Dynamic range true value

[plexuss]

MAAT uses R128 they say in their product into. if you want to get into the weeds, DR for audio should use PSR not PLR - Youlean from what I recall is only PLR so may not be the best tool for DR. but the OP has MAAT so thats all he needs. Personally I use Nugen Mastercheck Pro.

[mutantdog]

What's the difference between PSR and PLR?

[plexuss]

What's the difference between PSR and PLR?

PLR is a long-term (enire song) peak-loudness-ratio were as PSR is a short-term (3 second).

[mutantdog]

That makes sense.

So surely in Youlean, PLR-I corresponds to long-term while PLR-S is short-term..?

[plexuss]

That makes sense.

So surely in Youlean, PLR-I corresponds to long-term while PLR-S is short-term..?

You'd have to ask Youlean to explain the specifics of his meter, although that would be a good guess.

[Googly Smythe]

Loudness metering includes a Loudness Range component. All the meters I've seen have this component, including the Youlean.
This (I believe) corresponds to dynamic range...

https://www.pro-tools-expert.com/home-p ... ness-range

And where is Compyfox when you need him!

[mutantdog]

Not quite, the loudness range is more closely related to the difference between loud and quiet parts of a track (eg: chorus and breakdown). Whereas the dynamic range specifically addresses the difference between an averaged loudness value against the instantaneous peaks.

So if you take a track with a loudness range of, say, 10 and push it fairly hard into a master limiter the loudness range will stay roughly the same but the dynamic range will be reduced relative to the amount of gain increase.

[Doctor Doubledrop]

Happy to hear lot of interesting things.. "Nugen Mastercheck Pro", nice!
"protul expert tutorial! interesting too, let's download this Youlean ok.

DR for audio should use PSR not PLR

, not the opposite? I see PLR in the nugen audio analyzer.

So if you take a track with a loudness range of, say, 10 and push it fairly hard into a master limiter the loudness range will stay roughly the same but the dynamic range will be reduced relative to the amount of gain increase.

I'm not sure, let me try

[plexuss]

Happy to hear lot of interesting things.. "Nugen Mastercheck Pro", nice!
"protul expert tutorial! interesting too, let's download this Youlean ok.

DR for audio should use PSR not PLR

, not the opposite? I see PLR in the nugen audio analyzer

PSR

[Compyfox]

And where is Compyfox when you need him!

Working... away from KVR. (no joke, I'm quite busy as of late! And I have no headspace for KVR right now)

I saw this thread out of pure chance. And you folks got the most basic info covered. There is technically no need to repeat things. But I do hope that my addition to the thread clears the remaining questions.




To answer a couple of things:

The DR Meter was originally created by the "Pleasurize Music Foundation" (PLM in short), which Friedemann Tischmeyer lead. The original version was created by (according to conversations I had throughout the years) "Algorithmix", and the later OS ports were then handled by Brainworx. The PMF always had this tool as part of a bigger "campaign" to spread "awareness regarding loudness and better dynamic range". It did work initially - but rather of giving that meter away for free, it was more and more locked away behind a paywall until even old users couldn't access it anymore without investing a yearly fee.

After plenty of request, Brainworx set the meter "free" (aka: re-released it), added a ton of FR's and called it bx_meter. Years later, Tischmeyer teamed up with old fellows from Algorithmix days, revived the plugin again under "MAAT Digital" - and upped the price in the process, not to mention tm'd certain design features.

So basically, Tischmeyer helped create a meter that showed the "range between the average signal strength and the maximum signal strength" (or Crest Factor in order words) in a more snazzy realtime fashion, then called that thing "Dynamic Range Meter". Since he was the first to do so, and some of it is now trademarked, of course it is "standard". But it is not a widespread "standard" (as in ISO, etc), but a recommendation. I've lost track if there are more current white papers on this - but that's the basic gist of it.




Enter the different versions:

TT DR-Meter vs bx_meter vs MAAT DRMeter v1 vs MAAT DRMeter v2

The first three use pretty much the same(!) time windows. Plus minus xyz ms overlap for the mathematics. Which is the man crux of realtime RMS meters. Some use 250ms, 300ms, some 400ms, some 600ms, others 1000ms. Inertia also plays a huge role into the offsets. If that's not documented, you're a bit out of luck. There were huge debates throughout the years as well, as to which "ballistics" (inertia/rise/fall. however you want to call it) are the correct ones.

Rest assured, the original TT DR-Meter had no K-System ballistics (based upon the Dorrough 40A patent, so 600ms rise/fall), bx_meter included that with the K-System presets, MAAT DRMeter looks like a slightly refined version of TT DR-Meter (so the ballistics could have changed to 400ms - the now considered "go to" for "momentary" loudness; plus the inertia for the realtime DR measurement).

MAAT DRMeter v2 uses the ITU-R BS.1770-x specs (EBU R128 and ATSC A/85 are offsprings of the very same specification!). But compared to the "summed to mono specs" from ITU-R BS.1770-x (standard!), the meter also uses "measurement per channel" to be "more oldskool" (their words!) and have a better overview of what's happening "in the stereo field".

Personally I have a huge problem with the DR Meter "concept" these days, since to me(!!!), MAAT Digital rides on a very old and "gone" bandwagon and actively confuses users with that approach instead of sticking to modern day implemented standards/specs (ITU-R BS.1770-x). I mean, we've been fighting for over a decade to get rid of peak limiting, and here is DR Meter, reintroducing that again.

IMHO and all that. But who am I to talk.


The DROffline meter, which was only part of the original creation and is now reintroduced with the MAAT version, is basically offline analysis tools that takes the program stream, analyzes it from start to finish, then sets 100 fixed points (or was it more, don't remember the specs off hand) for loudness measurement and finally calculates an average value. Basically DR "Integrated". This is more accurate to the realtime readout - but has one caveat: the positions for the measurements aren't always the same. Back in the days I could analyze one track 5 times and had 5 different numeric .dot readouts. Hence the simplification to numbers without a decimal.

The offline meter also calculates the dB True Peak value. And the offline meter was one of the first "batch tool" of this kind to analyze a whole folder (like a CD in WAV format).




Now that we have THAT out of the door, what the funk is the difference between DR, PLR and PSR:

Simplified:

DR (Dynamic Range) = Crest Factor
Crest = Peak (in case of DRMeter dBTP max) − Average (RMS)

PSR (Peak to Short-Term Ratio)
PSR = dBTP (max) − SLk (max)

PLR (Peal to Loudness Ratio)
PLR = dBTP (max) - ILk




The biggest caveat of DR and PSR:

These measurement tools are program material dependent. That means... the value can shrink, the denser a "program stream" (a stereo mixdown, a podcast, etc) is mixed. One rock production for example could easily read DR14 while another could read DR8.

According to the "rule book" of the "DR Standard" (*cough*), the DR8 version would be inferior in dynamic range. Therefore worse sounding, not be as dynamic. But is that really the case or actually "worse"? What if a slight summing bus compression adds to the desired sound (if done tastefully of course)? What if you didn't over-process a snare drum to raise the DR value (transients add to the felt impact of a production).

Then if we talk peak limiting (aka not using a loudness target/suitable scale/specs as backbone for the DR analysis) instead of Loudness Normalization, that issue even ads up.

This system is massively flawed IMO. And the PSR calculation suffers from the very same issue. Yet the "creators" of this (DR-Meter it's PMF/MAAT/F. Tischmeyer, for PSR it's MeterPlugs/Ian Shepherd) insist on it to be the best since sliced bread. Sorry, but after all these years, I just can't agree to that.


Yes, DR/PSR/PLR give you an overview over how "dynamic" your program material theoretically is. And offering that is actually quite genious. But that alone shouldn't be the only deciding factor if a production, a music track or a CD is "correctly mastered". It is an additional indicator - nothing more.

