Lossless comparison: Difference between revisions

From Hydrogenaudio Knowledgebase
(table consistency with the codecs section)
(Table: Error handling, Tagging ambiguity, OS support (Monkey's ++). Each codec: several fixes. Overhaul still overdue?)
 
(18 intermediate revisions by 7 users not shown)
Line 21: Line 21:
! width="90px" | [[#OptimFROG (OFR)|OptimFROG]]
! width="90px" | [[#OptimFROG (OFR)|OptimFROG]]
! width="90px" | [[#True Audio (TTA)|TTA]]
! width="90px" | [[#True Audio (TTA)|TTA]]
|- <!-- *** Encoding speed is very fast if > 150x, fast if >75x, average if >40x, slow if >20x, very slow if <20x *** -->
|- <!-- *** Encoding speed is very fast if < 2%, fast if < 5%, average if < 10%, slow if < 20%, very slow if < 50% *** -->
  <!-- *** Numbers are taken from most recent CDDA comparison under 'External links' *** -->
| Encoding speed{{ref label|speed|A|A}}
| Encoding speed{{ref label|speed|A|A}}
| style="background: #00FF00" | very fast
| style="background: #00FF00" | very fast
Line 27: Line 28:
| style="background: #00FF00" | very fast
| style="background: #00FF00" | very fast
| style="background: #00FF00" | very fast
| style="background: #00FF00" | very fast
| style="background: #CCFFCC" | fast
| style="background: #CCFFCC" | fast
| style="background: #FFCC66" | slow
| style="background: #00FF00" | very fast
| style="background: #00FF00" | very fast
|- <!-- *** For decoding speed thresholds are doubled, i.e., very fast if >300x, fast if >150x etc *** -->
| style="background: #00FF00" | very fast
| style="background: #FFCC66" | moderate
| style="background: #00FF00" | very fast
|- <!-- *** For decoding speed thresholds are one step down, i.e., very fast if < 1%, fast if < 2%, average if < 5% etc *** -->
| Decoding speed{{ref label|speed|A|A}}
| Decoding speed{{ref label|speed|A|A}}
| style="background: #00FF00" | very fast
| style="background: #00FF00" | very fast
| style="background: #CCFFCC" | fast
| style="background: #CCFFCC" | fast
| style="background: #00FF00" | very fast
| style="background: #00FF00" | very fast
| style="background: #FFCC66" | moderate
| style="background: #CCFFCC" | fast
| style="background: #CCFFCC" | fast
| style="background: #00FF00" | very fast
| style="background: #FFCC66" | moderate
| style="background: #FFCC66" | slow
| style="background: #FFCC66" | average
| style="background: #FF9900" | very slow
| style="background: #CCFFCC" | fast
| style="background: #CCFFCC" | fast
|- <!-- *** Thresholds for compression are at 56% and 58% *** -->  
|- <!-- *** Thresholds for compression are at 51% and 53% *** -->  
| Compression{{ref label|speed|A|A}}{{ref label|comp|B|B}}
| Compression{{ref label|speed|A|A}}{{ref label|comp|B|B}}
| style="background: #CCFFCC" | 57.0%
| style="background: #CCFFCC" | 52.0%
| style="background: #CCFFCC" | 57.8%
| style="background: #FFCC66" | 53.2%
| style="background: #CCFFCC" | 57.1%
| style="background: #CCFFCC" | 52.6%
| style="background: #00FF00" | 56.0%
| style="background: #00FF00" | 50.5%
| style="background: #00FF00" | 55.1%
| style="background: #00FF00" | 50.7%
| style="background: #FFCC66" | 58.4%
| style="background: #FFCC66" | 53.8%
| style="background: #00FF00" | 54.6%
| style="background: #00FF00" | 49.6%
| style="background: #CCFFCC" | 56.6%
| style="background: #CCFFCC" | 52.1%
|-
|-
| # presets
| # presets
Line 70: Line 71:
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #FF9900" | no
|-
|-
| Tagging
| Tagging{{ref label|taggingambiguity|E|E}}
| style="background: #00FF00" | Vorbis tags
| style="background: #00FF00" | Vorbis tags
| style="background: #00FF00" | iTunes
| style="background: #00FF00" | iTunes
| style="background: #CCFFCC" | ID3/APEv2
| style="background: #00FF00" | APEv2
| style="background: #00FF00" | APEv2
| style="background: #00FF00" | APEv2
| style="background: #00FF00" | APEv2
| style="background: #00FF00" | APEv2
| style="background: #00FF00" | ASF
| style="background: #00FF00" | ASF
| style="background: #CCFFCC" | ID3/APEv2
| style="background: #FFCC66" | APEv2 or ID3
| style="background: #00FF00" | ID3v1/2 or APEv2
| style="background: #FFCC66" | APEv2 or ID3
|-
|-
| Hardware support  
| Hardware support  
Line 136: Line 137:
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #FFCC66" | no{{ref label|tak_os|F|F}}
| style="background: #FFCC66" | no{{ref label|ffmpeg_dec|F|F}}
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #FF9900" | no
| style="background: #FFCC66" | no{{ref label|ffmpeg_dec|F|F}}
| style="background: #FF9900" | no
| style="background: #FF9900" | no
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
Line 147: Line 148:
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes{{ref label|multichannel_ape|E|E}}
| style="background: #00FF00" | yes{{ref label|multichannel_ape|G|G}}
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
| style="background: #FF9900" | no
| style="background: #FF9900" | no
| style="background: #00FF00" | yes
| style="background: #00FF00" | yes
|-
|-
| OS support
| OS support{{ref label|OS_encode|H|H}}
| style="background: #00FF00" | All
| style="background: #00FF00" | All
| style="background: #00FF00" | All
| style="background: #00FF00" | All
| style="background: #00FF00" | All
| style="background: #00FF00" | All
| style="background: #CCFFCC" | Win/Wine
| style="background: #CCFFCC" | Win/Wine
| style="background: #00FF00" | All
| style="background: #CCFFCC" | Win {{ref label|OS_ape|I|I}}
| style="background: #CCFFCC" | Win/Mac
| style="background: #CCFFCC" | Win/Mac
| style="background: #00FF00" | Win/Mac/Linux
| style="background: #CCFFCC" | Any desktop
| style="background: #00FF00" | All
| style="background: #00FF00" | All
|-
| Fits in container
| style="background: #00FF00" | Ogg, MKV, MP4, CAF
| style="background: #00FF00" | MKV, MP4, CAF
| style="background: #00FF00" | MKV
| style="background: #CCFFCC" | -
| style="background: #CCFFCC" | -
| style="background: #CCFFCC" | -
| style="background: #CCFFCC" | -
| style="background: #00FF00" | MKV
|}
|}


{|
{|
|-
|-
|{{note label|speed|A|A}} Speed and Compression are based on '''each encoder's default settings''' and taken from [http://www.audiograaf.nl/downloads.html this comparison].
|{{note label|speed|A|A}} Speed and Compression are based on '''each encoder's default settings''' and taken from the last CDDA results of [http://www.audiograaf.nl/downloads.html this comparison].
|-
|-
|{{note label|comp|B|B}} The Compression ratio is compressed size/uncompressed size * 100. So, lower is better.  
|{{note label|comp|B|B}} Lower is better: Compression ratio is compressed size/uncompressed size * 100%.  
|-
|-
|{{note label|error|C|C}} Error handling means that a codec can detect a corruption (flipped bit) in a file and warn the user about it, but it will still decode the rest of the file.
|{{note label|error|C|C}} Error handling means that a codec can detect a corruption (flipped bit) in a file, warn the user about it, protect output against full-volume static, but it will still decode the rest of the file.
|-
|-
|{{note label|error_ape|D|D}} The official Monkey's Audio decoder does not support decoding through errors, but this may be achieved with FFmpeg or Winamp, though likely not, when the "Insane" preset is used.
|{{note label|error_ape|D|D}} The official Monkey's Audio decoder does not support decoding through errors, but this may be achieved with FFmpeg (or Winamp).
|-
|-
|{{note label|multichannel_ape|E|E}} Since version 4.86
|{{note label|taggingambiguity|E|E}} Tagging ambiguity is generally not desired, as one application might ignore tags written by another, or even reject files.  Most APEv2-compatible formats can be tagged with ID3, although not recommended unless for use with a hardware player that cannot read other tag sets, and some applications like [[Exact Audio Copy]] can attach an ID3 chunk to any format &ndash; including FLAC, which does not support it. All formats but OptimFROG and TTA state either a unique choice of tag set or a clear preferred choice.
|-
|-
|{{note label|tak_os|F|F}} Unofficial (but properly working) open source decoder is available as part of ffmpeg
|{{note label|ffmpeg_dec|F|F}} Unofficial (but properly working) open source decoder is available as part of ffmpeg
|-
|{{note label|multichannel_ape|G|G}} Since version 4.86
|-
|{{note label|OS_encode|H|H}} What platform the codec supports or can be played on. Also, most operating systems may offer native support for FLAC and ALAC (Android only FLAC).
|-
|{{note label|OS_ape|I|I}} Currently, Monkey's Audio is available for Windows. Monkey's Audio version 3.99 (stereo, up to 24 bits) is available on all desktop platforms and can be played back on mobile platforms.
 
