Set audio levels

As audio originates from many different sources, invariably the levels vary, sometimes by up to 5db. Annoyingly, even CDs rip at varying levels.

One solution of course is to use Audition and level each file, which is the option I am using at present, but it’s very time consuming. Another is to prefade each item before it is played. This I do in live assist, but obviously this is not a solution in auto mode.

In the analogue past to get round this issue, we would level every record before cueing it up.

It would be great if it was possible to set the level (higher or lower) in the PFL/properties dialogue. So the workflow would be check properties, set cue, ramp and fade points and then set the level. Ideally there would be a digital PPM meter and fader to adjust the level. The data could be stored in the mmd file or database in the case of version 3.

Personally, I think this would be better solution than an Automatic Gain Control. They tend to normalise everything and to increase the levels of quiet, slow, songs so that they end up too loud. I prefer to make levelling decisions myself.

What do you think?

In case you haven’t noticed, this is already possible in v3. There’s a slider in the PFL dialog.

Thanks Torben, I didn’t know that. I was planning to install version 3 today, so this is an extra incentive.

Is there a way to import existing mmd data into PostgreSQL?

I have to say that I think the idea of a fader within mAirList is a step backwards. That’s what mixing desks are for!

But to Andrew: levelling issues should IMHO be handled at the ripping stage (or data import if the tracks have been ripped on a different PC) and not left until the mAirList prep. stage. Unless you’re using a non-MP3 format, I recommend the freeware program MP3gain for this job. Ignore its ‘standard’ mode of operation, because that gives a consistent ‘average’ level and is really for home or car listening. Instead, after loading your tracks, press Ctrl+M to find the PEAK level of each track, then press Ctrl+X to APPLY those peak levels. Yes, you will still have to do some gain adjustment on playout, but at least you can set the desk knowing that nothing will ‘go over.’

PS: “PPM meter” is a tautology (PPM = Peak Programme METER). :wink:

BFN
CAD

Hi Cad,

You are of course right on two counts:

  1. Level control should be applied on acquisition. But my music library (mostly wav files, some mp3s) has been built over many years and from numerous sources. So if there is an easy way to sort my level issues, then I want it.

  2. PPM meter is tautologous!

Regards

Andrew

Yes, they are automatically imported when you load the songs into the database for the first time with the Synchronize tool.

Great, thanks. I should have guessed you would have already sorted that.

First off, if you want to keep the unmodified originals, remember to work only with COPIES. Either that, or burn the originals to DVDs (or memory sticks, or memory cards, or … )

For the MP3s, as I already said, look no further than the excellent freeware program MP3gain. After install, click Options, Advanced, and tick the Enable “Maximizing” features box before clicking OK. You can then load up a big bunch of MP3s and use the Ctrl+M, Ctrl+X ‘trick’ to force it to perform a WAV-style peak normalize to 100% (which it confusingly refers to as 89.0dB—blame the movie industry for that weird standard, apparently >shrug<) on all your MP3s. You will likely find that most of them are ‘over-peak’ by anything up to about 9dB (!).

I use MP3Gain myself for this, and have done for some years now. I rate it as one of my three must-have MP3 utilities (the others being MP3DirectCut, which is freeware; and MP3-Tag Studio, which is cheap but not free.)

For the WAVs, there are several programs which will batch-normalize them. Most are cheapware (up to about $25), but read on.

If you are totally strapped for cash, one slightly desperate option would be to use Audacity. Basically, you open the first WAVfile, Select All, then Effects/Normalize, THEN import another bunch of WAVs and Output them to separate files. (Allegedly. I just read this on the cdfreaks Forum! And personally, I loathe Audacity.)

The only true freeware solution which I’ve seen which can batch-normalize WAVs comes from those well-known Aussies, NCH. I know that many people loathe their DLLs, but in our case all we’ll be wanting is to convert from WAV to WAV but adding a Normalize whilst doing so (if that’s not too bewildering a concept?). They have a freeware program called Switch which claims to be able to convert up to 32,000 files per batch. (No, I haven’t personally tried this program!) You can read all about it at http://www.nch.com.au/switch/ and decide for yourself; I’d say it would fit your particular bill. Look at their screenshots and all will be clear. :wink: Provided you do a System Restore first, or use some utility like Returnil, you can always give Switch a whirl with a few files, see if it’s right for you, and have the option of dumping it afterwards if it isn’t.

Hope that helps?

BFN
CAD

Thank you CAD.

I’ve been using Audition and manually amplifying files as necessary, adjusting the db increase to peak at approximately -2db/0db.

I did experiment some time ago with batch normalisation. For me, the problem with this is that it over amplifies songs/music that are meant to be quiet.

I do want there to be a natual sounding, range of levels and so far manual adjustment has been the only way that I can successfully achieve this.

