flat sound after ableton audio export

[DennisJ]

hello!
recently i decided to mash up heads will roll (atrak remix) and kidsos (wippenberg remix) to have something new instead of playing the atrak remix all the time.

however, after i exported the track and loaded it up in traktor, it sounded flat.



i don't know really what to do, i bought the .wav from kidsos but this didn't help either..

greets

[charo]

the volume or amp of the track you made needs to be brought up the level of your mp3 collection.

you can see in your picture that the wave amp of your song isn't as high as the other track.

i'm not sure about it in ableton but is there volume or normalize export options?

otherwise a program like audacity could do it.

[DennisJ]

yeah, there is a "normalize" option but that doesn't help..

maybe i also have to get a .wave of heads will roll? because that's a 320.

[mpetersen3]

it isn't about wav vs. 320, they are just different levels of accuracy of the waveform relative to the theoretical perfect version of the track, it shouldn't be affecting the average volume, most likely it is just that one track has been mastered higher than the other. What has happened is that since the two tracks volumes have similar peaks the peaks added (logarithmically) and made the combined track have louder peaks, as a result the main level was brought down to avoid clipping. To correct this you could use a compressor set to peak or transient detection and have it compress the loudest parts of the track so that the track can retain a higher level. However when this is done it will weaken the dynamics of the track, but this brings on a whole debate about mastering. So it's up to you to decide, but it could well help you to add a compressor with a high threshold and shortish attack to reduce your peaks, and then you can bring up your master level.

ps. The issue I mentioned above is that by making music louder we lose the subtle distinctions between loud parts and quiet parts. Think of how classical music has almost inaudible parts when compared with the crashing magnificence of its drops, while modern music is more of an avg. volume level. So you can make the quietest parts louder but at some point the track will sound even flatter.

p.ps on second thought too much compression will cause flatness, so if the peak parts of the track are sounding flat it might be due to the individual waveforms having phased frequencies that are destroying each other, thus resulting in flatness. You could try watching each track with a graphic eq. and then watching the mashup. If that is the problem you could try to make stems from each track and then figure out which frequencies from which stems are causing destructive interference and use a graphic eq to cut out some of the frequencies and make the mix sound less flat because of muddied bass, mids or highs.

[DennisJ]

wow, great post thank you!

another question which fits good to this thread:
when exporting a file i have some so many options and i really don't know what some of them are doing.

what is normalizing doing? is it clever to choose a sampling rate above 44100? what about the bit rate? i have a audio2dj soundcard, so would it be clever to chose 24bit/96khz? or even 32 bit?

greets

[mpetersen3]

normalize is a word in almost all math science disciplines that means 'to average'. The basis of normalizing is to give the sound an average level of loudness. However typical normalizing works via a method called rms (a mathematical operation for averaging useful for sinusoids where values are positive and negative, where squaring the value to be averaged prior to averaging removes the negative, just leaving the magnitude of displacement) (i believe although it can vary for normalization methods possibly).

but i digress, normalizing is averaging the volume, compression could achieve this like I mentioned above, but in a different route. Humans can only hear frequencies of up to about 20khz and the sampling frequency (the rate at which samples are taken) needs to be double the max frequency in your mix. This is why 44.1khz is used for cd's. Frankly I am not sure why they use 88.2khz sampling rate (I recall hearing that it had some use but I can't remember what, if anyone reading knows please inform me, mostly it is to satisfy the egos of rich audiophiles afaik). 24/96khz will be fine, the improvement to 32 is negligible to your ears. In short, none of those options need be changed significantly from their defaults and if the mix is sounding off it isn't because of them. most likely it is frequency clash. which can be fixed to a varying degree based on the amount of time and effort. My advice to you would be to put a graphic eq on the finished track and try to boost certain frequencies. even though for pro applications it would be much more advantageous to use my above mentioned method (make stems analyze frequencies and cut/boost pre mix to eliminate clash) this method will take you far longer and you can still get the track to sound pretty good by using a compressor on your master that is set to bring down peaks not rms, then boost volume (but not too much or the loud parts of the track will lose punch because everything was too averaged (let me know if you get that)) and also use a graphic eq to boost the frequencies that sound flat before exporting.

One more thing is that I'm not sure if you understand the notion of signal chain, but in short if you put the graphic eq prior to the compressor in the chain you will be compressing sounds you have boosted, or you can do it the other way around and do eq boosting/cutting after the compressor.

This might be too much for you if you don't have any experience with this stuff, but it will help to read and understand it, at the least try finding a graphic eq and playing around with it on your master bus. That should help make it more punchy. If you are feeling slightly ambitious (less so than compression) get a graphical display of the frequency spectrum (shows you the level of each frequency as the mix plays and look to see if there are any that are particularly low compared to those freqs on the individual tracks, if so this is due to phase cancellation, and boosting those COULD help).

ed. I hate how when I stare at the reply box for too long and then look up i have a white box in my vision lol
also if anyone sees any inconsistencies please correct me (or knows about uses of 88.2khz fs)

[Weezi]

^^^^
Props to this massive post, helped me think about my mixing techniques

[n2hf1st]

There's an easier solution... turn the gain knob up in traktor

[janzak]

The track to the right looks like a compressed version of the one to the left, look at the peaks after about 50% into the track.

The one to the left has a section peaking at about 60%, going into a section around 80% then 100% but the one of the right is smack 100% all the time - making it sound louder bringing us into the loudness debate etc etc

So the left track has more dynamics and contrast than the right one but the right one sounds louder

I dunno if this is because you've manually adjusted the volume inside your mashup or whatever caused it but you could compress the track to make it louder or crank up the volume (in Traktor) to keep the dynamics of the track.

Personally I would go half way with both - making it 50% louder and raising the volume 50%

[sj03w4t]

Dont normalize after mastering! Not before as well..