TBProAudio releases AB_LM - Loudness Match and Gain Staging Plugin for Windows and Mac OS X

[Compyfox]

One very essential question:

AB_LM offers following features:

• 4 measurement modes: RMS, RMS +3dB, EBU R128 SL, EBU R128 ML

Does the RMS measurement use a weighting filter and L+R summing like your dpMeter again?

If so, I consider this a non-ideal use, as this is not according to RMS (raw loudness) specs.




And plugins like this feel like my toenails being ripped out. I mean... is the future really "individual track Gain Staging according to EBU R-128 MLk"? Who started this stupid trend anyway?! That's like the third ITU-R type meterng tool I see that offers "automatic gain settings" according to SLk or MLk, and is being advertised as "per-channel gain staging" and "setup your project accordingly".

Is this the new "you have to use Meter X in order to do it right" as it happened with the K-System Meter (during mixing! ) and the DR-Meter (also during mixing...)?


Not that the EBU R-128 specs can't be used used for anything else other than broadcast levels, but this is taking it to extremes.

[dusted william]

does this do the same thing as Ian Shepard's perception?

http://www.meterplugs.com/perception

[TB-ProAudio]

ok just to make sure we are on the same page here, i mean for the second instance of your plugin (ie the receiver) to have option of reporting the latency tested to the daw (which would be the latency of the au plugin for example, in the middle place between the two ab_lm)

http://www.voxengo.com/product/latencydelay/

that's the way to get around it for now, what i would do is add this after ab_lm and report the figure you have given. But it adds also an extra 200ms latency of it's own, so the voxengo is not ideal. You have the infastructure already built in to record the latency, just an option to report it would be amazing.
yes please do consider it, i don't think any customer could disagree that's a bad optional feature  

Yes, i got your point. But i think there is more to consider: A plugin with a certain internal delay could report any PDC value (higher, equal or lower) to DAW. Important is then the delta which could be either positive or negative. But negative PDC value are not allowed.
Tricky situation!
I have to shake my brain how to solve this...

PS i am confused about euphonia. do you want to discuss my questions here or another topic?

I think this is the right place: TBProAudio: Euphonia, optimal audio spectrum balancing plugin released

[ObsoleteAcc99]

One very essential question:

AB_LM offers following features:

• 4 measurement modes: RMS, RMS +3dB, EBU R128 SL, EBU R128 ML

Does the RMS measurement use a weighting filter and L+R summing like your dpMeter again?

If so, I consider this a non-ideal use, as this is not according to RMS (raw loudness) specs.




And plugins like this feel like my toenails being ripped out. I mean... is the future really "individual track Gain Staging according to EBU R-128 MLk"? Who started this stupid trend anyway?! That's like the third ITU-R type meterng tool I see that offers "automatic gain settings" according to SLk or MLk, and is being advertised as "per-channel gain staging" and "setup your project accordingly".

Is this the new "you have to use Meter X in order to do it right" as it happened with the K-System Meter (during mixing! ) and the DR-Meter (also during mixing...)?


Not that the EBU R-128 specs can't be used used for anything else other than broadcast levels, but this is taking it to extremes.

Roland, instead of saying such awful things, why not put your questions nicely? Toenails being ripped out?

I'm not even using it for that. I am using it to monitor my plugin effects with exact same volume a/b ing and testing latency of plugins. Why not look at the things it does as was designed?

I am sure if there are actual problems with it, that the developer will be receptive if it is asked normally.

[Compyfox]

Oh, so testing the other meter relevant tools and then asking if AB_LM has the same flaw in design is equal to saying awful things now?

You should know me by now, Theo. I do take a closer look under the hood if we talk metering tools, and you also know why this is important to me. I did get in touch with the developer in private a couple of weeks ago since I was in need of a simple RMS Bargraph, and he could not tell me in a reasonable way why his dpMeter uses a weighted RMS analysis instead of z-weighting (aka: unweighted) - like it's the specs of a RMS meter.


