the broadcast stage? would that enter the equation?

[eduardo_b]

interesting. rem's orange crush from original green cd measures a 13, from the best of remaster measures a 7. but why? just because they want to suck? i just don't get this. i'm not really hearing any transient problems. am i deaf?

I think you're likely to notice more volume, less dynamic range before any transient issues. The band my have also set a limit on limiting, and the record company said okay because REM makes them major amounts of money.

i edited my comment and included a question on mastering to broadcast levels. is this to eliminate or at least limit the possibility of additional dynamic processing at the broadcast stage? would that enter the equation?

It depends. In theory, they could determine the rms of the music and then apply additional compression or not depending on the reading. The compressor could be set so it won't trigger high rms content, I suppose. If done correctly, the limiting applied to everything going out would prevent listeners from having to ride gain on the volume control from one song to the next.

[Compyfox]

K-12 doesn't suck at all. It is more quiet compared to modern music indeed, but you have best of both worlds: Dynamic Mix, enough loudness to be heard (which is subjective).

Personally I use K-12 all the time, but with the twist that I use up the full amber zone (the +4dB over -12dB as VU/RMS Reference 0-point).

Since I digged a bit into this DR-Meter however, I found out that K-14 and K-12 Mixes are still hella quiet, and they will remain like that. But you can have a full desired dynamic.

At the moment this system is still a bit confusing, and I didn't get suitable results personally. But chances are that everything will be set and pinned down by the end of the year.

For now:
K-14 (including the amber zone) is okay, but quiet.
K-12 (including the amber zone, but not "abusing" the amber zone) hot enough, but no competition so far.

If you really can't live with the !quiet" recording, go as high as K-12 and use up the amber zone whole not overshooting it (aka having -8dB RMS max) - however this is NOT practicable, but still a compromise to most radio stuff.


You material should not be altered after that. But it's still dynamic enough to not be squashed via a radio station.


EDIT:
Sloppy english lately... Sorry about that. Maybe it's time for a break after all...

[camsr]

9dB RMS of dynamic range is good enough for most music. Most playback systems including the iTunes have their own maximization features so if your listener wants it loud they can have it loud as well as correct.

[Compyfox]

I do agree here. A good compromise indeed.
(K-12 with amber zone used up to +3dB for mezzoforte passages)

Even though I do prefer K-12 AZ+2 as of late.

[tonAP]

*IF* there is (also) interest in loudness metering in general (is there?), then the follow could be of some interest. So i'll jump another one:

148. Wave Candy: added new metering modes for loudness visualization.

Well i've got the answer:

It adds several loudness measurements systems, including the only one I believe to be working more or less, "Leq(A)"

Do you know what this stuff is about? Leq(A), Leq(RLB2), ITU BS.1770?
It's all about years of serious analysis, hard stuff in search of the best algorithms for the more effective loudness measurement.

BUT, since the whole thread is *just* all about K-meter..
then sorry: i'm obviously running OT.

[Compyfox]

It's not just "all about" - it's part of it.


Speaking of which... I did some listening tests again, and it looks like it that GR-14 is really for stuff from the 70ies. Most Pop Productions I've heard are indeed DR8 to DR10. Alone this meter is just confusing along with the planned volume correction, since an arrangement dynamic can differ greatly in terms of production.

So chances are that even if your track is K-12 (which is loud), but you want to keep a huge Dynamic Range (DR14), stuff must be uber-clean, the stereo space cleaned up, nothing distorted or bass heavy and the vocals 3 times as loud as the arrangement. In short: back to the 70ies.

I can live with K-12 AZ+2 and DR-10 however. But, according to specs, this will be adjusted "down" by 4dB again to be as loud as DR14 - which is not really possible.

IMO this system is actually good for producing and postproduction, while mastering you can use the K-System or meters that are calibrated for cinemas. It defintiely needs to work hand in hand with something, else it's simply driving me nuts.


Still a lot to learn, still a lot to try out. Hope I'll get some "aha" moments while being at the fair in a couple of weeks.

Oh and so far, I didn't encounter a DR3 or DR4 mix btw. Mostly DR5 and DR6, which would be K-5 if that would even exist.

[Cordelia]

Off the top of my head I can see so many problems here.

CDs are not the distribution system of the future, so a label on artwork is an idea from the past, most mp3s don't come with artwork.

Is this system going to be applied on a song by song basis, or will one DR number be applied to an entire CD? I don't think I need to spell out the trouble with trying to apply one DR number, and most disturbingly a final gain adjustment, to a group of songs that may not share the same dynamic range. If applied on a song by song basis, then the artistic decisions about how one song on a CD will flow into the next will now be out of the artist's hands.

As someone pointed out earlier, this will break "compatibility" with current and older CD releases, which will be much louder than the DR releases. Listening to a mix of the two might be a surprising and painful headphone experience.

And on a personal level, although I am against "loud for the sake of it" just as much as anyone, I wonder about those who have appointed themselves the deciders of "proper" dynamic range in all music. What is the proper dynamic range for an album like Metal Machine Music, for instance?

As a guideline, as an educational tool, I think this system is OK, if flawed. As a mandate...

[tonAP]

Is this system going to be applied on a song by song basis, or will one DR number be applied to an entire CD?

The "Discriminative Rating" will be applied to your whole CD

[fanfarecircuit]

guys, sorry... aes17 for k metering or not?