Is it just me, or are Logic's meters essentially useless?

I track something. The Logic meter peaks at say -2db during the track. I do not get a clipping warning window. The track is in every respect fine, and sounds fine.

But Logic's meters on playback (with no plug-ins or any other changes made to the track at all) are now happily going 2 or 3 dbs into the red.

This happens consistently, on practically every recording I make. I've just stopped taking any notice of the playback meter readings and instead use tham as a guide only, relying more on the meters on my console.

Should I just carry on assuming that 2 to 3 dbs "over" is no problem as long as it sounds OK?

--------------------
Dynamite with a laser beam...

I've never managed to get much sense out of them. Why not try the PSP Vintage Meter from here for reference, it's free.

What's your pan law set to?

You know what they say - if it sounds ok, it is ok.

I seem to remember from the MacProVideo tutorial that 'overs' on individual audio and instrument channels are fine to a degree - there is a certain amount of headroom. You just need to ensure you aren't clipping the outputs.

C

--------------------
www.2b-media.co.uk

There's something like 1500dB of headroom in a 32float mixer. (And that's not a typo!)

Just don't clip the outputs, and pay attention on plugin gains, as some designs use a 24bit fixed implementation, so if you have high channel levels or outputs from previous plugins, these *can* distort.

But the meters in Logic here have always proved accurate enough for me. If you have a stereo signal on a stereo channel with the fader at 0dB, and your master is at 0dB, the levels should be indentical.

If they are not, you may have some other routing funkiness going on (for example - you've sent that channel to a bus which is also sending to the main output, thus raising the volume)

Mono channels can vary in overall level when panned depending on the pan law.

Quote cwillsher:

---

You know what they say - if it sounds ok, it is ok.

I seem to remember from the MacProVideo tutorial that 'overs' on individual audio and instrument channels are fine to a degree - there is a certain amount of headroom. You just need to ensure you aren't clipping the outputs.

C

---

All depends. I did a quick 'mastering' job (assembly basically) for a mate, in Logic, and noticed some of his tracks looked pretty hot, but everything sounded 'ok', and Logic wasn't showing clipping.

So... I finished the job and sent it off to the CD manufacturer- who promptly sent it back, saying it was 'over' and unprintable ! So I invested in Elemental's Inspector (now Roger Nichols and no doubt some idiot name...) and ran the same files through it. Mayhem broke out, so to speak- multiple 'overs' with some blocks of 40 or so samples all over the limit, on at least 3 tracks. No wonder it came bouncing back.

So- yes, Logic's meters are not exactly on the case. Probably best to keep your peaks bouncing around -3db at max...

--------------------
http://anotherfineday.bandcamp.com/ http://anotherfineday.co.uk http://apollomusic.co.uk

Whatever you record is exactly what Logic plays back. If there is a seeming level gain during playback, the signal must be arriving not only once, but twice, at your output channel. Check your routing, there has to be an error somewhere (happens easily).

Quote desmond:

---

What's your pan law set to?

---

-3 db compensated....

It's not a routing problem. I've got lots of carefully put together templates, but I almost never use busses or auxes because I do so much on the console... most of the time Logic is a glorified tape recorder, for me.

I can set up a song with one channel, record myself saying hello, and watch the meter peak at X db (the little orange box above the meter records this and displays the peak throughout, and after, the recording process, till you press play or whatever).

I play back, it peaks at a different level.

--------------------
Dynamite with a laser beam...


Edited by Steve Hill (08/04/07 08:28 PM)

Hmmm... just done a quick test. Recorded a mono track centre panned at 0db. Panned it hard left and it played back at +2.9db...

--------------------
Dynamite with a laser beam...

Hi
In the Apple Soundtrack Pro forum there is a corresponding thread about STP's level meters being as unreliable - the Mixer level at zero gain reading lower than the Track level for a reference signal.

I havea theory

in terms of metering.... it's possible that the playback metering detection, is effectively a bit faster as it can look ahead on the source data , and therefore, shows more transient information, but as it's fairly slow display wise... it appears as a higher overall level.
The input metering cannot detect and display as much as it cannot look ahead on the signal source.

answer... use more accurate external metering and ignore Logic.


and track with more headroom......... it's not like you're short of dynamic range to play with....

--------------------
if you don't know who i am, i aint gonna tell you.

Hi Max. I quite agree about headroom and usually I do... my 0db test was with an oscillator to prove a point.

