Rendered Song Is Low Volume And Has Strange Dc Offsets

Help, Support & Bugs
1

I find the volume to be noticeable lower on a rendered song than on other “generic” songs i have in digital format on my computer. The Renoise master level meter is sometimes reaching clipping indication although i try to avoid it, since i guess it’s common practise. But how do i make the song louder without clipping and distortion? I run the ‘sox’ utility in linux to get some statistics about my wav and an other track that is loud but not distorted in some way.

my_file.wav:

  
 Overall Left Right  
DC offset -0.048049 -0.047935 -0.048049  
Min level -1.000000 -0.966156 -1.000000  
Max level 0.683258 0.678162 0.683258  
Pk lev dB 0.00 -0.30 0.00  
RMS lev dB -10.42 -10.42 -10.41  
RMS Pk dB -7.74 -7.81 -7.74  
RMS Tr dB -29.75 -29.75 -28.99  
Crest factor - 3.21 3.31  
Flat factor 4.44 0.00 6.02  
Pk count 16.5 11 22  
Bit-depth 16/16 16/16 16/16  
Num samples 1.66M  
Length s 34.500  
Scale max 1.000000  
Window s 0.050

other_song.wav:

  
 Overall Left Right  
DC offset -0.000039 -0.000032 -0.000039  
Min level -1.000000 -1.000000 -1.000000  
Max level 0.999969 0.999969 0.999969  
Pk lev dB 0.00 0.00 0.00  
RMS lev dB -9.46 -9.54 -9.38  
RMS Pk dB -5.48 -5.62 -5.48  
RMS Tr dB -184.13 -184.13 -182.23  
Crest factor - 3.00 2.94  
Flat factor 8.63 6.14 10.00  
Pk count 12.9k 10.5k 15.3k  
Bit-depth 16/16 16/16 16/16  
Num samples 26.9M  
Length s 609.080  
Scale max 1.000000  
Window s 0.050

Although all these numbers dont tell me much i can clearly se differences.

The other strange things is how the sound representation looks in audacity, not centered at all. I have checked all my samples used and the are centered in the sample editor of Renoise:

Is this bad? What is it and what did i do wrong?

Thanks for any help.

2

Renoise uses 6db headroom for rendering (this explains the softer output). If you don’t desire the headroom, add a gainer on the master and raise it +6db.
The DC offset i have no idea, are you pulling your material through effects on the master-track? that one of these effects are balancing out the output on the DC equator is one cause i could think of.

The other would be wrong recorded samples through bad shaped equipment or bad earth-wired recording lines.
All are just speculations though.

3

I tried adding gainer to master channel while rendering. +6 db gave me a distorted file, +3db sounded better.

All samples are “recorded” into renoise sample editor through JACK and vst effects. That’s the only thing I can think of causing it, but i dont know why and why it’s bad. It doesnt even sound strange, just looks bad?

4

To be honest, I am not quite sure why you have such a massive DC offset!

I can, however, tell you a few things I learned about volume etc…

I am not quite sure what everything in your readout means. I monitor my levels in real time using RN digital inspector, I now try to get my songs to hover between 10-12rms. RMS a is much more accurate guide to perceived loudness than peak. Most dance, pop, and rock songs today are considerably louder than 10-12rms…sometimes as loud as 3 rms. I used to master my songs to around 6rms and after some reasearch, I changed my mind and decided to focus on clarity and punch as opposed to sheer volume. A well mixed and punchy song at 12 rms may sound just as loud and better than a muddy one at 9rms.

If you do want your songs as loud as commercial tracks, simply use the maximizer in Renoise.

This will ensure that your songs don’t exceed 0db while boosting quiter elements of your track, thus increasing perceived volume(rms)

Leave the threshold at .0020/ceiling at .0000(unless of course you want to experiment or hear good advice telling you otherwise) and boost to taste. Be wary though, not to apply more than 3db an instance as it may noticably degrade sound quality. Many use standard compression (maximizing and limiting are extreme forms of compression)to aid this process and get good results. I prefer to keep things simple, when I used to make loud masters I would use one instance at .0020, no boost, to smooth over peaks, than another w/ 3 db boost, than maybe another for an extra 2b…

More important than loudness is quality of your mix. Important factors such as punch and clarity can have a huge effect on how your songs compare to your commercial/professional counterparts. Get rid of the low end on anything that is not a kick drum or sub/bass…It depends on what element you have chosen to be the lowest element in your track.

There is a LOT more to be said on the subject of mixing/ mastering, and I am sure many will have other/different things to add.