(I've been writing about this for years at this point...)




Enter ITU-R BS.1770-x

Now adapted and considered "standard" at this point are the ITU-R BS.1770-x specs (or EBU R128 S1, ATSC A/85, OP-59, yadda yadda - basically same thing, just slightly different thresholds for ILk and dBTP). One big box of tricks that covers all this. So no need for crazy nonsense and trademarked "DR measurement features" as being the only solution to mix audio.

The rules are simple for streaming services (Tidal, iTunes, Spotify), video platforms (Youtube, Vimeo) and TV stations (Netflix, local TV stations, etc), not to mention DAB+ (Digital Audio Broadcast):

Don't exceed a certain loudness value (ILk) according to the used platform (for most streaming services mostly -16LUFS, for most TV stations -23LUFS or -24LUFS), and do not(!) exceed -1dBTP (in some cases even -2dBTP - France and Asia comes to mind). Done.

Although certian CD releases are still driving up to -9LUFS, while I say that -12LUFS should NEVER be exceeded these days). Want to keep it simple and use a great learning tool, Nugen Audio MasterCheck Pro was mentioned already - they are up to date with all current applications and specs (no joke, something changes, they sometimes adapt at the same day!)




"But what about DR/PLR/PSR"

Simple, if you used (for example) -16LUFS as target loudness, you theoretically have a PLR value of 15, and PSR value of 12 to 10 (forte fortissimo passages can easily reach up to -11LUFS SLk). That is plenty of "dynamic range".

DR alone doesn't declare which "loudness target" was used - so it always needs a backbone like RMS = -14dB (unweighted), K-System K-14 or K-20, or LUFS. Unless you go the Peak Limiting route and don't care about a more healthier loudness value




"So which meter is ultimately correct"?

Short story - NONE.
As in - "there is no one true meter that all other have to adhere to"

Each LUFS meter can still have an offset of +-0,2LU (ITU-R specs) and +-0,2dBTP (due to different maths/code used to declare that value). So keep that in mind. Which is within specs btw.

And if we talk the "original DR Meter", then it gets even more tricky (as mentioned above) because of the different declared time windows, the inertia of the "realtime DR readout", etc. Heck, even Needle VU's have a 2% overshot, the Dorrough 40A RMS LED bargraph meter wasn't save of that either.

Do not go by strict values - else you just go insane with all of this. Learn each meter, find each ideal "hotspot", adjust your edits.




A lot of stuff to learn/understand.
This is nothing that comes overnight though.

Hope that clears the remaining questions. Sorry in advance if I can't respond any further




EDIT:
Nugen Audio MasterCheck Pro shows BOTH PSR (top rectangle window) and PLR (meter bargraph)

[cturner]

According to the "rule book" of the "DR Standard" (*cough*), the DR8 version would be inferior in dynamic range. Therefore worse sounding, not be as dynamic. But is that really the case or actually "worse"? What if a slight summing bus compression adds to the desired sound (if done tastefully of course)? What if you didn't over-process a snare drum to raise the DR value (transients add to the felt impact of a production).

In all fairness to Ian Shepherd, from the Meter Plugs Dynameter manual:

Some sounds naturally have low PSR values without any additional processing. For example: smooth, sustained sounds, “sine wave” bass sounds, solo flute or keyboard pads, and long-held sung notes. Synthetic sounds like virtual instruments, and audio from sample libraries may also have naturally low PSR readings.

If your music has “naturally” low PSR values without using compression or limiting, you don’t necessarily need to be concerned. However, check that no unwanted clipping is occurring, for example at an analog input, or in a fixed-point DSP host or plugin. This is especially true if you expected the sound to have a high PSR - for example drums or percussion.

