K system metering system anyone?

Hi all
A friend of mine and I are designing a 4 ch preamp.
He kind of insist on having VU meters on the thing for each channel, I say, with digital recording systems more common nowadays, we need more of a PPM type of metering.
But then again, I say that we might not even need metering at all!

I use a RME firewire800 converter unit and Nuendo3 software. With this system I have plenty of meters, with quite a few choices of ballistic and resolution.

Some may argue, and I may concur, that getting the signal well controlled between boundaries right at the preamp is the way to go.
Hmmmm.... yeah, I guess...
But getting the proper ballistics from a mid range meter , like the Siffam, needs quite a bit of circuitry.

Anyway to cut the discussion short, I think nowadays, the best metering system in this age would be Bob Katz 'K system' metering.
I have an article on this somewhere, but while I dig for it, is someone have knowledge about this system and even, perhaps, a design for it?

If I remember well, mr Katz system specifies quite a different attack and release time for the ballistic but more importantly, it is referenced to DBFS scale.
Which makes a lot of sense....

Any comments, thoughts, diatribe, criticism or anything on the subject? :

Luc

blue_luke said:

Hi all
A friend of mine and I are designing a 4 ch preamp.
He kind of insist on having VU meters on the thing for each channel, I say, with digital recording systems more common nowadays, we need more of a PPM type of metering.

Click to expand...

I tend to concur with you. The presence of mechanical VU-meters on a large number of mic pres is comfort therapy IMO. I understand the beauty of a nice 65mm or larger well-illuminated meter, but it has no justification in a digital recording environment

But then again, I say that we might not even need metering at all!
I use a RME firewire800 converter unit and Nuendo3 software. With this system I have plenty of meters, with quite a few choices of ballistic and resolution.

Click to expand...

Not entirely true, because all parts of the chain do not clip at the same time. A dramatic scenario would be the mic pre clips/distorts at +12dBu, the subsequent compressor clips at +20 and the converter clips at +24. You would never know by the DAW meters that the mic pre is clipping/farting. So you have to calibrate and know the level diagram of the whole signal path. And listen!

Some may argue, and I may concur, that getting the signal well controlled between boundaries right at the preamp is the way to go.

Click to expand...

Not necessarily. Several mic pres have a gain trim that potentially could lead to the input stage distorting with an otherwise normal output level.  Most of the times, it is not a serious problem because metering is not used alone - it is a complement to auditive monitoring. There are so many possibilities of making an intermediary element of the signal path distort, that it would be necessary to meter all of them, which is impractical.
But I must say that I like having a meter at the output of my outboard pres, as much as I like having a PFL button in my mixers.

But getting the proper ballistics from a mid range meter , like the Siffam, needs quite a bit of circuitry.

Click to expand...

Then you may end up with a 10msec peak-meter, which is definitely not fast enough for proper level assessment.

Anyway to cut the discussion short, I think nowadays, the best metering system in this age would be Bob Katz 'K system' metering.
I have an article on this somewhere, but while I dig for it, is someone have knowledge about this system and even, perhaps, a design for it?

Click to expand...

The K system is a monitoring scheme where the monitors are calibrated to produce a pre-defined SPL when the electrical signal has a standardised value. The purpose of it is that the level at which the sound engineer is working is the same that the level in the theatre, home or whatever listening environment the final listener will be in.
It has no relevance at all for a mic pre.

If I remember well, mr Katz system specifies quite a different attack and release time for the ballistic but more importantly, it is referenced to DBFS scale.

Click to expand...

This is true; there is a peak-meter with 0 integration time (one-sample of 0dBfs is displayed as 100%) and an average meter (similar to a VU-meter) which represents the subjective loudness. One can argue that an RMS meter would be more representative of the loudness, but in fact ther is not much difference. A Zwicker meter would be more appropriate, but its real-time implementation requires huge processing power.

Bob's metering system involves more than just a standardized monitor SPL. He also suggests a different headroom or nominal 0VU level re: full scale, for different output media.

Using a microprocessor based metering system one could probably handle "loudness weighting" (Zwicker meter?), if it is really desirable. To be honest I have never considered it. I have actually coded up real time RMS computation for a Peak/VU meter I did, but removed it again when I didn't see a visible difference (from simple average) on complex waveforms. A level dependent response weighting would involve a little more number crunching but is probably doable. 

Bob has posted here from time to time, so maybe he can add to this discussion. 

I will cast my vote for both PK and average/VU data, but if I had to pick only one for today's environment I'd choose peak, to "see" clipping and worry about loudness in the mix by listening.