== Codecs ==
== Codecs ==


These are the most popular lossless codecs, in alphabetical order:
The most popular lossless codecs, in alphabetical order:


=== Apple Lossless Audio Codec (ALAC) ===
=== Apple Lossless Audio Codec (ALAC) ===
https://alac.macosforge.org/trac
https://alac.macosforge.org/trac


[[ALAC]] is a codec developed by Apple for usage in [[Apple iPod|iPod]] and AirPort Express.
[[ALAC]] is a codec developed by Apple and used across their hardware and software platforms.


'''ALAC pros'''
'''ALAC pros'''
* [[Open source]] (encoding and decoding via FFmpeg and [[CueTools|CUETools]], decoding only via [http://craz.net/programs/itunes/alac.html a standalone decoder])
* [[Open source]] (encoding and decoding via macosforge/refalac, FFmpeg and [[CueTools|CUETools]])
* Fast encoding
* Fast encoding
* Fast decoding
* Fast decoding
* Hardware support ([[Apple iPod|iPod]], AirPort Express)
* Hardware support ([[Apple iPod|iPod]], AirPort Express)
* Software support (iTunes, Quicktime)
* Software support (iTunes, Quicktime)
* Independent encoder implementation available: ffmpeg
* Independent encoder implementation available: ffmpeg, CUETools
* Streaming support
* Streaming support
* Tagging support (QT tags)
* Tagging support (QT tags)
Line 199: Line 217:


''' ALAC cons '''
''' ALAC cons '''
* Limited software support
* No error detection/robustness<ref>[http://www.hydrogenaud.io/forums/index.php?s=&showtopic=33226&view=findpost&p=862031 HA forum post discussing ALAC robustness]</ref>
* No error detection/robustness<ref>[http://www.hydrogenaud.io/forums/index.php?s=&showtopic=33226&view=findpost&p=862031 HA forum post discussing ALAC robustness]</ref>
* Doesn't support [[RIFF]] chunks
* No hybrid/lossy mode (and not [[LossyWAV]] compatible)
* No hybrid/lossy mode (and not [[LossyWAV]] compatible)
* Not very efficient


''' ALAC Other features '''
''' ALAC Other features '''
* Fits in the [[MP4]] container
* Fits in the [[MP4]], [[Matroska]] and CAF containers


=== Free Lossless Audio Codec (FLAC) ===
=== Free Lossless Audio Codec (FLAC) ===
Line 218: Line 235:
* Very good hardware support (Android, Marantz, Sonos, [http://xiph.org/flac/links.html many others])
* Very good hardware support (Android, Marantz, Sonos, [http://xiph.org/flac/links.html many others])
* Very good software support
* Very good software support
* Independent encoder implementations available: flake/ffmpeg, FLACCL
* Independent encoder implementations available: flake/ffmpeg, FLACCL (providing GPU-based operation, potentially extremely fast)
* Error robustness
* Error robustness
* Streaming support
* Streaming support
* Supports [[multichannel]]. Limited to 8 channels. Channel mask in [https://docs.microsoft.com/ru-ru/windows/win32/api/mmreg/ns-mmreg-waveformatextensible WAVEFORMATEXTENSIBLE] is supported but support is not included in official specification. With reference encoder undocumented option --channel-map=none is needed to encode some non-standard layouts (e.g. 4.1; FL,FR,FC,BC), but no special options are needed with ffmpeg's encoder.
* Supports [[multichannel]]. Limited to 8 channels. Channel mask in [https://learn.microsoft.com/en-us/windows/win32/api/mmreg/ns-mmreg-waveformatextensible WAVEFORMATEXTENSIBLE] is supported.
* Supports [[high resolution]]s
* Supports [[high resolution]]s
* Tagging support (FLAC tags)
* Tagging support (FLAC tags)
Line 230: Line 247:
''' FLAC cons '''
''' FLAC cons '''
* No hybrid/lossy mode (but is [[LossyWAV]] compatible)
* No hybrid/lossy mode (but is [[LossyWAV]] compatible)
* Does not handle 32-bit float and there is no encoder that can render to 32-bit integer
* Does not handle 32-bit float


''' FLAC Other features '''
''' FLAC Other features '''
* Supports embedded CUE sheets (with [http://flac.sourceforge.net/faq.html#general__no_cuesheet_tags limitations])
* Supports embedded CUE sheets (with [http://flac.sourceforge.net/faq.html#general__no_cuesheet_tags limitations])
* Includes MD5 hashes for quick integrity checking as standard
* Includes MD5 hashes for quick integrity checking as standard
* Fits the [[Ogg]] and [[Matroska]] containers
* Fits the [[Ogg]], [[Matroska]], [[MP4]] and CAF containers


=== Monkey's Audio (APE) ===
=== Monkey's Audio (APE) ===
Line 243: Line 260:


''' APE pros '''
''' APE pros '''
* Open source (from version 10.18, Windows)
* High compression
* High compression
* Fast encoding
* Fast encoding at default setting
* Good software support
* Good software support
* Supports [[multichannel]] (since version 4.86). Limited to 8 channels. Channel mask in [https://docs.microsoft.com/ru-ru/windows/win32/api/mmreg/ns-mmreg-waveformatextensible WAVEFORMATEXTENSIBLE] is supported
* Supports [[multichannel]] (Windows, since version 4.86, channel count later increased.) Channel mask in [https://docs.microsoft.com/ru-ru/windows/win32/api/mmreg/ns-mmreg-waveformatextensible WAVEFORMATEXTENSIBLE] is supported
* Supports [[high resolution]]s
* Supports [[high resolution]]s
* Simple and user friendly. Official GUI provided.
* Simple and user friendly. Official GUI provided.
* Java version (multiplatform)
* Multiplatform [https://jmac.sourceforge.net/ Java implementation] available (3.99 version)
* Error robustness/decoding up to -c3000 (High compression)<ref>http://www.hydrogenaud.io/forums/index.php?showtopic=98984&st=0&p=821420&#entry821420</ref>
* Tagging support ([[APEv2]], optionally [[ID3v1]])
* Tagging support ([[ID3v1]], [[APE tags]])
* Supports [[RIFF]] chunks
* Supports [[RIFF]] chunks (only in the GUI encoder)
* Pipe support (Windows version, other OSes only in a [http://shnutils.freeshell.org/shntool/ special] 3.99-based version)
* Pipe support (only in a [http://www.etree.org/shnutils/shntool/ special] version)


''' APE cons '''
''' APE cons '''
* Problematic license (source provided, no modification or redistribution rights)
* Slow decoding. Higher compression levels are particularly CPU intensive.
* Slow decoding
* Limited hardware support (Rockbox, some Cowon players); potentially poor battery life due to CPU-intensive decoding (see [http://www.rockbox.org/wiki/SoundCodecMonkeysAudio MP3 player benchmarks])
* No hybrid/lossy mode (and not [[LossyWAV]] compatible)
* No hybrid/lossy mode (and not [[LossyWAV]] compatible)
* Limited hardware support (Rockbox, some Cowon players); poor battery life due to complicated decoding (see [http://www.rockbox.org/wiki/SoundCodecMonkeysAudio MP3 player benchmarks])
* Older versions (for non-Windows platforms) released under a problematic license
* Higher compression levels are extremely CPU intensive


''' APE Other features '''
''' APE Other features '''
* Includes MD5 hashes for quick integrity checking
* Includes MD5 hashes for quick integrity checking (without decoding)
* Supports APL image link files (similar to CUE sheets)
* Supports APL image link files (similar to CUE sheets)


Line 278: Line 294:
* Supports [[high resolution]]s
* Supports [[high resolution]]s
* Hybrid/lossy mode
* Hybrid/lossy mode
* Tagging support ([[ID3]], [[APE tags]])
* Tagging support ([[APEv2]], [[ID3]])
* Supports [[RIFF]] chunks
* Supports [[RIFF]] chunks


Line 291: Line 307:
''' OFR Other features '''
''' OFR Other features '''
* Supports 32bit float streams
* Supports 32bit float streams
* Includes MD5 hashes for quick integrity checking
* Fast verification (without decoding) or full decoding verification
* Optional MD5 hash


=== Tom's verlustfreier Audiokompressor (TAK) ===
=== Tom's verlustfreier Audiokompressor (TAK) ===
Line 304: Line 321:
* Error robust
* Error robust
* Supports [[multichannel]]. Limited to 6 channels. Channel mask in [https://docs.microsoft.com/ru-ru/windows/win32/api/mmreg/ns-mmreg-waveformatextensible WAVEFORMATEXTENSIBLE] is supported
* Supports [[multichannel]]. Limited to 6 channels. Channel mask in [https://docs.microsoft.com/ru-ru/windows/win32/api/mmreg/ns-mmreg-waveformatextensible WAVEFORMATEXTENSIBLE] is supported
* Supports [[high resolution]]s
* Supports [[high resolution]]s. Limited to 192 kHz sampling rate.
* Tagging support
* Tagging support ([[APEv2]])
* Supports RIFF chunks
* Supports RIFF chunks
* Pipe support  
* Pipe support  
Line 315: Line 332:
* No hardware support
* No hardware support
* Average software support
* Average software support
* Doesn't support Unicode (yet)
* Unicode only in GUI


''' TAK Other features '''
''' TAK Other features '''
Line 327: Line 344:
''' TTA pros '''
''' TTA pros '''
* [[Open source]]
* [[Open source]]
* Supports [[multichannel]]. Reference encoder/decoder is limited to 6 channels. ffmpeg's encoder/decoder is limited to 16 channels. Channel mask in [https://docs.microsoft.com/ru-ru/windows/win32/api/mmreg/ns-mmreg-waveformatextensible WAVEFORMATEXTENSIBLE] is '''not''' supported
* Supports [[multichannel]]. Reference encoder/decoder is limited to 6 channels. ffmpeg's encoder/decoder is limited to 16 channels. Channel mask in [https://docs.microsoft.com/ru-ru/windows/win32/api/mmreg/ns-mmreg-waveformatextensible WAVEFORMATEXTENSIBLE] is '''not''' supported (but implemented in 3rd-party foobar2000 component).
* Supports [[high resolution]]s
* Supports [[high resolution]]s
* Tagging support ([[ID3]]v1, ID3v2 or [[APEv2]])
* Tagging support ([[ID3]]v1, ID3v2 or [[APEv2]])
* Embedded CUE sheets support
* Embedded CUE sheets support
* Error robustness
* Pipe support
* Pipe support
* Average compression
* Fast encoding/decoding
* Fast encoding/decoding
* Symmetric algorithm
* Ultra low latency


''' TTA cons '''
''' TTA cons '''
* No error detection nor error handling (neither in reference implementation nor ffmpeg)
* Tagging ambiguity: tagging applications differ on what tag format to choose, and might not read each others' tags
* Reference implementation is picky on input files and rejects a range of TTA files as well.
* No hybrid/lossy mode
* No hybrid/lossy mode
* Doesn't support [[RIFF]] chunks
* Doesn't support [[RIFF]] chunks
Line 345: Line 361:
''' TTA Other features '''
''' TTA Other features '''
* Fits the [[Matroska]] container
* Fits the [[Matroska]] container
* Password protection
* Optional password protection


=== WavPack (WV) ===
=== WavPack (WV) ===
Line 362: Line 378:
* Supports [[high resolution]]s
* Supports [[high resolution]]s
* Hybrid/lossy mode
* Hybrid/lossy mode
* Tagging support ([[ID3v1]], [[APE tags]])
* Tagging support ([[APEv2]]; ID3 is possible but not recommended)
* Supports [[RIFF]] chunks
* Supports [[RIFF]] chunks
* Ability to create self extracting files for Win32 platform
* Pipe support
* Pipe support
* Good software support
* Good software support
* Works with Android (Through third party software, such as VLC.)
* Independent encoder implementation available. (FFmpeg WavPack)


''' WV cons '''
''' WV cons '''
* Limited hardware player support ([http://www.rockbox.org/ RockBox])
* Limited hardware player support ([http://www.rockbox.org/ RockBox])
* More than one tagging method allowed (ambiguity possible)


''' WV Other features '''
''' WV Other features '''
Line 378: Line 394:
* Accept audio files bigger than 4GB
* Accept audio files bigger than 4GB
* Includes MD5 hashes for quick integrity checking
* Includes MD5 hashes for quick integrity checking
* Can encode in both symmetrical and asymmetrical modes.
* Fits the [[Matroska]] container
* Fits the [[Matroska]] container


Line 384: Line 399:
https://msdn.microsoft.com/en-us/library/ff819508(v=vs.85).aspx
https://msdn.microsoft.com/en-us/library/ff819508(v=vs.85).aspx


WMA Lossless is the lossless codec developed by Microsoft to be featured in their Windows Media codec portfolio.
[[Windows_Media_Audio#Windows_Media_Audio_Lossless|WMA Lossless]] is the lossless codec developed by Microsoft to be featured in their Windows Media codec portfolio.


''' WMAL pros '''
''' WMAL pros '''
* Streaming support
* Streaming support
* Very good software support
* Hardware support (Microsoft Zune, [http://en.wikipedia.org/wiki/Gigabeat Gigabeat V and S line from Toshiba])
* Supports [[multichannel]] audio and [[high resolution]]s.
* Supports [[multichannel]] audio and [[high resolution]]s.
* Tagging support (proprietary)
* Tagging support (proprietary)
Line 395: Line 408:


''' WMAL cons '''
''' WMAL cons '''
* Rather low efficiency
* Abandonware? To the extent that certain Windows preview versions have been unable to decode losslessly.
* Limited hardware support (Microsoft Zune, Toshiba Gigabeat S and V. Both discontinued and obsolete. Rockbox, for 16-bit stereo files only.)
* Limited software support outside of the Microsoft Windows operating system.
* Low efficiency
* Closed source
* Closed source
* No hybrid/lossy mode (but is [[LossyWAV]] compatible)
* No hybrid/lossy mode (but is [[LossyWAV]] compatible)
Line 412: Line 428:
http://www.lossless-audio.com/
http://www.lossless-audio.com/


LA features an extremely high compression (on par with OptimFrog highest modes, but a bit faster), but it hasn't been updated for more than 10 years. Furthermore, backward compatibility is not guaranteed, so using it for archiving might pose a few problems. It isn't able to cope with file corruption either, software support is very limited and isn't open source.
LA features an extremely high compression (on par with OptimFrog highest modes), but never made it out of beta stage and hasn't been updated for more than 20 years. Furthermore, backward compatibility is not guaranteed, so using it for archiving might pose a few problems. It isn't able to cope with file corruption either, software support is very limited and isn't open source.


====MLP/Dolby TrueHD====
====MLP/Dolby TrueHD====
Line 418: Line 434:


====MPEG-4 ALS====
====MPEG-4 ALS====
MPEG-4 ALS is the successor to LPAC, which it was based on. It has been as a ISO standard and there is a reference encoder/decoder, but like TTA, it does not have features that make it stand out from other codecs, nor backing by a large organisation, so it hasn't much software and no hardware support.
MPEG-4 ALS is the successor to LPAC, which it was based on. It has been as a ISO standard and there is a reference encoder/decoder, which also supports floating-point signals. It never gained much popularity. FFmpeg can decode some ALS files.


====MPEG-4 SLS====
====MPEG-4 SLS====
Line 429: Line 445:


====Real Lossless====
====Real Lossless====
Part of the Real codec suite, Real Lossless too hasn't any very special features that make it stand out. Just like WMA Lossless and Apple Lossless, it was created to fit in a codec suite, but unlike WMA Lossless and Apple Lossless, there is no hardware support and software support is limited. Compression is on par with most other codecs, but it is rather slow to encode.
Just like WMA Lossless and Apple Lossless, it was created to fit in a codec suite, but unlike WMA Lossless and Apple Lossless, there is no hardware support and software support is limited. Compression is on par with most other codecs, but it is rather slow to encode. It is considered obsolete, and Real contributed to the creation of MPEG-4 ALS.


====Oddball formats====
====Oddball formats====
There are a few archaic formats of which encoders and decoders are hard to get by. Most of those would have disappeared by now, but some of them are being preserved for posterity at [[User:Rjamorim|rjamorim]]'s
There are a few archaic formats of which encoders and decoders are hard to come by. Most of those would have disappeared by now, but some of them are being preserved for posterity at [[User:Rjamorim|rjamorim]]'s [https://www.rarewares.org/rrw/ ReallyRareWares].


* Advanced Digital Audio (ADA)   
* Advanced Digital Audio (ADA)   
Line 458: Line 474:
''Sorted based on last '''update''' date.''
''Sorted based on last '''update''' date.''


* [http://www.audiograaf.nl/downloads.html Martijn van Beurden's comparison] - tries to compare all codecs and settings with a balanced pool of music (last updated 2015-01-05)
* [http://www.audiograaf.nl/downloads.html Martijn van Beurden's comparison] - tries to compare all codecs and settings with a balanced pool of music, with separate sections on hi-res and multichannel material (last version 2023-08-29)
* [http://www.squeezechart.com/audio.html Squeezechart audio] - tests as much codecs as possible, but not all their settings and with a limited test corpus (last updated 2013-10-31)
* [http://www.squeezechart.com/audio.html Squeezechart audio] - tests as much codecs as possible, but not all their settings and with a limited test corpus (last updated 2013-10-31)
* <s>[http://synthetic-soul.co.uk/comparison/lossless/index.asp Synthetic Soul's comparison] (last update 2007-07-28)</s>
* <s>[http://synthetic-soul.co.uk/comparison/lossless/index.asp Synthetic Soul's comparison] (last update 2007-07-28)</s>

Latest revision as of 08:20, 20 May 2024

The lossless comparison page aims to gather information about lossless codecs available so users can make an informed decision as to what lossless codec to choose for their needs.

Introduction

Given the enormous number of lossless audio compressor choices available, it is a very difficult task to choose the one most suited for each person's needs. Some people take into consideration only compression performance when choosing a codec, but as the following table and article shows, there are several other features worth taking into consideration when making a choice.

For example, users wanting good multiplatform compatibility and robustness (e.g., people sharing live recordings) would favour WavPack or FLAC. Another user, looking for the very highest compression available, would go with OptimFROG. Someone wanting portable support would use FLAC or ALAC, and so on. En fin, this is not a matter worth getting too worked up about. If you later find out the codec you chose isn't the best for your needs, you can just transcompress to another format, without risk of losing quality.

