Three decoders, measured against FFmpeg
A decoder is the part of a player that turns a music file back into sound. Orpheic has its own for MP3, for FLAC, and for DST, the lossless compression inside SACDs.
The code was written by Claude, an AI model from Anthropic, from the published standards, under our direction. Nothing was accepted on its word. Each claim on this page was measured, and the code and the tools to repeat the measurements are open, under the MIT licence.
How much faster
How many times faster each decoder runs than FFmpeg's decoder for the same format. Decoder alone, one thread, the same data for both. Each bar is the middle value over the files, on a desktop computer (Intel i9-14900K) against FFmpeg 9.0.2, the newest release.
FLAC2.37 times
MP3, fast path2.25 times
DST, stereo2.01 times
DST, five or six channels1.81 times
FLAC: 8 files, from 2.12 to 2.47 times. MP3: 32 files, from 2.18 to 2.57. DST: five excerpts of SACDs.
On other machines
| Desktop PC, FFmpeg 9.0.2 | MacBook (Apple M4), FFmpeg 9.0.2 from Homebrew | Raspberry Pi 5, FFmpeg 7.1.5 from the system |
| FLAC | 2.37 times | 3.45 times | 1.58 times |
| MP3, fast path | 2.25 times | 2.76 times | 2.27 times |
| DST, stereo | 2.01 times | 2.06 times | 2.89 times |
| DST, five or six channels | 1.81 times | 1.60 times | 2.74 times |
Each number is against the FFmpeg that most people on that system have. Against the FFmpeg 6.1.1 that Ubuntu 24.04 ships, the desktop numbers are 2.25, 2.38, 1.89 and 1.79.
What these numbers do not say
- MP3 has two paths, and the number is for the fast one. The exact path, which Orpheic uses to play music, is written for accuracy and is several times slower than FFmpeg. It still runs hundreds of times faster than the music.
- FLAC and MP3 were measured on eight freely licensed recordings, from orchestra to rock and jazz. Anyone can download the same recordings and repeat the measurement.
- DST was measured on five excerpts of our own that we cannot give out. FFmpeg's DST decoder also turns DSD into PCM and ours does not, so the numbers leave that step out. With it left in they are a little higher: 2.08 and 1.88 on the desktop computer.
- On the Raspberry Pi 5, FLAC gains least, and 24-bit files least of all: 1.27 to 1.45 times.
The same sound
Speed is worth nothing if the sound changes. So every decoder was compared with FFmpeg's on real music.
- FLAC. FLAC is lossless, so the two must agree exactly. On 95 files, 676 million samples, no sample differs.
- DST. Also lossless. On 24 disc areas, 81,000 frames, no bit differs.
- MP3. The MP3 standard is arithmetic on real numbers, so two correct decoders differ in the last digits. Each decoder was compared with the same arithmetic done in double precision. On the 32 MP3 files made from the eight free recordings, the largest gap for Orpheic's was 0.42 millionths of full scale. One step of 16-bit sound is 73 times larger than that.
MP3: distance from the exact result
The same measure for four decoders, on five other files of our own, in millionths of full scale. Shorter is closer. None of these can be heard.
Orpheic0.18
minimp30.48
mpg1231.18
FFmpeg1.34
These five files are not the 32 above, which is why Orpheic shows 0.18 here and 0.42 there. The other three decoders were not measured on the 32.
16,000 MP3 files from a real library
Files picked at random from one listener's collection of about 130,000, made by every kind of encoder over 25 years. Both decoders read each file, and every frame was compared.
| The same as FFmpeg in every frame | 15,109 files |
| Differ only in rounding | 602 files |
| Have damaged spots | 200 files |
| Other difference | 10 files |
| Left to FFmpeg (Layer II, MP3 inside WAV, other) | 79 files |
At a damaged spot the two differ on purpose: FFmpeg drops the frame or plays noise, Orpheic plays silence there.
The same bits on every machine
All three decoders give the same output, bit for bit, on Linux with two compilers, on macOS, on a Raspberry Pi 5 and on Windows. This was checked by taking a fingerprint of every sample: 33 MP3 files, 11 FLAC files and 5 files of SACD frames.
For FLAC and DST this follows from the arithmetic, which uses whole numbers only. MP3 needs cosines, sines and powers, and each system answers those a little differently in the last digit. So the MP3 decoder never asks the system. Every such value is written into the code, computed once to 60 digits.
Who gains in practice
Playing an MP3 or a FLAC file never waited on the decoder. There the gain is less work for the processor. With SACDs the decoder is real work for a small machine, and there the gain is room to spare.
- Old computers. On a laptop from 2007, with five or six channels, FFmpeg's decoder runs at 1.19 times the speed of the music and keeps a whole processor core 84 percent busy. Orpheic's runs at about twice that speed and needs 45 percent. For stereo, FFmpeg's runs at 3.5 times the speed of the music and Orpheic's at 7.4. Measured.
