Maximum output level


My original wav file exported from an Ardour sessions had a general signal level far above 0db with peaks even up to 6db. It was too much.
Next I used JAMin EQ and its limiter, so now I have my sound even louder, but not going beyond 0db.
I want to increase loudness of the whole song a little bit still, but cannot see an option in JAMin’s limiter to go beyond the 0db level.

So now I have a single track in Ardour with my music, and I’d like to ‘scale’ the whole signal level up, so the maximum reaches e.g. 2 or 3 db.

Is there a special plugin to achieve this or should I use a regular amplifier-like plugin?

B. Regards,


Thanks for your response.
You’re right, I know the basics, but I’m still missing lots of knowledge in this area.
So far have I noticed, that even I had not used JAMin before, and recorded my stuff on an audio CD or played it as mp3 on a few different devices, I had not heard any clips (peaks had been around 4 db), however I had not been satisfied with a general volume level (it was too quiet and dim).

A few days ago, I used JAMin on my last song, and the general effect was very good, but I found this still not as loud as it could be when comparing to other commercial recordings (I listen to metal, and try to compose it in my home ‘studio’).
And I used JAMin after I mixed and exported all tracks from Ardour to a .wav file (so I used JAMin on a single file with my song, and recorded it again as a single stereo track in Ardour with the signal level max at 0db).

What did you mean by “If you just want more limiting, why not set the thresholds lower in JAMin”? What thresholds did you mean and what kind of effect I could achieve with that? As far as I remember I had my limiter settings in JAMin set to 0db (the input level), and the output level was set to 0db, but it’s not possible to go beyond, and set the output to let’s say 2db.


Hi Tomek,

I assume you know the basics of how digital audio works, but it’s worth pointing out that Ardour, JACK, and friends work with a floating-point digital representation of audio. This basically means that, for any in-the-box processing, headroom is not a concern - you can use pretty much any level you like as your “nominal” level. However, when your audio leaves the system (on a CD, MP3 file, or out from your sound card), it has to “fit” into the fixed, limited range of numeric values supported by the medium, e.g. -32768…+32767 for a 16-bit medium like CD-DA. Anything louder than that will be clipped, with the out-of-range part of the waveform simply being thrown away. This clipping happens at the “0 dB” or “full scale” mark on many peak-reading DAW meters, so I’m not sure why you would want to go “beyond” the 0 dB level. If you just want more limiting, why not set the thresholds lower in JAMin?

If your Ardour mix is regularly exceeding 0 dB, I would think you would hear a lot of distortion (if the master buss is sent straight to your sound card). What style of music are you engineering, and how loud is your monitoring level? -20 dB FS (RMS) is a reasonable sort of level for a pre-mastering mix (with peaks approaching but never quite reaching 0 dB FS).

It may be worth mentioning that the 0 VU nominal analog level and 0 dB FS (peak-reading) are nothing like the same measurement. 0 dB FS (peak-reading) is more like 20 dB above 0 VU (and VU meters are averaging, not peak-reading). 0 dB FS (peak-reading) is not a “nominal” or “target” level - it’s a “do-not-exceed-ever” level. :slight_smile:

Hope this helps!

1 Like

you don’t need to set your output louder than 0db. If you lower the limiter threshold and/or turn up the input gain, the limiter will squeeze some of the dynamic range out of your recording making it sound louder overall (without exceeding 0db).

Tomek, the issue of loudness is an interesting one, and I think you hit on something with “quiet and dim”. Human hearing is much more sensitive around the range of 2-5 kHz, and very much less sensitive for low frequencies. Have a look at the chart here:

Bass frequencies therefore require a lot of energy to sound loud, and this can eat into your available amplitude. Compression and/or limiting can be useful here in managing the bass - the compression will reduce the amplitude, but you probably won’t hear that much of a difference. At the same time, you can make the recording sound louder by boosting higher frequencies with EQ, without significantly increasing the overall amplitude of the signal. This can make it sound louder without reducing its impact. (Too much though and it can start to sound thin, harsh or tinny.)