JR

Yeah my understanding of the K system was that it was mainly a mastering application.  This would help balance the perceived loudness of different things with each other.  If you put up an uncompressed source, say something symphonic performance of an average volume throughout, next to a peak limited to hell modern rock mix, level match them, the "LOUD" modern rock mix begins to sound weak next to the other.  So if everybody mastered their work using the k scale, then all media would fit better together without making one stand out or stand down to other sources.  It seems to me kinda like an attempt at creating a mastering level standard so everything works together as opposed to just letting the producer tell the mastering engineer "Louder please!  Make it louder!  Squeeze all the dynamics out of it so its freakin loud!!!!!"  An attempt at regaining dynamic content that was lost thanks to the loudness wars.

So yeah, I don't really see an application for this on a mic pre.....

Almost everything there is to know about the K system is there:
http://www.aes.org/technical/documentDownloads.cfm?docID=65
As phishman wrote, it is strictly for mastering purposes. The different K scales are aimed at different reproduction systems, based on estimated max SPL capacity, estimated nouse level, media dynamic range and listener's expectations.
Check also the TT Dynamic Range website
http://www.dynamicrange.de/fr/fr/meter-spex
Their concern is domestic and professional playback; they seem to advocate a minimum 12-14dB headroom between RMS and peak, purely on a subjective point of view.
None of these are relevant to electrical/digital operating level in the recording chain.

This makes an interesting reading on the subject.

First a 'mea culpa' I suggested mr Katz 'K system' metering without really knowing what it is all about!!

(Won't do that again... blush!)

And after a bit of thinking, D'enfer is right when he says that we want to avoid overload at any stage.
It makes sense after all, once a signal is clipped, it stays clipped forever in the rest of the chain.
So We will probably end up with something like an LED that blinks at clip level, on each stage.

But I still feel we do not need meters directly on the preamp, after all, many good vintage preamps and broadcast equipment had very little in terms of metering and no one complained.

Luc

One thing I've done on a few consoles that I though was useful was using bi-color LEDs. Green for signal present and red for several dB before clipping. In a console just that information can cover a lot of what people use meters for. If you are putting multiple channels in one box signal present could be useful.

Another common practice..  you can diode "or" together multiple circuit nodes to sample for clipping, so you don't need multiple clip LEDs per channel, even if there are multiple gain stages that could clip. You can always figure out which stage is clipping by backing off later stage gain or EQ boost.

Note: for very tweaky ultra low distortion design, be aware that the diode sampling presents a non linear load so can actually distort high source impedance circuit nodes. If a cap hold is being charged directly from the sampling diode this can transiently distort waveforms.  Also if charging a hold cap, be sure the ground path for that cap does not corrupt a clean signal ground.

So first do no harm.. while trying to prevent harm.

JR

Thanks John for the advices!
I know quite a bit about metering, but just did not study enough about the K system.

I have designed a bicolour led many years ago and found that it made a small tick every time it fired.
That little circuit taught me to treat with much greater respect the layout of ground traces and cabling in a console!
Green for -20, orange (both green/red) for -6 and red for +4.
In the end I did not like it, the difference in colour between red and orange was correct, but in practice, it proved to be too subtle.
Finally, i opted for just a red LED that flashes more rapidly, the more you are heavily overloading.
That went right.
To completely eliminate the tickings, I had to use a +/- 15V supply rail and that is something I want to avoid in this new design, a two stage tube preamp that already needs +HV, heaters, phantom and a logic supply rail for all the relays...
So probably just an led from the logic rail will do, but one for each stage.

Yes, that's pretty much my experience too.. the both colors on is not very useful for working with dynamic music signals.

I have used the both color LED mode effectively for a status indicator on a product with three modes A, B or A+B, where it logically and visually made sense, but the red and green on does not compute as less than red, for a level indicator, IMO. 

Yes, laying out a console will give you new respect for the details of where PS and ground currents go..  Hardest simple circuits to execute.  ;D

I have seen consoles dedicate a rail just to LEDs.. I like to power them between the +/- 15v rails, but for two-lead bi-color LEDs, you often need to drive them + and - re a middle V.  The SMT bi-colors I am playing with now bring out all 4 leads so are easier to manage, but I'm not as involved with consoles these days. Aren't they obsolete yet? . 

JR

JohnRoberts said:

but I'm not as involved with consoles these days. Aren't they obsolete yet? . 

JR

Click to expand...

That's a whole new subject indeed!!
I am an old school engineer (57yo) I like faders and buttons way more than mouse clicks an sub-menus!
4,000 buttons+ and 48meters is 'natural' for me! 3 level of sub-menus is VERY annoying!!  ;D
But that's only me! 
Luc