4 ms·

Streaming services don't compress dynamic range, rather the opposite - because every streaming service employs loudness normalization the loudness wars have lar

Streaming services don't compress dynamic range, rather the opposite - because every streaming service employs loudness normalization the loudness wars have largely ended. Compressing your track to hell in an effort to get it to sound louder doesn't work anymore because every streaming service will just normalize it anyway. Spotify at least uses 320kbps vorbis for their "high quality" setting which is at the point where I don't believe it's humanly possible to distinguish from a lossless encoding.

PaulDavisThe1st 4y ago
> Compressing your track to hell in an effort to get it to sound louder doesn't work anymore because every streaming service will just normalize it anyway. This is a bit of confused remark. "Normalization" means applying some constant gain factor to the signal so that the loudest level is 0dBFS. Compression applies a varying gain factor to the signal to meet some parameters. Normalizing audio will generally make it louder (unless the loudest level is already at 0dBFS), but it does not change the dynamic range of the signal. The "loudness wars" were all about using compression, and this does change the dynamic range of the signal (you end up with less difference between the quietest and loudest parts of the signal - hence the term compression). You can still "sound louder" by using compression, even if the peak volume is still 0dBFS.
- initplus 4y ago
  Modern volume normalization isn't done based on the highest peak of the track. Instead they normalize based on the average perceived loudness of the whole track (to a level below 0 so there is headroom for peaks). It's intentionally designed to avoid the exact issue you describe. If this was not done users of streaming services would have to be constantly adjusting the volume to deal with perceived volume differences between tracks due to different levels of dynamic range compression.
  - _carbyau_ 4y ago
    I don't have to change the volume between shows. But for a while, every action show required I down volume the noisy bits and up volume the talking bits. I understand dynamic range is a thing and people with reasonable stereos and a place to use them should enjoy. But in an apartment with a sleeping toddler I simply want to hear it without waking my kid. I now run: Windows Loudness Equalisation, Clear Voice on the sound bar, every show is subtitles on.
  - PaulDavisThe1st 4y ago
    Sure, peak-to-N-dBFS is not what the streaming services do. But what they do is not normalization. I'm the author of a DAW that allows normalization, compression and loudness control. The first one is more different from the last two than they are from each other.
- whyoh 4y ago
  >"Normalization" means applying some constant gain factor to the signal so that the loudest level is 0dBFS. Not necessarily, although that has been perhaps the most common use of that term so far. But you can normalize to a lower peak level or to something else... https://en.wikipedia.org/wiki/Audio_normalization https://en.wikipedia.org/wiki/Audio_normalization
Nav_Panel 4y ago
> the loudness wars have largely ended I've been digging up a lot of remastered 80s albums from the last 5-7 years (mainly Japanese stuff), and I can 100% tell you that this is not the case. The engineers may not be sucking every last decibel out of the tracks anymore, but they're still compressed to hell, dynamic range often 5-10dB less than the original version straight off the 80s-era CD. If you think about the way normalization works, there's still benefit to slapping a master limiter on top of it and pushing it to the point where normalization doesn't require any gain, because at least then you have control over the final mastering chain. But I don't think this is the reason, and I don't really know why they still juice it hard. Maybe for car stereos, or because phone amps are weak, or because they're not thinking about Spotify specifically.
brokenmachine 4y ago
>because every streaming service employs loudness normalization the loudness wars have largely ended This is a joke, right? Although I applaud their efforts, it has just created a second loudness war on how to maximize loudness despite the normalization.

PaulDavisThe1st 4y ago

> Compressing your track to hell in an effort to get it to sound louder doesn't work anymore because every streaming service will just normalize it anyway. This is a bit of confused remark. "Normalization" means applying some constant gain factor to the signal so that the loudest level is 0dBFS. Compression applies a varying gain factor to the signal to meet some parameters. Normalizing audio will generally make it louder (unless the loudest level is already at 0dBFS), but it does not change the dynamic range of the signal. The "loudness wars" were all about using compression, and this does change the dynamic range of the signal (you end up with less difference between the quietest and loudest parts of the signal - hence the term compression). You can still "sound louder" by using compression, even if the peak volume is still 0dBFS.

initplus 4y ago
Modern volume normalization isn't done based on the highest peak of the track. Instead they normalize based on the average perceived loudness of the whole track (to a level below 0 so there is headroom for peaks). It's intentionally designed to avoid the exact issue you describe. If this was not done users of streaming services would have to be constantly adjusting the volume to deal with perceived volume differences between tracks due to different levels of dynamic range compression.
- _carbyau_ 4y ago
  I don't have to change the volume between shows. But for a while, every action show required I down volume the noisy bits and up volume the talking bits. I understand dynamic range is a thing and people with reasonable stereos and a place to use them should enjoy. But in an apartment with a sleeping toddler I simply want to hear it without waking my kid. I now run: Windows Loudness Equalisation, Clear Voice on the sound bar, every show is subtitles on.
- PaulDavisThe1st 4y ago
  Sure, peak-to-N-dBFS is not what the streaming services do. But what they do is not normalization. I'm the author of a DAW that allows normalization, compression and loudness control. The first one is more different from the last two than they are from each other.
whyoh 4y ago
>"Normalization" means applying some constant gain factor to the signal so that the loudest level is 0dBFS. Not necessarily, although that has been perhaps the most common use of that term so far. But you can normalize to a lower peak level or to something else... https://en.wikipedia.org/wiki/Audio_normalization https://en.wikipedia.org/wiki/Audio_normalization

Nav_Panel 4y ago

> the loudness wars have largely ended I've been digging up a lot of remastered 80s albums from the last 5-7 years (mainly Japanese stuff), and I can 100% tell you that this is not the case. The engineers may not be sucking every last decibel out of the tracks anymore, but they're still compressed to hell, dynamic range often 5-10dB less than the original version straight off the 80s-era CD. If you think about the way normalization works, there's still benefit to slapping a master limiter on top of it and pushing it to the point where normalization doesn't require any gain, because at least then you have control over the final mastering chain. But I don't think this is the reason, and I don't really know why they still juice it hard. Maybe for car stereos, or because phone amps are weak, or because they're not thinking about Spotify specifically.

brokenmachine 4y ago

>because every streaming service employs loudness normalization the loudness wars have largely ended This is a joke, right? Although I applaud their efforts, it has just created a second loudness war on how to maximize loudness despite the normalization.