- Discs with five or six channels. They are the first to break up on a modest machine, and the first to gain. Measured.
- Small boxes that sit next to the DAC (the device that turns the file into sound), such as a Raspberry Pi. On a Raspberry Pi 5, with five or six channels, FFmpeg's decoder runs at 2.35 times the speed of the music on one core. Orpheic's runs at 6.7 times. Measured on a Raspberry Pi 5. Older and slower boards were not measured.
- Anything that reads whole albums, such as measuring loudness or converting music for a phone. The time falls by the gains above. Follows from the measured speeds.
- Energy. Decoding the same SACD takes about half the processor energy: 1.8 to 2.1 times less than FFmpeg 6.1.1 over five excerpts. Measured on the desktop computer only, one core, the processor alone and not the whole machine.
- Quiet and battery-powered machines. Less work for the processor should mean a fan that stays off and a battery that lasts longer. Not measured.
- Phones that play to a USB DAC. The same decoder is built for phones as well. Not measured on a phone.
- Other players. The three decoders are plain C++ with nothing else to install, under the MIT licence.
How much of one processor core each decoder keeps busy while an SACD plays. Shorter is better.
Raspberry Pi 5
Stereo, Orpheic4.8 percent
Stereo, FFmpeg14.7 percent
Five or six channels, Orpheic14.9 percent
Five or six channels, FFmpeg42.6 percent
A laptop from 2007
Stereo, Orpheic13.5 percent
Stereo, FFmpeg28.2 percent
Five or six channels, Orpheic45.2 percent
Five or six channels, FFmpeg84 percent
Above 100 percent a machine cannot keep up and the music breaks up. Three discs on each machine.
Memory
Kilobytes of memory each decoder asks for while it works. Orpheic's needs a little more than FFmpeg's for MP3 and FLAC, and much less for DST. All of it is small next to the player itself, which uses about 66,000.
| Orpheic | FFmpeg |
| MP3 | 50 | 42 |
| FLAC | 64 | 54 |
| DST, stereo | 48 | 150 |
| DST, five or six channels | 48 | 232 |
Other MP3 decoders
Time to decode one minute of music, in thousandths of a second, on the desktop computer. One file at 128 kilobits per second. Shorter is faster.
Orpheic9.4
FFmpeg24.4
mpg12325.87
minimp327.02
FFmpeg, mpg123 and minimp3 were only run, as finished programs. Their code was not read.
How this was measured, and what a careful reader should know
- FFmpeg, mpg123 and minimp3 were used as finished programs, through their public interfaces.
- Everything runs on one thread. FFmpeg's own command-line tool can use several threads for FLAC.
- The speed table at the top comes from one tool, on the same free recordings, on three machines. The charts for processor share, memory, the other MP3 decoders and the 2007 laptop are from earlier measurements, made between 12 and 15 November with the tool inside the player and our own test files. They are not comparable one to one with the table.
- mpg123 and minimp3 cut the stream into frames inside the timed part. Orpheic and FFmpeg receive frames already cut. That favours Orpheic slightly.
- The exact result used for MP3 accuracy is Orpheic's own double-precision reading of the standard. It is a witness, not an outside authority. In accuracy, minimp3 is in the same class as Orpheic.
- FFmpeg's DST decoder returns PCM, so its time includes turning DSD into PCM. Orpheic's stops at DSD. That step is 4 to 7 percent of FFmpeg's time. With it taken out, the stereo gain on the desktop computer is 1.97 instead of 2.05.
- FFmpeg 9.0.2 on the desktop computer was built from the signed release with its default settings. On the MacBook it is the one Homebrew installs, on the Raspberry Pi 5 the system's own package. The 2007 laptop was measured against the FFmpeg 6.1.2 that ships with Orpheic.
- Measured from 12 to 15 November 2026.
Check it yourself
The three decoders and the small tools that repeat the main checks are in one repository. Each tool prints one fingerprint per file. The same file must give the same line on your machine as on ours.
github.com/orpheic/orpheic-decoders
Where the code comes from
The MP3 and FLAC decoders were written from their standards, and checked against other decoders that were only run, never read. The DST decoder is the second one written for Orpheic. The first was written from the standard and compared, byte for byte, with the standard's reference decoder. The present one keeps its header reading and rewrites the rest for speed.
Nobody can show what an AI model remembers from its training. So the code was also compared with other decoders for the same formats. The reviewers were AI models too: separate ones, which had not written the code and were asked to look for copying. What is shared is what the standards themselves print. The one part of the MP3 decoder they found organised like another decoder was rewritten.
Orpheic itself is not open yet. We are preparing to open the whole program.