Please help me understand metering

[Roy Slough]

I am attempting to follow guidance from YouTube for metering using additional plugins (I am preparing a Stereo wav for uploading to a streaming website)
Specifically here using MV's VU meter and SPAN's meter.

In Cakewalk meter (Track and Master) the peak shows -3db and the RMS around -12db

In the SPAN my master is averaging around -6db with the peak line at -3db (no spikes no clipping)
However at this level the MV meter is well over the red area with the overload LED permanently on.
To get the MV Meter to around 0db with no overload I have to reduce the gail by -8.7db, this reduces the average on the SPAN to around -12db.

Any simple advice on what os going on and which I should pay more attention to. 
Obviously me ears tell me to go with the SPAN and ignore the MV VU meter, as the song sounds much better louder.

Perhaps this question should be in Q&A but I put it here as it specifically references plugins.

[bitflipper]

When two meters disagree, it always comes down to settings within the meters being different.

Sometimes, it's because they are using different reference values. For example, some let you specify a reference other than 0dBFS. This is really just a convenience, allowing you to easily set a target below full scale, e.g. when mastering for streaming services versus writing to a CD. It just means that "0dB" is something other than the actual maximum value that digital audio supports (0dBFS). But it can lead to confusion when you see two meters giving different readings.

It can also be because of averaging, length of the RMS window, and FFT bucket size (called "Block Size" in SPAN). For example, a 50ms RMS window will yield different values than a 500ms window. These things can affect apparent accuracy, but they serve a purpose. That's why you can't worry too much about which one is "right". You can achieve your goals with any meter, as long as you know what it's telling you.

Then there are "weighted" meters, which try to measure sound the way human ears perceive it. Weighting can radically effect the values you see.

Then there are meters that offer "True Peak" readings. That just means they are oversampled internally, allowing greater accuracy between individual samples. That will very often result in different peak values. True peak can be as much as 6dB higher than what Cakewalk's meters show, especially after conversion to MP3. SPAN shows true peak values at the bottom of the UI. Just make sure the TP values don't go too much over your chosen target; if they do, you'll want to lower your limiter threshold a little.

Sonar/CbB built-in meters are accurate and as un-colored as a meter can be. You can specify the RMS window, but that's about it. No weighting, no "true peak" option and full scale is always the reference. For the most part, what you see is what's really happening. You can generally trust the -3dB peak values, but RMS comparisons are only valid between tracks within the project. Cakewalk's default 50ms RMS window is a widely-used standard, but you should be wary of using RMS readings as an indicator of suitability to streaming applications, especially as an indicator of loudness.

Streaming services rely on LUFS values to decide whether your song needs to be turned up or down on playback. Even if two of them use the same LUFS threshold, LUFS, like RMS, is averaged over a period of time. Fortunately, LUFS has been standardized to "short" (comparable to RMS), "momentary" and "integrated" windows. "Integrated Loudness" has been standardized to 400ms with no weighting, so it's the most consistent when used alongside transparent meters such as Cakewalk's (in a well-balanced mix, anyway).

If the target is YouTube or Spotify, for example, you'll want to shoot for around -14 on the LUFS scale. This correlates reasonably well with -14dBRMS/FS as shown in the Cakewalk meters. It's usually OK to have a slightly lower value (e.g. -16 or -18), because the streaming player will turn up your volume automatically. But beware going too far over the desired levels (e.g. -6), as that can be damaging to your sound.

There's a whole lot you can say about metering, and these are just thoughts off the top of my head. SPAN has a bunch of options, each of which will yield different results. The closest to Cakewalk is the "DBFS" option, but I'd suggest trying out the "LUFS-SL" option for comparison. I'm not going to tell you to go out and buy an expensive metering suite like iZotope Insight because you can manage quite well with what you've already got. At the end of the day, if you upload it to a streaming service and play it back, and it doesn't sound radically different than it did in your DAW, then you can claim success.

 

[Glenn Stanton]

one critical piece to all of this - ears vs meters - calibrate your system. run a pink noise -12db mono in CW/Sonar - no effects anywhere - verify your CW/Sonar meters are all showing -12db including the H/W meters. next up - set your monitors to your preferred listening level at your listening position, with this -12db pink noise - i set mine to 75db - all my speakers, headphones, mono speaker, tv etc all produce 75db from the CW/Sonar feed - usually i have to adjust the various bits in my Windows listening chain to get the right values. then use your media players etc - and set them to output 75db (not changing the Windows volume, just on each app).

mark the position on your system knobs, or make notes, take a photo of the positions etc etc so you can quickly return to your settings.

once you're done, whatever source is playing the -12db pink noise, you're getting 75db at your listening position. not a perfect calibration because different sources have different frequency responses, so you might after some experience tweak those devices slightly to your ears - but not right away - you need to let things "bake in" for a bit. 

now you'll find it much easier to set your levels in your master. as bit noted - some media services set levels regardless so making your stuff overly loud can result in lower volumes and squashed sound because of their limiting/compression etc they'll apply.

and if you want it louder (so it's "sounding better") turn up your volume knob on your system. 🙂 when you're done hyping yourself, reset the position of your volume control.

it also helps to do the opposite, turn things way down to see what is coming across or not. also, a low level (-20db for example) noise track can help to identify what is coming across or not as a preliminary static mix.

[Roy Slough]

Thanks BitFlipper,  All that information does help me understand what is happening.  I am not going to delve into what all the different metering options are in SPAN as it doesn't really help my process and I don't do this for a living any more, This is more a hobby now using my experience and knowledge from before (see note to glenn next)

Thank Glenn,  How do I check the 75db at my listening position 
In the 80s I was an aspiring musician in a band, I had my own home studio with professional monitors (NS10s etc.), 16 track tape, & mixing desk. All white noise balanced and the room treated. We never made it...
Now I'm in my 60's and just helping one of my ex-singers work up his new songs. I use whatever equipment I have left working (no NS10s anymore) with new software which is so much better than the analogue I could afford back in the day. My only expense so far has been new headphones. I do check my work on a variety of devices - HiFi (yes I have one which still works), Sound Bar,  Headphones and even in the car!!!!

Even though I have a very limited budget there is advice I would like on improving my output/monitoring. Currently I either use the headphone output (sound bar & headphones) or a USB Behringer audio device (V old now) which gives me phono output to feed other music systems.
I would prefer a better audio device, which I could connect all these to and switch between them. I dislike making the switchover on the computer is a bit laborious and the output level needs adjusting each time I switch, so poor for comparing.  - Any suggestions on what I should look at. It needs to be USB as I am working on a laptop so no internal sound cards possible. Also a reasonable price is essential.

[Glenn Stanton]

simply - using your phone (smartphone) with an sound level meter - will suffice in many cases, and set it up about where your head is. my preference (and what i use) is a sound level meter which has weighting options (mainly use A and C) to compare. i also use a tripod to hold it in place and a boom mic to set the headphone levels (initially, so i don't have to move things, but later i just stick the meter into the headphones if i'm only check those - i have several brands & models, so i have several profiles setup for them).

[fret_man]

I  used to work for a company that made the microphones for iPhones. Please be aware that there is (or used to be) a +/-3dB tolerance (3 sigma, IIRC) on the sensitivities of those microphones at 1kHz. As the frequencies get higher (or lower), this tolerance grows even larger (up to +/- 5dB at the extremes, IIRC). That was 5 years ago, but I doubt things are much different now-a-days.

[Roy Slough]

Thanks Glenn & Fret_man  Luckily I don't have an iPhone so it will be an Android - I think a slight inaccuracy will be OK as long as it is consistently so, this will make the systems "match". 

[bitflipper]

On 8/20/2024 at 1:02 AM, Roy Slough said:

a slight inaccuracy will be OK as long as it is consistently so

Bingo.

The main purpose for calibrating your speakers is to achieve consistency. Personally, the levels proposed in the K standard are a little too loud for my tired ears. Doesn't matter. However, I can clearly hear a difference between projects I made before reading Bob Katz's Mastering Audio and after. It was a paradigm-shifting read.

[57Gregy]

On 8/17/2024 at 3:00 PM, bitflipper said:

these are just thoughts off the top of my head

Imagine the info that could be imparted if he thought about it for a while! 😁

[bitflipper]

22 hours ago, 57Gregy said:

Imagine the info that could be imparted if he thought about it for a while! 😁

Sounds like every elementary school teacher who wrote comments on my report cards.

[Starship Krupa]

On 8/18/2024 at 3:25 AM, Roy Slough said:

I would prefer a better audio device, which I could connect all these to and switch between them. I dislike making the switchover on the computer is a bit laborious and the output level needs adjusting each time I switch, so poor for comparing.  - Any suggestions on what I should look at. It needs to be USB as I am working on a laptop so no internal sound cards possible. Also a reasonable price is essential.

An oft-asked question.

The bare minimum is 2 mic inputs, no MIDI, must have factory provided ASIO driver. In that category there are multiple options under $100.

The one I currently recommend is the Mackie Onyx Producer. It has 2 mic inputs, as well as 5-pin MIDI jacks to go with any MIDI controller. Dedicated headphone volume. Reputable brand, ASIO driver, comfortably under $100. The only thing that makes me prefer my PreSonus Studio 2|4 is that the Studio 2|4 has bar graph meters instead of just the peak overload LED's, which make it easier to dial in a good hot signal that doesn't go over. Not essential, but handy. Noise floor and headroom isn't as big a deal in these days of 32-bit digital recording. You can record it cool and normalize it.

For metering, the plug-ins I favor are LEVELS from Mastering the Mix, and (free) dpMeter 5 from TBProAudio and MLoudnessAnalyzer from MeldaProduction. The MeldaProduction one comes as part of their FreeFX bundle, which also includes other useful tools for your task, such as a noise generator and an oscillator (for doing sweeps).

The important thing to me was settling on one meter and learning to trust it. This was dpMeter 5 at first, then LEVELS once MtM gave me a free license. They all attempt to combine a bunch of readings into one UI, which made it confusing as hell. The advantage LEVELS has is that rather than trying to put all of that on one screen, it has different screens for different meters. You can switch from LUFs to Peak to Stereo Balance, one at a time.

Curious: what headphones are you mixing on? You can tell by my sig that I am a studio headphones geek.

[Roy Slough]

On 8/24/2024 at 4:17 AM, Starship Krupa said:

An oft-asked question.

The bare minimum is 2 mic inputs, no MIDI, must have factory provided ASIO driver. In that category there are multiple options under $100.

The one I currently recommend is the Mackie Onyx Producer. It has 2 mic inputs, as well as 5-pin MIDI jacks to go with any MIDI controller. Dedicated headphone volume. Reputable brand, ASIO driver, comfortably under $100. The only thing that makes me prefer my PreSonus Studio 2|4 is that the Studio 2|4 has bar graph meters instead of just the peak overload LED's, which make it easier to dial in a good hot signal that doesn't go over. Not essential, but handy. Noise floor and headroom isn't as big a deal in these days of 32-bit digital recording. You can record it cool and normalize it.

For metering, the plug-ins I favor are LEVELS from Mastering the Mix, and (free) dpMeter 5 from TBProAudio and MLoudnessAnalyzer from MeldaProduction. The MeldaProduction one comes as part of their FreeFX bundle, which also includes other useful tools for your task, such as a noise generator and an oscillator (for doing sweeps).

The important thing to me was settling on one meter and learning to trust it. This was dpMeter 5 at first, then LEVELS once MtM gave me a free license. They all attempt to combine a bunch of readings into one UI, which made it confusing as hell. The advantage LEVELS has is that rather than trying to put all of that on one screen, it has different screens for different meters. You can switch from LUFs to Peak to Stereo Balance, one at a time.

Curious: what headphones are you mixing on? You can tell by my sig that I am a studio headphones geek.

Thanks, I have the input devices sorted, probably not the best But I have a Behringer UM2. It has been mentioned that this would be troublesome but as long as I use a USB2 and not a USB3 port it works OK.  My "issue" is that my output monitoring is either via the headphone output OR a really old Behringer UCA200 USB which gives me RCA stereo input & output. so switching between monitoring systems is a minor irritant.

Metering, Obviously I am monitoring the track from source using the Cakewalk meters and gain staging to some degree (Once I begin adding FX and have adjusted the fader I stop messing with the gain - but I do ensure any FX have enough gain to drive them) At the master fader end I am using SPAN to watch for clipping and the overall LUFTS (and frequencies,stereo etc)

Headphones: after some research I purchased DT770 Pro 32ohm. I chose these as they had good reviews and reported a flat frequency response. For my setup I was given to believe the 32ohm would be better given the low output, but I still have to boost to nearly full to really hear everything - This works for now but is one of the reasons I am exploring a different output device
a) so it is easier to switch between the 3 systems I monitor with
and b) to perhaps give more output for the headphones.