Using EBU R-128 metering tools for "gain staging" is just a no-no if we talk individual channels. 99% of the tools on the market that use reference levels are setup to 0VU = -xyz reference level values, and once the program signal kicks in (read: non 1kHz sine signal), the meter is offset if we talk weighting filter and different ballistics than a basic VU. If we talk mix downs as preparation for CD authoring - that's a different thing. That the ITU-R BS.1770-x specs would also be better suited for the task, is also just a minor detail (in order to be EBU R-128, we talk about "fixed" reference level of -23LUFS, and specs according to ITU-R BS.1770-2 // most recent is BS.1770-3 however - and this also allows for custom color codes and custom reference levels -- but, details!).

This however is also my main issue on a lot of audio related web-boards - there is Metering Tool X, it's new, it's advertised with buzzwords - suddenly it's being declared as the "only tool to use" for what it was never made for. I've read that over and over and over again. On the Reaper boards the "K-System v1" meters are considered a holy grail and used for both gain staging and mastering, on another music-making entry-level forum I read the same about EBU R-128 meters. It's driving me insane!


Yes, I am all for education - but I DO have a problem if the main feature set is for a metering engine, that is not made for this type of setup. This in turn results in spreading wrong/false information and people going ham on the forums if we talk "how to to xyz the right way". Which then also bleeds over to audio engineering schools and music magazines, which is then spread to the masses as defacto "fact"... you get the idea.





If the devs doesn't want to answer - fine with me.

But in all honesty - this whole EBU R-128 metering thing from a lot of devs recently, just look like riding the bandwagon and a possible quick cash in. Another dev recently confirmed me that as well - he clearly told me "I'm not skilled regarding metering tools". Wonderful, isn't it?

YMMV.

[TB-ProAudio]

One very essential question:

AB_LM offers following features:

• 4 measurement modes: RMS, RMS +3dB, EBU R128 SL, EBU R128 ML

Does the RMS measurement use a weighting filter and L+R summing like your dpMeter again?

AB_LM uses the same pre-filter and summing algos for ML/SL measurement as dpMeter, RMS measurement is without pre-filter.

If so, I consider this a non-ideal use, as this is not according to RMS (raw loudness) specs.

Hmm, i do not get your point: what's wrong with it?

[TB-ProAudio]

Oh, so testing the other meter relevant tools and then asking if AB_LM has the same flaw in design is equal to saying awful things now?

Sorry for insisting, but could you please be more specific about the "flaw" you are mentioning? Thank you!

he could not tell me in a reasonable way why his dpMeter uses a weighted RMS analysis instead of z-weighting (aka: unweighted) - like it's the specs of a RMS meter.

Hmm, sorry for any confusion. Just for clarification:
All TBProaudio products including dpMeter and AB_LM use pre-filter for EBU R128 measurement modes according to specification.
All RMS measurement modes are without pre-filter (unweighted)!

I hope it is now clear how we implement RMS and EBU-modes.

[Compyfox]

AB_LM uses the same pre-filter and summing algos for ML/SL measurement as dpMeter, RMS measurement is without pre-filter.

Hmm, sorry for any confusion. Just for clarification:
All TBProaudio products including dpMeter and AB_LM use pre-filter for EBU R128 measurement modes according to specification.
All RMS measurement modes are without pre-filter (unweighted)!

I hope it is now clear how we implement RMS and EBU-modes.

Okay, so RMS is "unweighted", but what about summing of the signal? I do remember from previews contacts, that you mentioned that in dpMeter it's "L+R Summed", which is not according to standard RMS meter specs.

If so, I consider this a non-ideal use, as this is not according to RMS (raw loudness) specs.

Hmm, i do not get your point: what's wrong with it?

Once more - RMS metering specs (similar to VU) are:
- unweighted
- channel individual
- time frame of 300ms

To my understanding with dpMeter, is that the channels are summed. Which in turn results in an offset (3dB higher!) of the the RMS meter.


THAT is my main concern.