Just to summarize, however, if you want to get a song as loud as a track you are comparing it to…open two instances of renoise…

In the first, open your song. In the second, an mp3/wav of the track you are comparing to(put a gainer on the master[this instance only] set to 6db boost to compensate the automatic 6b lost in imported audio files in Renoise)

Have an rms reader open in the master track of both Renoise instances…use the maximizer to get your track to the same rms as the reference track.

Hope this helps!!

BTW…DC offset may not sound bad…it does, however, rob you of headroom as one side of your waveform will be using most of the available headroom!

EDIT: I would be remiss not add that when assessing loudness readings using an rms tool, factors like sub-bass and low end can hinder percived loudness! Make sure your mix is balanced as low end esp sub bass tends to eat up headroom more than other frequencies I find…so you could have two songs at 10rms and one still sound quiter because it is not balanced correctly…again back to mixing…

5

I added an EQ to master channel and tried to -20 dB everything below 20 Hz. I also added a Maximizer with +6 dB boost.

It now looks like this. Looks better, but still a little bit strange with more concentrated dark blue areas on the lower part than on the upper part.

Will i be fine with this result?

6

I think it is not a good way to mix a track after a “good” looking waveform. Especially if you start making tracks.
Train your ears. Compare your tracks with good sounding tracks, you like.
A good balanced track often dont need a “massive” mastering stage.
Take care about every instrument and its place in the whole. You can first start with a minimum limit of instruments and place it as good you can in the mix. With training you will mix better.

Have a look at the waveforms after you have made a good step forward in mixing.

Nothing is finally, everything can be.

This is what helped me.

7

Bambooli is absolutely right…don’t worry about the waveform…use your ears, compare your track directly with a track you like in the style you are making.

He also said a well balanced track does not need a massive mastering stage, this is correct and a variation of what I said earlier.

I am currently working on one of my first tracks to use my new approach of low rms/massive punch and clarity…it works…

I even compared it some loud commercial tracks around 3-4 rms…my track hovers between 10-12 and still sounds almost as loud because I have taken care to mix for clarity and punch…perceived loudness is just that…perceived…a lot of what is raised are background sounds…it sounds louder and more “happening” at first…

If you have taken care to imbue your piece with punch and dynamics, yours will sound better, even when playing the other track at its louder volume…you will especially notice the difference if you play both tracks at the same rms…the other track will reveal its true squashed nature.

@eqing below 20hz…the filter is a much better tool imo for this…if you want to tidy up freq on the master, use the butterworth hp filter to cut below 20hz.

It won’t help that much, however, if you have tons tracks with low freq that you don’t need! ie…hi hats…can be quite liberal with cutting these as long as it doesn’t detract from the sound…almost every sound in a mix can use a hp filter…

sounds can also have unneccessary highs as well, I use lp filters all across the mix…are u starting to get the picture?

think of your piece as a canvas in physical space…every sound uses a part of this space…once a space is used up, paint somewhere else…ie use a different frequency or even panning etc…

8

Thank You, Tarek-FM! This gave me quite some inspiration ;)

9

Maybe I wasn’t clear enough with that I ofcourse not just look at the waveforms. But as it looked so unbalanced it made me unsure about if what I thought sounded “pretty OK” was totally wrong.

Thanks for all guidelines. I will start using lp/hp filters on all channels where possible to make instruments just play in the spectrum of frequencies as they “should”.

What I now start to understand is that i need a pair of monitors instead of using DJ headphones.

For anyone interested in how the part of the waveform first posted sounds like:

http://soundcloud.com/krixa/work-in-progress-messing-around-2

10

Look at your waveform, both before and after you added the EQ. Unfortunately you have not put it any time information for me to reference too.

Intro, even above and below zero. First change, still balanced. Main part kicks in and your waveform becomes skewed. What sounds do you bring in here? One of them is quite obviously the culprit.

As mentioned it is good practice to both high pass and low pass all sounds to cut out unwanted frequencies and keep your mix as clean as possible. I would guess it’s something rogue in your bass and a high pass filter (30Hz is a figure often used at mastering stage for DnB and that’s a more sub-bass heavy genre than most) may solve the DC offset.

11

track devices -> dc offset -> auto dc

12

try that or check if one of your samples has an dc offset. kick/bass, snare, clap, hihat, etc.

13

Ha! You are very welcome. Foo? did a lovely article on using Renoise filters to sculpt your mix, Beatslaughter also offered an intro to filters for beginners…filters

For anyone who is interested, check it out if you are not already aware of the benefits of filters(mixwise)! :walkman: