hey friends! not sure what else to do except turn to the board of wisdom for an answer. thanks in advance for your help

i have a dubstep tune that is 145 bpm (finished audio file, not in a program) that needs to play at 150 bpm permanently. using it in a premade custom edit with another track that is 150 bpm

I used fl studio to resize the song, and it is indeed the proper speed. it plays back perfect with my set and the other 150 track. but for some reason because i am using the time stretch tool in fl studio to adjust the songs speed, the volume spikes way over 0 db. Normalizing doesnt help and turning on a limiter makes it fairly quiet. It sounds like its the right volume when its peaking, which is even weirder. I cant find out how or where it is auto-gaining the track during the sample stretch.

Anyone else permanently change their song bpms? Even in other programs like ableton or audacity? i'd love to hear other methods of changing an exported tracks speed. Thanks guys!!

 

Have you tried the different algorithms?
What limiter are you using?

 

yeah, tried stretch and e3 to maintain the pitch. both produce the same result of increase in master volume as soon as the time knob gets changed

fabfilter, FL, and LA xlimit for limiters. thing is, i shouldnt need a limiter at all! it should just maintain the original audio file volume

 

Download REAPER [DAW] (it's technically free) and use that to change your tempo.

 

E2 transient is worth a shot, my first impression of that algo seems to me that is placets the transients at the original volume.

What it should do and what you want are 2 different things, essentially what it does is formant shifting.

Without going to deep ( I myself do not grasp the full ‘picture’.)
Formant shifting is making potatato soup of the frequency balance.

 

I tried this with Reaper by adjusting the playback rate in the audio file's settings to 1,034 (equals 145 -> 150) and found two very interesting things.
Firstly, the peak level increased by more than 2dB!
Secondly, with DMGAudio's TrackLimit it was chopped accordingly, no more clipping, but this didn't have any effect on the LUFS! ???

Tbh, I'm confused by both points.

 

Just tried, clips the same way.

 

That's exactly the difference between peaks and LUFS.

When you change the tempo of an audio file, it also gets resampled implicitly. Resampling introduces artifacts (or "noise") and if the original audio is normalized to (almost) 0 dBFS, these artifacts lead to clipping. But since this noise doesn't affect the perceived loudness, the LUFS value stays the same.

So if you don't want the signal to clip, you either need to reduce the gain before changing tempo, or limit the signal again afterwards.

 

This will lower the track's perceived volume significantly and you'll have to raise it afterwards again and add a limiter to catch the clipping. So, no win, but both ways work.

I'd change the tempo add a limiter with threshold and output set to the original peak (at least this is the way TrackLimit works), render it and you're done.

 

it's technically free - To Try not to use permanently.

 

I think what's happening is that Fruity's timestretch is affecting the volume, and it's realtime so the normalize function is not seeing the peaks properly. If you turn down the stretched track a few db, export to wav, and then normalize the Wav in audacity or whatever editor you like, it should be just fine.

 

that could be a very good explanation

 

This is on principle the same that @Olaf said and does not work. At least not in Reaper.

 

What's funny @celtic3342?

 

Yes, of course. If you reduce the gain, it will be quieter. But it won't clip and you can use a brickwall limiter to raise it again (and take care of the peaks). That's the idea behind this. There might be software that automatizes it – i.e. with a build-in limiter – but I don't know.

 

thanks for all the input guys!!

 

Yeah. Don't be afraid to lower the volume. Lowering and then raising the volume back up does not effect the quality of the audio (if you don't clip it first); it will be the same (it's digital). Normalizing - on the other hand - does affect the quality of the audio.

 

Last edited: Aug 18, 2019

The clipping is the result of an artifact caused by the time stretching. You could prevent it by using pitch shift instead, but that would obviously change the pitch and it wont fit in your mix as nicely.