And that the plugin is advertised to use "EBU R-128 specs" for individual signal gain staging (ITU-R BS.1770-x type/based-upon meters are NOT made for mixing!). Can we also please start to call the specs ITU-R BS.1770-x, since this is what R-128 based upon, not to mention that it does not(!) have a fixed reference level, specific set gate and color codes? Thanks.

[Lesha]

So, I was right when I said that dpMeter shows RMS level incorrectly?

http://www.kvraudio.com/forum/viewtopic ... 3#p6254593

http://www.kvraudio.com/forum/viewtopic ... 7#p6274777

[Compyfox]

If you look at this post:
http://www.kvraudio.com/forum/viewtopic ... 5#p6270035

Then I'd say "yes".


Your other RMS tests need to take two things into consideration:
1) the AES-17 +3dB offset
2) you used White Noise rather than a Sine Wave - but you basically tested the full spectrum rather than a single frequency, so it's fine

[Lesha]

Thanks

[TB-ProAudio]

Okay, so RMS is "unweighted", but what about summing of the signal? I do remember from previews contacts, that you mentioned that in dpMeter it's "L+R Summed", which is not according to standard RMS meter specs.

yes, dpMeter uses summed RMS value.

What's wrong with summed (L+R) RMS value? We and others use it every day in our music/speech productions.
BTW: ML/IS/IL in EBUR128 are also summed values!

[Compyfox]

Meter standards exist for a reason.

With EBU R-128, or ITU-R BS.1770-x for that matter, summing the signal is part of the specifications to compensate two things, while simplifying the readout ("mono" bargraph):
a) the weighting filter (k-weighting) - which in turn handles the flaw of VU/RMS meters responding too strong on low frequency content, and raising the dangerous-to-the-ear frequency spectrum from 1kHz and above pre metering
b) the 400ms time frame for MLk and 3s time frame for SLk


So what do you mean by "we and others use it every day"?

I do not consider myself as "we" or "the others" - unless you mean with "we" the Reaper boards (where I've also read the most horrendous things regarding metering tools). I use the ITU-R specs for mastering and CD authoring exclusively - actually with my own concept (K-System v2). Not for anything else. And this was/still is it's main purpose - checking broadcast loudness and loudness normalize accordingly.


The Dorrough Meter (40A and the digital successor 2x0 and 3x0, Dorrough were among the first to call this meter a "RMS bargraph meter") does NOT use summed analysis
A VU doesn't use summed analysis (IEC 60268-17) - unless you run a stereo signal through a "mono" needle meter
A PPM doesn't use summed analysis (IEC 60268-18 digital and IEC 60268-1 Type 1-4, also IEC 60268-17)

The K-System meter is based upon the Dorrough 40A (actually, it's 1:1 600ms analysis window, but a shifted reference level/scale)
DR-Meter give a readout of the "crest factor" (average signal strength to maximum signal strengh), also non-summed (especially in regular level meter mode)
LRA Meters give out the "range" between lowest average loudness to highest average loudness (see histogram feature in ITU-R Meters, or Shepherd's Dynameter)


The ITU-R BS.1770-x meters are based upon the "best of all worlds", and is a successor to LEQ(m) (who actually used summed analysis before it turned into ITU-R BS.1770-x)
The R-128 Meter is, like the ATSC/85 specs, just a fine tuned and fixed(!!!) setting based upon the ITU-R BS.1770-x specs. The main genealogy starts there.


So the only meter that is "summed" in terms of analysis, is the loudness analysis of the ITU-R BS.1770-x specs.
NOT RMS (read: raw loudness) meters.

There is not even a mention about it in the AES-17 white paper (not even the most recent version from 2015) which declares something else than the +3dB offset during calibration, and how to calibrate RMS meters in general. Your RMS meter(s - plural!), due to the fact that it sums the signal, is off by 3dB. The RMS+0 value reads +3dB higher, making it equal to the RMS+3 setting (aka AES-17). So 1kHz -18dBFS at RMS+0 should read -21dBFS, your meter reads -18dBFS. While 1kHz -18dBFS at RMS+3 should read -18dBFS, your meter reads out -15dBFS.

