PSOrvia

Learn · PSOrvia Player

What is missing after compressing: lossless and lossy

You hear the same piece lossless and in four levels of a lossy format, and you see in the difference what is missing.

Difficulty
medium
Duration
approx. 10 minutes
You need
PSOrvia Player

The audio example at the end

This is what the final state sounds like. Below it you can download the file in lossless form.

The audio example at the end

0:00

Keyboard: Space plays or pauses, arrow keys jump 5 seconds, Home and End go to start and end.

Lossless (FLAC) File in the house format (Opus)

A lossless format such as FLAC returns every value exactly as it was. A lossy format such as Opus throws away parts that the ear hears poorly, and therefore needs much less space. The lower the bitrate, the more is missing. Here you hear the same piece lossless and then at 96, 48, 24 and 12 kbit/s, encoded with the PSOrvia Opus encoder and converted back with the PSOrvia decoder. The “Difference” view shows you what is lost.

Headphones or good speakers. Each level is also available as a real Opus file to download; the player plays it. The audio examples are not brought to equal loudness: the lower the bitrate, the slightly quieter the piece gets, because parts are missing.

For comparison

Starting point: lossless

0:00

Keyboard: Space plays or pauses, arrow keys jump 5 seconds, Home and End go to start and end.

  1. Step 1: 96 kbit/s: hardly audible

    The starting point is lossless, every value as at rendering. The first level is an Opus file at 96 kbit/s, roughly a tenth of the data size of the original. Many people hear no difference here. The “Difference” view nevertheless shows that something is missing, above all very quiet parts and the topmost highs. You can test this yourself in the player: “Listen for approval” (Settings › Extras) plays versions switchably at the same spot and has an ABX test in which you guess eight times whether X is the same as A or B.

    In the player
    File formatOpus
    Bitrate when encoding96 kbit/s

    Audio example

    Hardly any difference from the lossless original. Anyone listening closely may find slightly less shine on the cymbals.

    The “Difference” view colors what is missing or added: at 96 kbit/s almost everything stays calm. “Wiper” shows both spectrograms side by side.

    Step 1: 96 kbit/s: hardly audible

    0:00

    Keyboard: Space plays or pauses, arrow keys jump 5 seconds, Home and End go to start and end.

    Exactly this file

    Listening review in the player with two waveforms one above the other, original and Opus version at 96 kbit/s
    “Listening review” in the player: the lossless original and the Opus version at 96 kbit/s as waveforms.
  2. Step 2: 48 kbit/s: the highs get thin

    Half the bitrate, half the data size. The encoder must now decide what it keeps, because the lossy method never gives back what was left out. It saves first on parts that are quiet or that sit next to louder ones. You see it in the spectrogram as areas getting darker in the highs.

    In the player
    File formatOpus
    Bitrate when encoding48 kbit/s

    Audio example

    The sound is still pleasant, but a little duller. Cymbals and glockenspiel lose shine.

    With “Wiper” you drag the dividing line across the spectrograms of 96 and 48 kbit/s: at the top right, in the highs, it gets darker.

    Step 2: 48 kbit/s: the highs get thin

    0:00

    Keyboard: Space plays or pauses, arrow keys jump 5 seconds, Home and End go to start and end.

    Exactly this file

    Listening review in the player with two waveforms one above the other, original and Opus version at 48 kbit/s
    “Listening review” in the player: the lossless original and the Opus version at 48 kbit/s as waveforms.
  3. Step 3: 24 kbit/s: veil and gaps

    At 24 kbit/s the encoder is left with only a small part of the data. It cuts off the high frequencies and reproduces the rest more coarsely. The difference is now clear: the piece loses shine, transients such as the hits get softer, and a veil settles over the sound.

    In the player
    File formatOpus
    Bitrate when encoding24 kbit/s

    Audio example

    Dull and somewhat smeared. The drum hits have less bite, the highs are almost gone.

    “Difference” now colors large areas; the “Spectrum” shows that the curve falls off above a few kHz. The “Waterfall” shows the missing highs as low hills.

    Step 3: 24 kbit/s: veil and gaps

    0:00

    Keyboard: Space plays or pauses, arrow keys jump 5 seconds, Home and End go to start and end.

    Exactly this file

    Listening review in the player with two waveforms one above the other, original and Opus version at 24 kbit/s
    “Listening review” in the player: the lossless original and the Opus version at 24 kbit/s as waveforms.
  4. Step 4: 12 kbit/s: only the framework is left

    At 12 kbit/s the limit is reached. For comparison: FLAC returns every value, but needs many times the data. The encoder narrows the bandwidth, keeps only the loudest parts, and the piece sounds as if through a telephone. For speech so little is often enough, for music it is too little. This is how you hear what the bitrate is good for: it is the price you pay for space and data volume.

    In the player
    File formatOpus
    Bitrate when encoding12 kbit/s

    Audio example

    Narrow and tinny. The bass is there, the highs are largely missing, and everything sounds somewhat tinted.

    The “Spectrum” breaks off above a few kHz; the “Difference” shows how much is missing. “Afterglow” shows during playback that the highs stay empty.

    Step 4: 12 kbit/s: only the framework is left

    0:00

    Keyboard: Space plays or pauses, arrow keys jump 5 seconds, Home and End go to start and end.

    Exactly this file

    Listening review in the player with two waveforms one above the other, original and Opus version at 12 kbit/s
    “Listening review” in the player: the lossless original and the Opus version at 12 kbit/s as waveforms.

Listening quiz

Can you hear which step is behind it? Listen to the three sounds and choose the one that belongs to the named step. None of this is saved.

1. Which sound results after step 2, at 48 kbit/s?

2. Which sound belongs to step 3, at 24 kbit/s?

3. Which sound is the last one, after step 4, at 12 kbit/s?

Continue