It may seem strange, but I would recommend not using recent commercial CDs as a guide to how loud your final recordings should be. I’m sure you’ve heard of the “loudness wars”? In my opinion (and I’m not alone), recent releases have gone well beyond the point of compromising sound quality for the sake of intrinsic loudness, even for metal.

Remember that you cannot control the loudness at the listener’s end. They’re the ones with the ultimate master volume control, and also their hearing will adjust to the overall level of the material. If you can retain the punch and dynamics of the recording in the mastering, that will still translate no matter what the playback level. Also, try comparing the “before” and “after” sound, adjusted so that they sound equally loud, and only then try to judge which sounds better.

Regarding limiter thresholds: compression and limiting work by making the loud stuff quieter. The threshold is the point at which that gain reduction kicks in - in the case of a limiter, you’re basically saying “I don’t want my signal to have anything louder than this point”. Lowering the limiter threshold will squash the transients (and eventually also the dynamics) in your recording more. That does enable you to apply more gain post-limiter for increased overall loudness, but it comes at a cost.

Rock on… :^)

Thank you very much for your comments. I played a little bit with a few settings in JAMin yesterday.
When I increased the limiter input from 0 to 1 db, and the limiter output set to 0 or -0,2 db, it was louder, but some negative influence of the change was heard in particular parts of a song (it was acceptable, however not as good as on the 0db limiter input level).
So next I returned to 0 db in the limiter input, but increased the general input from 0 to 4 db, and decreased the boost from 4,8 to 4. I think it was the best balance between loudness and quality (at least for this song of mine).

Btw, would you be so kind to take a look at my JAMin settings? Maybe you could advise some changes to make my sounding even better? (I must say I’m very satisfied with what I managed to get, however lack of more advanced knowledge surely stops me from getting the most of the sounding, even in my home ‘studio’).

The following settings are from a session before I started the discussion:

And those are from yesterday tests:

Thank you very much for your helpful comments!

How you set your tools depends on the signal you feed into them. So posting the settings alone usually wouldn’t help that much.

But one thing that is very evident is that you boosted high and low end up to 9 dB. As I usually would consider mastering a more subtle task it makes me think that there might allready be something wrong with your source material. If this is a multitrack and not a 2-track live recording you should consider revisiting your session and try to fix it there first.

Maybe I took a wrong direction and an aim from the beginning - my idea was to get rid of the dim from the source material, and make it as loud as possible without loosing quality (at least for my ears).
Because of my lack of knowledge I did not put too much attention to what levels I set particular frequencies to, but rather focused on what I heard after having played with those knobs and curves, even they were out of ‘scale’.

Thanks for your suggestion, I’ll try to go down with them a little :slight_smile:

My material has many tracks with different instruments and effects, including drums and vocal.
Instruments are guitars and synths played by either qsynth or zynaddsubfx.

If it helps, in JAMin, the limiter “Input” slider is just a gain control - you have 20 dB of boost/cut there if you need it. The slider in the limiter box labelled “Limit” what I was referring to as the threshold.

Note that the signal chain in JAMin is a bit strange too (at least on my system): the master Output fader seems to be before the Boost and Limiter elements. So, actually, there is no post-limiter gain control - the limit you set is the final limit for the signal. (I think the meters aren’t actually in the same point in the signal chain as the faders they are next to, either.) AFAICT, it goes:

Input gain -> EQ -> Compressors -> "Master" Output fader ->
    Boost (distortion) -> Limiter Input gain -> Limiter -> out

So, probably just set the Output to 0 dB and the Limit to 0 dB (or maybe a tad below) as you’ve done, and leave them there. Then you can adjust the overall output level with the limiter’s Input control - just remember it’s a tradeoff! With the limiter engaged, you would never expect to see the output level go above 0 dB.

The other thing I should mention is that the meters in JAMin are all peak-reading. These have limited use for mastering: you might want to try jkmeter (based on Bob Katz’s K-System design) as an external master level meter. It shows not only peaks but also average levels, with a zero-point that corresponds to “LOUD”. Average levels correspond much better to how loud the mix actually sounds - peak meters are really just for watching for clipping. The K-14 meter (with 14 dB of headroom) is for typical pop/rock music, and the K-20 (20 dB headroom) is for higher dynamic range stuff like classical music. Grab the source from the source (if your distro doesn’t have a package):

I’d second the C.L.A.'s opinion that mastering should be a subtle spit-and-polish process rather than a major overhaul of the sound. :^) Would you be willing to post a clip of a recording, so we could check out the mix in terms of levels and tonal balance?