Your meter is/your meters are therefore not according to decade old specs - PERIOD!



And this is what pisses me off with so called "specialized metering tools" and their developers... they don't see anything wrong (especially those that just code, yet don't have a grasp on the material behind it). And as somebody that has this topic as pet peeve, I will call the developer in question out on that. Imagine building a bridge in imperial system, while others use the metric system. Is it still okay if the bridge was built with a 20cm offset? I don't think so (see "Mars Climate Orbiter/Mars Surveyor '98 Orbiter incident).

This is what I tried to tell you with my last mails (November 2015).




So once more:
Why do you use a "summed" signal for RMS measurement?
What do you expect to accomplish here?

You make an already flawed RMS meter (responding stronger to bass intensive material, has an offset depending on the "measurement time window") even more flawed. I really don't get it.

[ObsoleteAcc99]

does this do the same thing as Ian Shepard's perception?

http://www.meterplugs.com/perception

yep, i never knew what that plugin in did, but from what i can tell they are more or less identical except TB gives the actual latency figure rather than just internally measuring it. And more metering standards. So save yourself 100 bucks!

[ObsoleteAcc99]

Oh, so testing the other meter relevant tools and then asking if AB_LM has the same flaw in design is equal to saying awful things now?

You should know me by now, Theo. I do take a closer look under the hood if we talk metering tools, and you also know why this is important to me. I did get in touch with the developer in private a couple of weeks ago since I was in need of a simple RMS Bargraph, and he could not tell me in a reasonable way why his dpMeter uses a weighted RMS analysis instead of z-weighting (aka: unweighted) - like it's the specs of a RMS meter.


Using EBU R-128 metering tools for "gain staging" is just a no-no if we talk individual channels. 99% of the tools on the market that use reference levels are setup to 0VU = -xyz reference level values, and once the program signal kicks in (read: non 1kHz sine signal), the meter is offset if we talk weighting filter and different ballistics than a basic VU. If we talk mix downs as preparation for CD authoring - that's a different thing. That the ITU-R BS.1770-x specs would also be better suited for the task, is also just a minor detail (in order to be EBU R-128, we talk about "fixed" reference level of -23LUFS, and specs according to ITU-R BS.1770-2 // most recent is BS.1770-3 however - and this also allows for custom color codes and custom reference levels -- but, details!).

This however is also my main issue on a lot of audio related web-boards - there is Metering Tool X, it's new, it's advertised with buzzwords - suddenly it's being declared as the "only tool to use" for what it was never made for. I've read that over and over and over again. On the Reaper boards the "K-System v1" meters are considered a holy grail and used for both gain staging and mastering, on another music-making entry-level forum I read the same about EBU R-128 meters. It's driving me insane!


Yes, I am all for education - but I DO have a problem if the main feature set is for a metering engine, that is not made for this type of setup. This in turn results in spreading wrong/false information and people going ham on the forums if we talk "how to to xyz the right way". Which then also bleeds over to audio engineering schools and music magazines, which is then spread to the masses as defacto "fact"... you get the idea.





If the devs doesn't want to answer - fine with me.

But in all honesty - this whole EBU R-128 metering thing from a lot of devs recently, just look like riding the bandwagon and a possible quick cash in. Another dev recently confirmed me that as well - he clearly told me "I'm not skilled regarding metering tools". Wonderful, isn't it?

YMMV.

This plugin is basically used in mastering.. check out perception.. i definitely wouldn't be using it for gain staging the mix as such.. this plugin is about , um, level *perception* to make sure one doesn't make effects decisions on volume increase placebo.

I'm a bit concerned about you saying the RMS meter is wrong but that can be ignored i guess and one can use any other RMS meter in conjunction with the plugin. Is the ebu 128 on TB's plugin working according to specs? that's the most important one for mastering