As always there is limited budget and an output device may fall behond other essentials

Thanks again for the advice/info  

[Starship Krupa]

4 hours ago, Roy Slough said:

I have a Behringer UM2

Nothing wrong with Behringer interfaces in general, except that their very bottom of the line (the UM2) doesn't have an ASIO driver. ASIO gives best performance (monitoring latency, timing accuracy) when overdubbing. Until you get something else, be sure to use WASAPI Exclusive, which is the closest thing to ASIO you'll get with the UM2.

If you happen to use software that supports ASIO but not WASAPI (as I was astonished to learn is the case with Ableton Live 12), you'll get better results with ASIO2WASAPI than with the ASIO4ALL driver that is sometimes recommended.

Very nice headphones, BTW, I really like my Beyerdynamic DT880's. So comfortable.

[Roy Slough]

4 hours ago, Starship Krupa said:

Nothing wrong with Behringer interfaces in general, except that their very bottom of the line (the UM2) doesn't have an ASIO driver. ASIO gives best performance (monitoring latency, timing accuracy) when overdubbing. Until you get something else, be sure to use WASAPI Exclusive, which is the closest thing to ASIO you'll get with the UM2.

If you happen to use software that supports ASIO but not WASAPI (as I was astonished to learn is the case with Ableton Live 12), you'll get better results with ASIO2WASAPI than with the ASIO4ALL driver that is sometimes recommended.

Very nice headphones, BTW, I really like my Beyerdynamic DT880's. So comfortable.

Here is the thing, It was this year I was invited to mix some songs (we were in a band in the 80s/90s) So I looked around and discovered cakewalk by Bandlab was free and looked good to me. I later discovered the singer was using Reaper but by then I was committed to Cakewalk.
Whenever I tried changing the preferences to ASIO nothing worked but leaving it on WASAPI exclusive seemed to be OK (Laptop running Windows 11)
I dabbled with ASIO4ALL with little success and the WASAPI was not giving me any issues, so I have not explored ASIO any further.

In my experience if something is working don't mess with it. However, if someone can explain why ASIO is worth my exploring further I will listen.
If this was a desktop machine and I had a chance of utilising a soundcard, maybe this would be the catalyst.
OR if I were to purchase a USB device that used ASIO again that would be a consideration.
Meanwhile I will use my spare finances for things I need more (e.g. A melodyne license ) 

Please advise if I have misunderstood the ASIO vs WASAPI situation...

[Glenn Stanton]

ASIO works best when paired with the interface it was designed for. a "generic" ASIO is probably a crap-shoot in many cases. why would you choose a proper ASIO driver paired with the interface it was designed with? because ASIO will provide the most direct path from the IO to the computer processing. however, if the IO unit itself has hardware latency which exceeds the possible latency settings, then it's not going to work out as expected.

and if (in your example of mixing previously recorded material), you don't need low latency, you want high latency - one to reduce resource consumption on the PC, and two enable better PDC behaviour as the shortest delay will be based on the longest plugin delay - so more buffer time will seem smoother. 

in recording overdubs or because you're monitoring via the DAW and plugins - there will be limits in how low things can get, and the tolerance of the performers to perform with those delays. some people - 10ms doesn't phase them, some people cannot handle 2ms...

(fwiw - sound is about 1.1ms / ft so if you're 10' from your amp, you have about a 9ms delay, hence tighter live performance often works best if the band members are close together. of course if you're only monitoring via headphones, then it matters much less since the mic or DI to the amp / instrument is typically sub-millisecond)

Edited August 30 by Glenn Stanton

[Starship Krupa]

8 hours ago, Roy Slough said:

I dabbled with ASIO4ALL with little success and the WASAPI was not giving me any issues, so I have not explored ASIO any further.

I wasn't suggesting you use ASIO2WASAPI with Cakewalk.

What I mentioned was that if you have other software that can't use WASAPI (such as Ableton Live), ASIO2WASAPI is a better bet than ASIOALL.

As you have seen yourself, Cakewalk works great with WASAPI. IIRC, Cakewalk, Inc. worked closely with Microsoft while WASAPI was in development, so no surprise there.

For interfaces that don't have their own native ASIO drivers, such as my laptop's internal Realtek CODEC, WASAPI is absolutely the best way to go for software that can use it. But there is software that can't use it, even though it's been around for an eternity in computer years. I don't know what Ableton's problem is. Considering that Live! must be rock solid in order to fulfill its intended use as a performance tool, continued failure to support WASAPI boggles my mind.

The time to explore ASIO is when you grab one of those Mackie interfaces. I've seen them as Amazon Refurbished for about $30.