Video summary

I Discovered Something Disgusting About Spotify

Main summary

Key takeaways

Technology

Overview

The video investigates a claimed myth about Spotify’s loudness normalization—a setting intended to make tracks play at a more consistent perceived volume across songs.

What Spotify Loudness Normalization Is (as described)

  • Spotify’s loudness normalization is described as ensuring tracks are played back at the same perceived loudness (i.e., the same “volume to the listener”).
  • It’s enabled by default.
  • The creator mentions a common claim from some audio engineers: turning it off should reveal the music closer to what the artist intended.

Experimental / Testing Approach

The creator runs controlled audio tests using:

  • Spotify Premium with lossless playback enabled
  • Recording the output using digital loopback into Reaper
  • Two captures of the same track:
    1. With loudness normalization on (Spotify default)
    2. With loudness normalization off, adjusted by 4.2 dB attenuation so both files should be the same loudness

Null Test Method

A null test is used to check whether the two files are actually identical:

  • One file is polarity-flipped
  • The flipped file is summed with the other
  • If the result is zero, the files match (meaning no audible differences)

Key Findings

  • Initial expectation: the files “should” null (be effectively identical) because it’s “just attenuation/volume control.”
  • Result: the first null test did not fully null as expected.
  • To investigate further, the creator increases gain dramatically (mentions boosting by about 70 dB) to expose differences.
  • Conclusion: loudness normalization does change the sound, at least slightly—contrary to the creator’s earlier skepticism.

Second Test: Volume Slider Effect

The creator also asks whether Spotify’s playback volume slider affects sound quality.

They run another null test comparing:

  • Spotify playback with loudness normalization off
  • Playback with custom gain applied (via a volume slider)

Result:

  • This second comparison also yields non-zero residuals.
  • The creator argues the residuals are likely due to testing methodology, such as:
    • a small volume mismatch
    • scaling effects
    • potential quantization/dithering noise
  • They emphasize that this is likely different from large, obvious distortion observed earlier.

Hypothesis About What Spotify Is Doing

The creator speculates Spotify may apply loudness normalization before encoding into the streaming codec—potentially following guidance similar to AES/ITU BS.1770-era recommendations (they mention “TD8” and AES recommending it in TD8).

  • They expect that with lossless enabled, artifacts/distortion shouldn’t appear.
  • Reading further into the referenced standard, they mention an alternative approach:
    • using metadata and the playback device’s native volume control
    • which, they argue, would be less likely to affect the actual audio content

Overall Conclusion + Recommendations

  • Objectively (as the creator interprets the tests): Spotify loudness normalization can change audio slightly.
  • The creator stresses that the differences are very small, but argues that tiny changes repeated across a production workflow can “add up” and degrade intended sound.

Recommendations

  • Don’t turn it off—keep loudness normalization on for more consistent listening.
  • If you truly care about audio quality, the creator says Spotify may not be the right platform.

Links / Callouts Mentioned

  • Mentions linking an “ultimate loudness tutorial.”
  • Mentions another video about a power supply (for unrelated “freak out” content).

Main Speakers / Sources

  • Primary speaker: the video creator/experimenter (speaking first-person throughout)
  • Technical references cited: AES / TD8 and loudness normalization guidance (referenced as standards documents)

Original video