How can people create songs that are “slowed + reverb” and it sounds high quality, but when I try to slow down a YouTube video, movie, sports footage or anything else it sounds like a bleeping demon through the audio?
In: 2
It’s the same principle as enlarging a digital photograph. If the source material is high-resolution, the enlargement will show more detail you weren’t able to see. If it’s low resolution, you’ll end up with pixels or you can try to enhance the zoomed-in picture digitally, which makes your computer guess and subdivide the pixels into smaller ones. There are a lot of ways to do this guesswork (detecting edges, recognizing objects from a dataset) and it is an active research field.
Digital sound is usually captured by recording samples of loudness. These can be seen as sound pixels. The trick is to choose a high enough rate of recording to enable it to be recognizable as speech or music. For streaming media like youtube, the sample rate is usually set intentionally low to make sure your video loads quickly. The sound pixels themselves also contain fewer possible loudness values, like screen pixels on old computer would also contain fewer possible colors.
When trying to playing back the same number of samples in a different time frame than the original recording , two things can happen. The recording has to squeeze the samples when it has less time to play the recording. In your case it has to stretch out the samples. The pitch will then be perceived differently by your brain: slowing down will lower the pitch and speeding up will make it higher.
Luckily, your computer has ways of fixing this, like the picture enhancement mentioned earlier. For slowing down, the computer has to to correct the pitch so that it sounds slower and not lower. This will introduce little gaps that will make the recording sound choppy. The computer will have to guess what to put into thise gaps. The more information it has to work with, the easier the guesswork will be. Digital studio recordings have as much information as allowed by the recording devices and will be thus easier to enhance.
Streaming audio not so much, and which makes the computer guess that the gap probably needs a screaming demon more often.
Edit: Downvotes? Really?
The alogrithm that YouTube website player uses to slow audio is low quality.
If you use a higher quality timestretch alogrithm then you can achieve good quality results.
Starting from a high quality audio source helps too.
Latest Answers