Thanks for this great description! It’s better to see how the signal flow looks like - after getting it all together maybe I’ll be able to do mastering more reasonably.

Sure, I can put some clips to download, but do you want them in original .wav or .mp3 (best quality)?
And including vocals or rather only a “clean” music to have a better picture? (I’m asking because I was mastering the final export, including vocals)

I’ll publish those clips in the evening, because I’m at work at the moment :slight_smile:


OK, I zipped 3 samples (before JAMin session, and after), so 6 in total; no vocals included :slight_smile:

[file size is 49 997 855 bytes]

Forgive me also my lack of skills in playing the guitar - this is my very first guitar which I bought half year ago, so haven’t had much time to practice… but any way, these samples come from my first guitar song :slight_smile:

Thank you!

Thanks for posting those - I like that orchestral metal vibe. :slight_smile: Nothing wrong with your guitar playing either!

First comments, on the pre-JAMin mix:

  1. It's already too loud, and is clipping. You could probably take the level down 5 or 6 dB on the master fader in Ardour; possibly more since you'll want to leave room for the vocals, which will probably add another 3 dB.
  2. The guitar sounds a bit fizzy (I'm guessing it's recorded direct?).
  3. The drums are lacking in punch, and sound rather diffuse. It's a nice sounding room/reverb, but a bit too ambient maybe.
  4. The bass instrument is a bit indistinct.

There’s no doubt that the things you’re doing in JAMin are improvements, especially for the guitar and drums, but I think you could get better results by applying them to separate tracks in the mix, rather than to the whole mix, especially the effects like the Boost distortion and the limiter. I’d try the following:

  • Give the snare drum some EQ boost in the 1.5-2.5 kHz region (and perhaps upwards of there as well) for a bit more crack.
  • Some EQ boost on the kick below 100 Hz is usually good, with similar boost in the upper midrange to the snare for punch and impact.
  • If you've added reverb to the drum track, try taking it down a notch.
  • Another layer of distortion on the guitar could be effective, and some EQ to take out some of the "fizz" in the highs and give it a bit more body.
  • Add some distortion to the bass track for a bit more grind and character, perhaps with some compression and EQ to give it a bit more presence.
  • The dullness is due to the overall shape of the spectrum, which gets quieter as you go higher. Try EQing the drums (or perhaps the whole mix) with progressively more gain as you go up the frequencies, starting at around 5 kHz. By 15 kHz you could afford to be boosting by about 12-15 dB, I'd say, and then have it taper off after that. I also noticed quite a bit of energy in your original mix around 160 Hz (which may be due to the kick drum or perhaps the guitar or bass) which you could try to take down a little.

Also, have you thought about double-tracking (or triple-tracking) the guitar part? I’ve always liked that layered stereo rhythm guitar sound in metal, but that’s a matter of taste really. :slight_smile:

For mixing, try turning up your monitor speakers rather than trying to boost the level using buss limiting and distortion, and try jkmeter for keeping the average levels reasonable so you don’t eat into your headroom - you need that to keep the punch of attacks, especially drums.

Hope this helps!

Oh, also, for looking at the frequency spectrum and comparing it with reference recordings for working out EQ, Baudline is a great tool.

It has an averaging spectrum analysis mode, and you can switch between different “destinations” to plot multiple curves on the same graph for easy comparison.

Really BIG THANKS for your comments!
I’ll try to proceed as you described, but for sure it will take me some time before I get it under control.

