Audio Quality Standards for Streaming

The safest streaming master is a native, lossless source file with the original sample rate and bit depth preserved, clean stereo channels where stereo delivery is required, and enough true-peak headroom to survive lossy encoding. There is no single loudness number, file format, or technical profile that governs every major streaming platform and every delivery route.

The most important distinction is between three different things:

1. Measurement: how loudness and peak level are calculated.
2. Playback normalization: how a platform changes playback level for listeners.
3. Delivery specifications: the source files and technical profiles a platform or delivery partner accepts or recommends.

Those categories are related, but they are not interchangeable. A platform’s playback reference level is not automatically an ingest requirement. A broadcast loudness recommendation is not automatically a music-streaming target. A platform’s preferred source format does not mean listeners receive that exact file unchanged.

The basic terms: LUFS, true peak, and sample peak

LUFS is a loudness measurement unit used with standardized loudness algorithms. It helps describe the perceived level of a complete program, track, or section of audio rather than the height of one isolated waveform sample. When a specification refers to integrated loudness, it is describing an average measurement over the relevant program or track.

Sample peak is the highest value represented by the individual samples in a digital audio file. It is useful, but it does not show every peak that may occur after digital-to-analog conversion, resampling, or lossy encoding.

True peak estimates inter-sample peaks: levels that can occur between the stored samples. This matters because encoding or other processing can create peaks between samples even when the file’s sample peaks appear to remain below a ceiling. The current in-force ITU recommendation for algorithms measuring program loudness and true-peak audio level is BS.1770-5. It defines measurement methodology; it does not establish one universal loudness target for all interactive music services. For metering, use a BS.1770-compatible loudness and true-peak workflow, and do not treat a channel-peak ceiling as a universal platform rule. BS.1770: Algorithms to measure audio programme loudness and true-peak audio level

The practical takeaway is simple: measure integrated loudness and true peak separately. A track can have a similar integrated LUFS value to another track while producing different true peaks, and two tracks with similar true peaks can have very different perceived loudness.

Spotify: source delivery and playback normalization are separate

Spotify strongly prefers FLAC and also accepts technically equivalent WAV files that meet its requirements. Spotify requires native stereo delivery at 44.1 kHz or higher. Its documentation says to preserve the native sample rate and bit depth of the master. A 24-bit master should remain 24-bit; 16-bit is allowed when no higher-bit-depth master exists. Do not make a separate 16-bit version by reducing a 24-bit master merely to fit an assumed platform preference. Audio file formats for Spotify

Spotify performs internal conversion and playback optimization, so the delivery file is not the same thing as the file a listener necessarily receives. The goal is therefore to provide a clean, technically valid source rather than to pre-compress the master for the service.

Spotify also documents loudness normalization during playback. It states that it adjusts tracks to -14 dB LUFS according to the ITU 1770 standard, subject to playback-context limits. The web player and some third-party devices do not use loudness normalization. In other words, -14 dB LUFS is a documented playback reference in Spotify’s normalization behavior, not a universal ingest rejection threshold for all music delivered to Spotify. Loudness normalization on Spotify

Spotify’s mastering guidance recommends targeting -14 dB integrated LUFS for a master and keeping maximum true peak below -1 dBTP for lossy formats. Spotify additionally says that louder masters should keep true peak below -2 dB to reduce encoding distortion. These are recommendations, not stated universal acceptance rules. They should be treated as a platform-specific mastering choice, not as proof that every service requires the same loudness target. Loudness normalization on Spotify

A useful Spotify decision is whether the master is being optimized for level, dynamics, or both. If the track is mastered louder than Spotify’s normalization reference, Spotify may apply negative gain during normalized playback. If it is quieter, Spotify may apply positive gain, depending on the playback context. A louder master can therefore lose its level advantage under normalization while still carrying more risk of encoding distortion. The exact listener experience also depends on whether normalization is active on the device or player.

Apple: ordinary stereo, Apple Digital Masters, Hi-Res Lossless, and immersive audio are different profiles

Apple’s documented source profiles cover more than one type of delivery. For ordinary stereo source audio, Apple accepts 44.1, 48, 88.2, 96, 176.4, or 192 kHz sample rates with 16-bit or 24-bit resolution, using approved WAV, ALAC, CAF, or FLAC formats. These are source-profile requirements within Apple’s delivery ecosystem, and special profiles can impose additional conditions. Apple Video and Audio Asset Guide

That list should not be interpreted as a command to upsample every project to 192 kHz. The appropriate source is the highest-quality native master that actually exists. Artificially changing sample rate or bit depth does not recreate information that was absent from the original production.

Apple Digital Masters is a separate program profile. Apple Digital Masters content requires 24-bit audio in an approved format and accepts the listed sample rates from 44.1 through 192 kHz. Apple advises against upsampling or bit-padding lower-resolution sources. These requirements relate to Apple Digital Masters badging and program eligibility; they do not mean that every Apple Music stereo asset must qualify for that program. Apple Video and Audio Asset Guide

Hi-Res Lossless is also a separate source profile rather than a universal instruction for every release. When a delivery route calls for a higher-resolution lossless asset, use the native high-resolution master and confirm that the approved format and resolution match the route’s profile. Do not assume that a special Apple profile changes the loudness behavior of ordinary stereo playback.

