Formats and Quality

Formats and Quality

All formats are defined centrally in config.py. This page explains what the individual entries mean and when each one makes sense.

---

Audio formats

Selection	Extension	Codec	Bitrate	When it makes sense
MP3 (128 kbps)	.mp3	libmp3lame	128k	Speech, podcasts, limited storage
MP3 (192 kbps)	.mp3	libmp3lame	192k	Good compromise for mobile use
MP3 (256 kbps)	.mp3	libmp3lame	256k	Default, fine for most purposes
MP3 (320 kbps)	.mp3	libmp3lame	320k	Maximum for MP3, largest files
AAC (256 kbps)	.m4a	aac	256k	Apple devices, iTunes, better than MP3 at the same bitrate
OGG Vorbis (320 kbps)	.ogg	libvorbis	320k	Free format, good quality
WAV (Lossless)	.wav	pcm_s16le	uncompressed	Further processing in editing and DAW software

---

Video formats

Selection	Extension	Resolution	Note
MP4 (1080p)	.mp4	up to 1080p	Default, plays everywhere
MP4 (720p)	.mp4	up to 720p	Roughly half the size of 1080p
MP4 (480p)	.mp4	up to 480p	For small screens and tight storage
MP4 (360p)	.mp4	up to 360p	Smallest variant
MKV (Best)	.mkv	highest available	Best tracks, including 1440p and 2160p
WebM (Best)	.webm	highest available	VP9 and Opus, smaller at comparable quality

The resolution values are upper bounds. If the requested step does not exist, the next best one is used automatically.

---

Why there is no FLAC

The Spotify downloader sources its audio from YouTube Music. That is a lossy source at roughly 256 kbps in AAC or Opus. Converting to FLAC would not recover the discarded information, it would only wrap a lossy signal losslessly. The result would be three to four times the size of the source and sound identical to a good MP3.

The comment in config.py puts it this way:

Source is YouTube Music (max ~256 kbps AAC/Opus), lossless formats are intentionally excluded

WAV is no quality gain for the same reason, but it is available because some programs require WAV as an input format.

---

What happens technically

Audio

1. yt-dlp downloads with the selector bestaudio/best, so the best available audio track.
2. The FFmpegExtractAudio postprocessor converts to the target codec.
3. preferredquality sets the target bitrate. For WAV the value is omitted, since PCM has no bitrate in the usual sense.
4. Optionally further postprocessors follow: EmbedThumbnail, FFmpegEmbedSubtitle, SponsorBlock plus ModifyChapters.
5. In the Spotify downloader, mutagen then writes the tags from the Spotify API.

Without FFmpeg, step 2 is skipped and the file stays in its source format.

Video

yt-dlp downloads the video and audio tracks separately in the best matching versions and FFmpeg muxes them into the target container. No re-encoding takes place, the tracks are only repackaged. That is why a video download is usually much faster than an audio conversion.

---

Metadata per format

converter.embed_metadata() picks the right tagger based on the file extension:

Extension	Library	Tag system	Cover art
.mp3	mutagen.id3	ID3v2.3 (TIT2, TPE1, TALB, TDRC)	APIC frame
.flac	mutagen.flac	Vorbis comments	FLAC Picture
.ogg	mutagen.oggvorbis	Vorbis comments	Base64 encoded metadata_block_picture
.m4a	mutagen.mp4	iTunes atoms (©nam, ©ART, ©alb, ©day)	covr
.wav	mutagen.wave	ID3 inside the WAV container	APIC frame

All cover art is embedded as JPEG. If embedding fails, the file is kept and the download still counts as successful, the error is silently ignored.

---

Recommendations

Use case	Recommendation
Music for phone and car	MP3 (256 kbps)
Music library on a computer	MP3 (320 kbps) or AAC (256 kbps)
iPhone, iPad, Apple Music	AAC (256 kbps)
Podcasts and audiobooks	MP3 (128 kbps)
Editing, DAW, sampling	WAV (Lossless)
Video for archiving	MKV (Best)
Video for sharing	MP4 (1080p)
Video with tight storage	MP4 (480p)