In a studio I pre-fade and set the level on a PPM so that the vocal or loudest part of a track will peak at 6, and I set my microphone level also at 6. The rational being that my voice shouldn’t dominate the music that I am presenting and if it is loud, then I shouldn’t be talking over it anyway.

Regards

Andrew

Hi Cad,
i have some questions about MP3Gain.
You said that it will better not to use the standard setting for average volume. As i understand, each level which cause a clip will normally not transfered to the output (also somewhere descirbed in the helpfile of MP3Gain). As i undestand this means that the “heared volume” is more constant over the different files.
If i use the maximum peak setting i’m a little bit affraid that diffrent songs have mostly different levels but, you’re right, no level which is higher than the set level.

But for using MaL in Auto mode i think that the standrd method is easier to handle, or? Means, if i have a short peak, ok, either will be cutted by the louder signal is so short that no one realize it. Or i’m wrong?

I use my mixer only at 3/4 capacity (full digital mixer) so taht in case of some peaks nothing will happend.
Also i must say, i’m not using for radio but if i take a look on the Meters it looks more constant as with the peak setting of MP3Gain.

So i’m not sure what is the better method for if MaL used in Auto Mode?

Greets

Piet, it depends on how your station (and presenters!) operate. Our presenters turn up everything ‘to the max,’ so it’s more important for me to have audio which won’t ever go ‘over peak.’

If you’re mainly using AUTO, then yes, using the ‘standard’ MP3Gain settings would make all tracks sound like they are at the same level. Having faders less than fully open will also reduce ay potential problems!

BFN
CAD

Hi cad,
thanks for this info.
I thought maybe there are some other special to use this method you described.
So before normalize in one of the ways it should be cleared before how the most moderators want to work (fader fully up or not) :slight_smile:

Greets

Andrew, I just noticed that you said:

Assuming a mono PPM (?), music should average PPM 4 and peak at PPM 5, and presenter mic should average PPM 5 and peak at PPM 6. That’s what the BBC taught me, anyway! :wink:

If you’re on stereo PPMs, things are somewhat more complicated.

BFN
CAD

The BBC standard I was taught was 4 for music and 6 for speech.

As soon as I got to IRL, it was 6 for the vocals and 6 for the mic. It means if you talk over the start of a song, the vocal comes in at the same level as your voice which I think sounds right.

Remember that IRL stations use a LOT more compression and ‘processing’ than the BBC do. They rely a lot more on the post-desk processing chain to get the levels ‘right.’

It was very instructive when all the commercial stations had a charity day a few years ago and were broadcasting the same programmes each day. Flicking through the stations, you could hear pretty major differences in how they had each set up their Optimods (or whatever processors they were using)! :o

BFN
CAD

Hi,
i followed quite intrested this discussion and maybe i have a little information why settings for Mic is 2ppm higher then music.
Simplified it is like this, each doubling of a soundsource (mono) will increase the sound level by 3dB. By the reason that a stereo signal is not exactly a doubling of 2 mono signals, it is a little bit less. Means from other hand, if you have a Mic signal (which produce only single mono signal) you must set it rouhgly 3dB higher to get it to the same level as an double mono signal.
Most full digital mixer can do this adjustment, as CAD say, in the Post processing with an digital offset (i do same way with my one). So there, as the operator, you use the same level seetings for both.
On analouge system it was mostly not possible to do this in post processing without huge external signal routing. So there the settings are different, or the gain is more open.

Greets

I agree with you, Cad and Piet,

we also use MP3Gain with much success. It all depends on how much compression you use on your audio. We prefer a moderate compression rather than the commercial stations. So our Audio is normally with 2-3 dB higher.
Having a 6dB Headroom on the mixer was always sufficient to overcome most problematic situations.
It really depends on your studio setup and the skills of your techs/moderators.
Ah yes, our studio is still fully analog.

regards:
-Serge-

Piet, that’s partly correct but not 100% correct.

Music has a higher AVERAGE level than speech. Therefore, to make them sound the same, you need to make the PEAK level of speech somewhat higher than music.

Ironically, this difference in average levels is easier to see on an old-fashioned VU meter (remember those?) than a PPM, because VUs meter measure average levels and PPMs were specifically designed by the BBC to measure peak levels. (The reason peak level was so important in the analogue days was that some signal transmission equipment, especially the equipment used to convey signals along landlines, would either cut out or blow up if the peak level was too high.)

Also, remember that PPM ‘units’ are 4dB and not 3dB. (In other words, PPM 6 = +8dBm, PPM 5 = +4dBm, and PPM 4 = 0dBm.)

BFN
CAD

Hi Cad,
thanks for explain. I think i got somewhere a wrong information about the PPm system :D:D.
But quite interesting, i was talking long time ago to another engineer about that and he explained me as (basically) as said before. Looks like that there are two ways to come to the same result.
And with the average voltage, quite logical.

Greets