[cturner]

The DROffline meter, which was only part of the original creation and is now reintroduced with the MAAT version, is basically offline analysis tools that takes the program stream, analyzes it from start to finish, then sets 100 fixed points (or was it more, don't remember the specs off hand) for loudness measurement and finally calculates an average value. Basically DR "Integrated". This is more accurate to the realtime readout - but has one caveat: the positions for the measurements aren't always the same. Back in the days I could analyze one track 5 times and had 5 different numeric .dot readouts. Hence the simplification to numbers without a decimal.

The old DR Offline Manual explains it thus:

In order to determine the official DR value, a song or entire album (16 bit, 44.1 kHz wave format) is scanned. A histogram (loudness distribution diagram) is created with a resolution of 0.01 dB. The RMS – an established loudness measurement standard – is determined by gathering approximately 10,000 pieces of loudness information within a time span of 3 seconds (dB/RMS). From this result, only the loudest 20% is used for determining the average loudness of the loud passages. At the same time, the loudest peak is determined. The DR Value is the difference between the peak and the top 20 average RMS measurements (top 20 RMS minus Peak = DR)

And the new Offline Mk II manual:

Technically, DROffline MkII splits the loudness values of the measured song into many thousands of different quanta, and feeds that information into a histogram, which provides needed statistics about the loudness distribution. Then, it gates or tosses out 80% of the material, taking the loudest 20% of the histogram or loudness events and measures the difference to the second loudest peak. Thus, DR is a hybrid between methods A and B, and is a bit closer to PLR then to LRA. That said, it is not LRA or PLR.

I forget what information besides DR the old version produced, but I never found it very compelling. The "Mk II" that's out now seems more useful, producing statistics that compare favorably with iZotope RX.

offline.png

If you have need of batch measurement of these kinds of statistics, at the current intro price, the application seems pretty compelling.

[Compyfox]

Basically the manual of DR Offline Mk II explains in super detail as to what each ITU-R BS.1770-x tool these days is doing "behind the curtain" already. If you're not skilled in that area, you'll walk around with giant ???'s above your heard.


The TL;DR version:
DR Meter Mk II and DR Offline Meter Mk II only introduce "DR metering" (in this case: PMF Dynamic range, which is probably: DR = dBTP - MLk) and the "per channel" loudness statistics. Nothing that declares a "defacto go to standard". (IMHO and all that)


Though it is interesting from the screenshot, that PSR is declared via SPPM (Sample Peak Programme Meter) minus Short Term Loudness k-weighted (SLk)... Shepherd uses dBTP minus SLk according to his page. Which brings me back to my point that "everybody does different", which makes the use of something unified like ITU-R BS.1770-x even more important IMO

[cturner]

Though it is interesting from the screenshot, that PLR is declared via SPPM (Sample Peak Programme Meter) minus Short Term Loudness k-weighted (SLk)... Shepherd uses dBTP minus SLk according to his page. Which brings me back to my point that "everybody does different", which makes the use of something unified like ITU-R BS.1770-x even more important IMO

I think you're referring to PSR (not PLR) here, aren't you? Shepherd's calculation is the same as Nugen's, who refers to it as "Short-term PLR" to connect it to the EBU standard.

What I found interesting in my poking around is that the PMF DR measurement constructs a ratio based on peak to the loudest 20% of the program, where the EBU standard (if I understand it correctly) gates the top 10% and bottom 95%. This is to exclude things like "background noise" and "loud gun shots" from the measurement, but reflects a film or broadcast viewpoint.

As you said, everyone does it a little differently, and at the end of the day, it's the standards in use at the point of distribution that matter most: iTunes, Spotify, etc. Working backward from one's intended use isn't a bad guideline.

[plexuss]

You have to keep in mind that trying to meter loudness is trying to come up with one number that is an agregate of every single human's hearings - it can't be done. The intent is come up with a way to estimate what people might hear based on a very statsically shakey model. So arguing about which standard or technique is moot - as long as the approach is within the ballpark of the already stastically precarious model, it's good enough. This is perceptual metering and as such is trying to fit a square peg into a round hole: measuring subjetivity.