My source session in Ardour has plenty of tracks:

  • particular instruments were recorded separately (different tracks)
  • bass is very weak, I know, but I don’t have anyone who could play the bass line, so I ‘played’ it using MIDI seq, qsynth and some soundfonts I had found over the internet
  • drums were separated a little too - kick, snare, toms, and hihats have their own tracks, so I can easily (for example) turn hihats up
  • guitars - two main tracks, mono, one a little bit panning right, the other one panning left, having the same riffs (but not ‘copied & pasted’, I played and recorded every riff twice on each of those tracks). The third track is centered (mono too), and it contains some guitar melody in a few parts of the song with a reverb effect on it
    btw, my guitar is connected to the POD X3 device, and next from the POD to my soundcard (Edirol FA-66) to the second (Hi-Z) input with one cable, so I record in mono, direct (I suppose if we think of the same thing :slight_smile: )
  • vocals

I also noticed the ‘fizz’ effect in some parts, but I did not know how to get rid of it (maybe I should reorganize the way I record guitars, but not sure how it should be done to have a desired and better effect).

OK, I’ll try to spend some time this weekend and next week to proceed according to your suggestions.
They are very helpful, so once again thank you very much for them!
I’ll let you know when I have anything new ‘masterized’ :slight_smile:


I tried to follow your directions, and made a few changes yesterday:

  • in Ardour, master fader was decreased by 8 db

  • in JAMin, for snares I increaed a little the EQ level around frequencies you had mentioned (file: gtr001_snares.jam)

  • did similar things to the kick (file: gtr001_kicks.jam)

  • drums had no reverb, so no modifications done there

  • for guitar tracks I added foldover distortion in Ardour with default settings (both params set to 0, any way it sounds a little bit better; when having set those params to some other values, the sounding was getting worse, so I stayed with “0”)

  • in JAMin I did some EQ modifications to guitar tracks (file: gtr001_guitars.jam)

  • for the bass track I did similar things as tor guitars’ one: foldover distortion + some EQ changes (file:gtr001_bass.jam)

  • Next I exported a part of the song from Ardour with the above changes, and applied the final JAMin session to the whole exported mix (file: gtr001_all.jam). In JAMin sessions for particular instruments I only played with EQ, while for the whole mix I also added some boost, input level, and EQ modification.

  • Additionally I exported the same part with vocals, and applied the gtr001_all.jam for the whole mix.

If you were so nice to have a look, the samples and .jam files are available under the link:

(file size sie 60026894 bytes)

I’m also wandering whether I should play with other settings in JAMin (e.g. compressor curves and related params) or rather stay with the current result.
I’m not sure in what way I could modify those to not break the work, and to not ‘go out of the scale’ like in the beginning when I was blindly moving different sliders.
Do you think I should focus on them now or is there still more to change in EQ or just leave it as it is?
And I hope there are no clips now… :slight_smile:

Thank you!

I’ve had no time to touch those tools yet (jkmeter, baudline), but I’ll try to have a look at them this week.

Hi Tomek,

Sorry, I should have been clearer about what I meant by fixing it “in the mix”. Certainly you can use JAMin on a track-by-track basis, but what I had in mind was applying the same kinds of effects on individual tracks using plugins within Ardour itself (for example, a parametric or graphic EQ plugin, compressor, distortion). I think it would be more convenient that way since Ardour will manage saving and reloading the plugins and their settings for you, and you would avoid e.g. exporting or bouncing or running multiple instances of JAMin and managing all the connections via JACK, which could be a hassle.

Thanks for posting the updated samples and files - I will take a listen when I get the chance (should be in the next day or two).

I was going to mention too, units like the POD can give pretty good results without the hassle of setting up a full guitar rig + room + mics, but the consensus seems to be that it won’t sound as good as the real thing. A certain “fizzy” quality is one of the most common complaints. Some EQ and additional room ambience (e.g. using an impulse convolver) may be able to help with that. Sorry I didn’t notice your guitar overdubs and panning: I did wonder if there was a second part in there, but wasn’t sure! :slight_smile: You can use a short (<30 ms) delay on one channel in addition to panning to increase the stereo width further (due to the Haas Effect).

I wonder if foldover distortion might be a bit harsh for the guitar processing: maybe try the TAP TubeWarmth or TAP Sigmoid Booster effects, which can be used fairly subtly.


Hi Chris!

