I have couple of questions.
I have 100mins of 16bit 44.1 PCM WAV audio I want to add into a DVD, so obviously I have to change the sample rate from 44.1 to 48000 for DVD, however I also need to slightly reduce the volume as when I open the 44.1 file in Audacity and simply export the file at 48000, I get results of 11 and 8 possibly clipped samples for each channel.
So normalizing the file to a lower dB (with DC Offset selected) should stop it from clipping once converted to 48000 (my first question, is it OK to use Audacity’s normalize function to change to a lower volume?, I would usually use amplify function, but thought I’d use normalize to use the DC Offsett feature also).
My second question, does it matter if I change the sample rate box at the bottom left corner from 44.1 to 48000 before I normalize + DC Offset, does it make any difference if I normalize + DC Offset when the box is still 44.1, then only change to 48000 as the last step before I export the WAV?, I’m pretty certain this will not make any difference at all?
Are you doing anything else to the sound in Audacity or just changing the sample rate?
If only changing the sample rate, have you checked to see if your DVD authoring application can change the sample rate? (some apps do this automatically when required).
Yes you’re right, though I’m using Muxman to combine video (m2v) and my new WAV PCM audio together, and it’s not accepting my PCM WAV as 44.1, unless I’ve already changed it to 48000.
So yes all I really need to do is change the 44.1 file to 48000, but I have the possible clipping problem as described in my first post (please also see my updated first post about the normalizing question).
I guess I should also ask the question is the clipping indicated really going to be a problem or should I just leave it?
For example the original 44.1 file has a peak volume of -0.05 / -0.66 dB with no possible clipped samples, once opened in Audacity and simply exported at 48000 with nothing else done the output 48000 file now has a peak volume of 0 / 0 dB with possibly clipped samples of 11 / 8.
Is it even worthwhile bothering to lower the peak dB volume before converting to 48000?
A very few samples like that will not be noticeable at all. In fact, they are probably not actually clipped. The red “clip lines” are “warnings” not “errors”. They indicate that the waveform touches or exceeds 0 dB. In your case it seems that occasional samples are touching 0 dB. It’s like redlining your car - the engine does not immediately explode, but if you push it too far or too often then there will be noticeable damage.
I actually got those clipping statistics from “analyze/statistics” in Adobe Audition, as I wasn’t sure where Audacity shows that info.
Going back to 2 previous questions I hope to clear up, does it matter if you don’t change the project rate at bottom left from 44.1 to 48000 until after normalising and DC Offset?, or does it have to be changed before normalising + DC Offset?
Also does normalizing to a lower peak volume than the original file work correctly? eg. changing the peak volume of -0.02dB to a new peak of -0.4dB?
The reason I ask is there is also an ‘amplify’ function, but I used normalize as at the same time I could correct some very slight DC Offset.
There is no difference which you do first.
In both cases the resampling occurs when you export to a new file. Note that for safety it is highly recommended that you export to a new file rather than overwriting the original. When the job is all done and tested, then you can delete any files that you don’t need any more.
Yes you can just Normalize to -0.4 with DC Offset enabled. When the peak level is calculated it allows for the DC offset correction (when DC offset correction is enabled).
Thank you for the info Steve.