If the playback meters are the more accurate ones, that's fine. It's when I start adding EQ, compression or whatever in the box, adding several dbs, that I need to rely on the playback meters.

--------------------
Dynamite with a laser beam...

Quote Steve Hill:

---

Hmmm... just done a quick test. Recorded a mono track centre panned at 0db. Panned it hard left and it played back at +2.9db...

---

Well, it should be +3dB, but there's only so much resolution in the numbers in the meter, and the meters aren't reading on every sample, they average, so every "sample" of the signal used to redraw the meter probably is averaged over 256-samples or so, causing a slightly innacurate reading, depending on where the meter samples are taken, and what the sample values are - it's like an averaging sample-and-hold circuit. Displaying metering on every sample across all channels would take up significant CPU, which is why is done like this - and Logic always gives priority to audio, and it's metering/GUI priority will fall if the audio needs all the CPU it can get.

Don't forget, Logic's meters *aren't* intended to be highly accurate sample-level displays - they are an indication of levels and track activity. For accurate levels, Logic provides more sophisticated metering plugs, or third party tools such as Inspector XL are invaluable - a good set of meters is a production must, imo. This is true for all DAW's as far as I know.

As for tomafd's mastering thing - don't forget the meters won't show reconstruction overs - it's quite possible to have sample values of significantly less than 0dBFS, and thus Logic's meters won't display overs, but those sample values cause the reconstructed waveform to go over 0dBFS, which was why your mastering guy rejected them.

This is yet another reason to print your mixes with headroom, or at the very least, use decent metering that can handle these sorts of complexities...

Thanks Desmond - I wasn't expecting massive accuracy. and as I said I rely far more on the meters on my desk anyway.

But you'd think Logic might issue some kind of health warning, would you not? I wonder how many people trust these displays as a matter of course?

Anyway, one object lesson learned: record with the pan pots in Logic more or less where you would expect the final pan position to be...

--------------------
Dynamite with a laser beam...

Bit viewer another plugin you can use.

Cheers.

http://www.macmusic.org/software/view.php/lang/en/id/3154/

Quote Steve Hill:

---

Anyway, one object lesson learned: record with the pan pots in Logic more or less where you would expect the final pan position to be...

---

Well, you're using the 3dB compensated setting, so you're *asking* Logic to raise your centre panned signals by 3dB.

If you don't want this, then use the -3dB pan law (not -3dB compensated).

See: http://logicquicktips.blogspot.com/2006/10/laws-of-pan.html

Interesting.For me, as far as I recall, the 'in' levels on my Fireface mixer and logic very nearly agree, and Logics outputs agree with what I put in.

I will have to do some checking. And I'm on the -3 pan law setting.

--------------------
PA stuff on FB

Logic has three pan law settings - 0dB, -3dB, and -3dB compensated.

The two -3dB settings are identical apart from a 3dB offset.

Using the compensated setting will keep hard panned mono signals at 0dB and push them up by 3dB in the centre - hence the +3dB discrepancy reported earlier.

Using the -3dB (non-compensated) setting, hard panned signals will be -3dB (rather than 0) with centre panned ones being 0dB (rather than+3dB).

This has been annoying me for ages!

I'm pleased it's not just me having these problems. I'm forever turning down channels even though they sound fine. I'll give up now and put it down to Logic just being buggy (which it most definitely is - Core Audio Overload? On 2 tracks!?!)

SJ.

Sorry folks, made a mistake, you cant really use bit veiwer for metering.

:-)

is this feature not available in logic express?

Quote refinerymusic:

---

is this feature not available in logic express?

---

What? The inaccurate meters feature?

Probably!

--------------------
Dynamite with a laser beam...

Quote Steve Hill:

---

Quote refinerymusic:

---

is this feature not available in logic express?

---

What? The inaccurate meters feature?

Probably!

---

Are you saying that your meter troubles are not down to user error, but caused by Logic itself?

not entirely unrelated - there was a long, long thread a while back about distortion at the two bus, lots of stuff about 32 bit float and all that, and how it shouldn't be a problem - but I think logic does start to sound 'orrible the closer you get to 0dbfs on the two bus, especially if you use a limiter or whatever - I reckon about -6dbfs is OK, above that it gets sorta fizzy

I'm not a Logic user, but this might help... -2dB peaks are a bit close to the edge aren't they?

I set the peak level at -12dB in Pro Tools when tracking. Gives me lots of head room. Might as well use those 24 bits for something! No clipping distortion and loads of room for mixing. Sounds a hell of lot better than -6dB.