Immersive audio has its own documented technical and loudness conditions. Apple’s immersive-audio profile specifies integrated loudness not exceeding -18 LKFS and true peak not exceeding -1 dBTP, measured using ITU-R BS.1770-4. This is an immersive and surround delivery requirement; it does not establish an Apple Music stereo loudness target. Apple Video and Audio Asset Guide

The operational lesson is to identify the Apple profile before mastering or exporting. A stereo master, Apple Digital Masters asset, Hi-Res Lossless asset, and immersive mix may require different validation steps. One number cannot safely stand in for all of them.

YouTube: distinguish partner delivery from ordinary uploads

YouTube’s official documentation describes multiple delivery contexts. For eligible label and distributor partners using YouTube Content Manager and DDEX ERN, FLAC or uncompressed WAV is preferred. A 320 kbps MP3 file is accepted as a fallback, and YouTube transcodes delivered audio. The documentation limits these features to that partner delivery route, so they should not be presented as universal requirements for every YouTube upload or every YouTube Music release. File format for resources

For YouTube music-video partner delivery, YouTube documents FLAC or linear PCM, recommends 48 kHz, recommends 24-bit while accepting 16-bit, and supports stereo delivery. Lossless source is preferred over a re-compressed lossy source because compressed delivery can degrade the result after further transcoding. This is a music-video delivery profile for Content ID music-video partners, not a documented YouTube Music loudness target. Encoding specifications for music videos

If you are preparing a music-video delivery package, 48 kHz and 24-bit are sensible defaults when they match the native source and the partner profile. If the native master is 44.1 kHz or 16-bit, do not invent quality by upsampling or padding. Confirm the exact route and preserve the best genuine source available.

The broader pattern is consistent across the platforms: send lossless or uncompressed source whenever the documented route prefers it, avoid unnecessary lossy conversion before delivery, and verify whether the specification applies to a partner feed, a particular content profile, or a general consumer upload.

Loudness targets that should not be treated as universal

Many mastering discussions repeat a single target for every service. Current authoritative documentation does not support that shortcut. Spotify documents a playback normalization reference around -14 dB LUFS and provides mastering recommendations tied to that behavior. Apple documents a separate immersive-audio loudness and true-peak profile, but that does not establish a stereo Apple Music target. The reviewed official YouTube sources document file and encoding specifications, not a general YouTube Music loudness target.

Broadcast guidance is another separate context. EBU R 128 describes broadcast loudness-normalization practice and gives -23 LUFS with stated tolerances. That can be useful when working in broadcast, but it should not be presented as a universal target for interactive music streaming. Applicability to a specific music service must come from that service’s own documentation. Loudness

Accordingly, do not turn commonly repeated Apple Music, YouTube Music, or Tidal loudness numbers into firm article claims without current primary evidence. A loudness number can be a mastering preference, a playback normalization reference, a profile-specific requirement, or a broadcast convention. Those are materially different uses.

A practical delivery workflow

Start by identifying the delivery route. Ask whether the release is ordinary stereo, Apple Digital Masters, Hi-Res Lossless, immersive audio, a YouTube partner music-video package, or another profile with its own documentation. If a distributor is submitting the release, check the distributor’s current technical intake rules as well as the platform profile because the delivery path may add conditions.

Next, locate the native master. Record its sample rate, bit depth, channel format, and file type. Preserve those properties where the platform permits them. Do not downconvert a 24-bit master to 16-bit simply because 16-bit is accepted, and do not upsample a lower-resolution source to suggest that it is high resolution.

Then create a lossless delivery file. For Spotify, FLAC is preferred and qualifying WAV is accepted. For Apple, use one of the approved formats for the relevant source profile. For eligible YouTube partner delivery, prefer FLAC or uncompressed WAV; for music-video partner delivery, use FLAC or linear PCM where available. Avoid sending a lossy file when a lossless source can be delivered.

Measure integrated loudness and true peak with a compatible meter. Keep the measurements attached to the correct profile. Spotify’s -14 dB LUFS reference and true-peak guidance describe Spotify playback and mastering recommendations. Apple’s -18 LKFS and -1 dBTP limits apply to the documented immersive-audio profile. Neither should be copied into an unrelated stereo or YouTube workflow.

Finally, audition an encoded preview if your mastering process supports it. This is especially useful when the master is loud, because Spotify warns that louder masters should keep true peak below -2 dB to reduce encoding distortion. The preview is a quality-control step, not evidence that a particular loudness number is universally required.

A compact route map looks like this:

Native master → identify profile → preserve genuine resolution → export lossless source → measure profile-specific loudness and true peak → check encoding risk → deliver through the correct route.

The central standard is not “master everything to one number.” It is disciplined separation: measure with a recognized method, deliver the best valid source, follow the requirements of the actual profile, and treat playback normalization as playback behavior rather than a universal submission gate.

Open Music Business provides educational information, not individualized mastering, legal, financial, tax, contract, or royalty advice. Platform documentation can change, and requirements may vary by delivery route, partner status, content profile, territory, playback mode, and date. Verify the current first-party specification before exporting a final release package.