I didn’t know there was a graphic EQ plugin for Ardour - I thought JAMin was more flexible and powerful, so I used it instead.
Do you recall the plugin’s name? I remember I saw TAP EQ <something>/BW LADSPA plugin in the manager. I can try to use it for drums, bass, and guitars to see how it goes.
And I agree, it wasn’t too convenient to ‘jamin’ particular tracks… :wink:

Regarding the ‘fizzy’ guitars - I don’t hear the ‘fizzyness’ too much when playing/recording one track, I feel it becomes more noticeable after recording the second guitar track, but (fortunately) only in a few parts of a song only.
I’ll try to move the second riff wave about 20ms to see what effect it has (I’m only not sure how to move a wave 20ms forward or backward in Ardour, but maybe I’ll find how :slight_smile: )
And the POD is the only choice I have - considering what kind of music I like, and having a wife and a baby around, I cannot afford to set up a ‘real’ environment :slight_smile:
But fortunately, I like my POD very much and plenty of different sounds it can give me - so either I’ll get rid of the ‘fizzyness’ somehow or just will live with it.

But I’m not sure how the EQ could be used to improve those fizzy guitars - did you mean an EQ in Ardour applied for particular guitar tracks, or in JAMin for mixed guitar tracks or maybe on the POD itself before the signal gets recorded?
And what is this ‘impulse convolver’ - is it a plugin or separate hardware to process the signal?

Sorry for so many questions, but I’m trying to get from the sound as much as I can (considering my environment, and not too wide knowledge in this area unfortunately).

You have helped very much so far, and I really appreciate it!
I owe you many beers now for sure! :slight_smile:


And if you could give those samples a try in your free time, it would be great! I think I already corrected much, but probably there’s still more to do better or in other way.

Hello daemonic_myst,

As a general comment, since I have not heard your samples yet, I would say that the pre-mastering mix is the most important part. Mastering is not exactly about loudness only.

I am no sound engineer by profession, but after 2.5 years of fiddling around with my own music, I realized that one should not make each single track (in a multi-track mix) too hot. After discovering the K-metering (as cited above), I found out that I was not on the wrong track (no pun intended) with my empirical approach. Get the mix right, it does not matter if it sounds too low. You want to achieve a nice dynamic and for that you need to leave some headroom for when you enter the mastering stage, you will need this headroom. Now, depending on your music genre, you can leave as much as 20dB (RMS), 14dB, etc.

Personally, since I am still in the composition phase, I do not care too much and still have hot-mixes (pre-master) that sound rather loud but when I am done with the composition, I will revisit EVERY mix so that I end up with a master headroom of about minimum 14dB. By headroom, I mean the gap between the master RMS level and 0dBFS. Only tyhen will I consider the mastering stage, where I’ll be able to exploit this headroom.

Note that peaks can be way higher than the RMS level. For that, the fast-lookahead limiter is a good tool. But I don’t want to sacrifice the mix dynamic. In fact, I don’t subscribe to the loudness philosophy, and on a rather aesthetic level, I like to see the waveforms showing something else than a big rectangle bar :lol:

Hi Thorgal!

Thanks for your comments.
By the master RMS level, do you mean an average signal level? In that case, if I lowered the master fader in Ardour to -8db, then I would have the 8db headroom - am I getting it right? I have found the following definition of RMS, but I admit it’s not clear for me (
But I’m still searching for a better description, so hopefully I’ll find something soon.

And by the pre-mastering mix did you mean to apply EQ (in Ardour) to every single track?


RMS is Root Mean Square, or standard deviation in physics jargon.

For digital audio signal, say your buffer has N frames (e.g. your jackd number of frames per period). You can use the discreet formula in the figure above to calculate the RMS for this buffer (or set of N frames). This value is between 0 and 1 in this case. You can then convert it to dB with a log10 based formula.

In general, the std deviation can be calculated for anything (apple size, stock value fluctuation over a period of time, etc).

For the attenuation on the master bus, I would think you can maybe remix your tracks so that you don’t have to attenuate massively on the master bus. That’s what I mentioned when I said that I avoid having hot tracks. To check your mix, send the post fader of the master bus to jkmeter (jack client widget that can be used for precisely this sort of things, see Fons Adriaensen’s webpage to get the source code) for example