I'll use the PT meters, but If I need real metering I'll use Waves PAZ, particularly on the master outs.

--------------------
madtheory creations
Synths and pianos for Kontakt

Quote tipex:

---

not entirely unrelated - there was a long, long thread a while back about distortion at the two bus, lots of stuff about 32 bit float and all that, and how it shouldn't be a problem - but I think logic does start to sound 'orrible the closer you get to 0dbfs on the two bus, especially if you use a limiter or whatever - I reckon about -6dbfs is OK, above that it gets sorta fizzy

---

What's possibly/probably happening is that as you start to get close to 0dB on your main outputs, you're actually clipping your output convertors from overs from the reconstructed waveform, even though the individual sample values are not clipping.

Headroom folks. There's reasons why it's a good thing, and as we have such high s-to-n ratios these days, it's even more of a good thing than it was in ye olde days of analog/tape...

Quote Steve Hill:

---

Quote refinerymusic:

---

is this feature not available in logic express?

---

What? The inaccurate meters feature?

Probably!

---

LOL! i meant the preference option to turn on or off the -3db compensation. i cant find it in logic express and im afraid it is not an available option.

I don't think Logic Express has user selectable pan laws. It probably just defaults to 0dB, as all earlier versions did.

It's under Song Settings -> Audio in Logic Pro

What're pan laws? I've never heard of them.

SJ.

http://logicquicktips.blogspot.com/2006/10/laws-of-pan.html

Quote desmond:

---

Quote tipex:

---

not entirely unrelated - there was a long, long thread a while back about distortion at the two bus, lots of stuff about 32 bit float and all that, and how it shouldn't be a problem - but I think logic does start to sound 'orrible the closer you get to 0dbfs on the two bus, especially if you use a limiter or whatever - I reckon about -6dbfs is OK, above that it gets sorta fizzy

---

What's possibly/probably happening is that as you start to get close to 0dB on your main outputs, you're actually clipping your output convertors from overs from the reconstructed waveform, even though the individual sample values are not clipping.

Headroom folks. There's reasons why it's a good thing, and as we have such high s-to-n ratios these days, it's even more of a good thing than it was in ye olde days of analog/tape...

---

yep, keep them levels down and beware of crap plugins on the two bus

Quote desmond:

---

I don't think Logic Express has user selectable pan laws. It probably just defaults to 0dB, as all earlier versions did.

---

It's true. Weak.

Quote simonmitchell:

---

What're pan laws? I've never heard of them.

SJ.

---

"Despite a great surge of popularity throughout the country, the instrument continued to enjoy official disfavor as being noisy and disreputable, associated with troublemakers and anti-colonial rabble-rousers. An anti-pan law was passed in 1949, giving the police the freedom to smash the drums and arrest anyone caught playing them. But the easy availability of the drums made it impossible to wipe them out."

http://www.stephenbrookes.com/music/

--------------------
Dynamite with a laser beam...

I just did a small experiment in Logic, with a Fireface 800 interface:

Logic's Test Oscillator, square wave at 0 dB, inserted on and audio track. All outputs set to 0dB, panned centre, -3 pan law. (Very funny joke, BTW!)

Channel Meter = -1.1
Output Meter = -1.1
Bus Meter = -7.1
Aux Meter = -4.1

Output to FF 800; FF Mixer input from Logic shows a -0.02 input.
The FF 800 out meter = -0.02
Patched this to a FF800 input, analogue.
FF 800 input meter shows -0.62
The Logic input meter from FF800 shows -1.7
The recorded signal in Logic shows -4.7 on the channel meter and -4.7 on output meter ... which then shows up as -3.62 on the FF 800 mixer.


Here are the white noise figures:

Test Oscillator, white noise at 0 dB, inserted on audio track. All outputs set to 0dB, panned centre, -3 pan law.

Channel Meter = -3.0
Output Meter = -3.0
Bus Meter = -9.0
Aux Meter = -6.0

Output to FF 800; FF Mixer input from Logic shows a -4.68 input.
The FF 800 out meter = -4.68
Patched this to a FF800 input, analogue.
FF 800 input meter shows -5.47
The Logic input meter from FF800 shows -2.2
The recorded signal in Logic shows -5.2 on the channel meter and -5.2 on output meter ... which then shows up as -8.46 on the FF 800 mixer.

Hmm...

--------------------
PA stuff on FB