Note: for latest comparison of lossless compression, scroll down to the Links section of this page.

Comparison Table

Features FLAC ALAC WavPack TAK Monkey's WMAL OptimFROG TTA
Encoding speed[A] very fast fast very fast very fast very fast very fast moderate very fast
Decoding speed[A] very fast fast very fast very fast moderate fast moderate fast
Compression[A][B] 52.0% 53.2% 52.6% 50.5% 50.7% 53.8% 49.6% 52.1%
# presets 9 2 > 10 > 10 5 1 > 10 1
Error handling[C] yes no yes yes yes[D] yes yes no
Tagging[E] Vorbis tags iTunes APEv2 APEv2 APEv2 ASF APEv2 or ID3 APEv2 or ID3
Hardware support very good good limited no limited limited no limited
Software support very good good good average good good average good
Hybrid/lossy LossyWAV no yes LossyWAV no LossyWAV yes no
RIFF chunks yes no yes yes yes no yes no
Streaming yes yes yes yes no yes yes yes
Open source yes yes yes no[F] yes no[F] no yes
Multichannel yes yes yes yes yes[G] yes no yes
OS support[H] All All All Win/Wine Win [I] Win/Mac Any desktop All
Fits in container Ogg, MKV, MP4, CAF MKV, MP4, CAF MKV - - - - MKV
A Speed and Compression are based on each encoder's default settings and taken from the last CDDA results of this comparison.
B Lower is better: Compression ratio is compressed size/uncompressed size * 100%.
C Error handling means that a codec can detect a corruption (flipped bit) in a file, warn the user about it, protect output against full-volume static, but it will still decode the rest of the file.
D The official Monkey's Audio decoder does not support decoding through errors, but this may be achieved with FFmpeg (or Winamp).
E Tagging ambiguity is generally not desired, as one application might ignore tags written by another, or even reject files. Most APEv2-compatible formats can be tagged with ID3, although not recommended unless for use with a hardware player that cannot read other tag sets, and some applications like Exact Audio Copy can attach an ID3 chunk to any format – including FLAC, which does not support it. All formats but OptimFROG and TTA state either a unique choice of tag set or a clear preferred choice.
F Unofficial (but properly working) open source decoder is available as part of ffmpeg
G Since version 4.86
H What platform the codec supports or can be played on. Also, most operating systems may offer native support for FLAC and ALAC (Android only FLAC).
I Currently, Monkey's Audio is available for Windows. Monkey's Audio version 3.99 (stereo, up to 24 bits) is available on all desktop platforms and can be played back on mobile platforms.

Codecs

The most popular lossless codecs, in alphabetical order:

Apple Lossless Audio Codec (ALAC)

https://alac.macosforge.org/trac

ALAC is a codec developed by Apple and used across their hardware and software platforms.

ALAC pros

ALAC cons

  • No error detection/robustness[1]
  • Doesn't support RIFF chunks
  • No hybrid/lossy mode (and not LossyWAV compatible)

ALAC Other features

Free Lossless Audio Codec (FLAC)

https://xiph.org/flac/

FLAC is a lossless codec developed by Josh Coalson. It's part of the Xiph multimedia portfolio, along with Opus, Ogg, Vorbis, Speex and Theora.

FLAC pros

  • Open source
  • Very fast decoding
  • Very fast encoding
  • Very good hardware support (Android, Marantz, Sonos, many others)
  • Very good software support
  • Independent encoder implementations available: flake/ffmpeg, FLACCL (providing GPU-based operation, potentially extremely fast)
  • Error robustness
  • Streaming support
  • Supports multichannel. Limited to 8 channels. Channel mask in WAVEFORMATEXTENSIBLE is supported.
  • Supports high resolutions
  • Tagging support (FLAC tags)
  • Supports RIFF chunks
  • Pipe support
  • Used by a few online stores

FLAC cons

  • No hybrid/lossy mode (but is LossyWAV compatible)
  • Does not handle 32-bit float

FLAC Other features

  • Supports embedded CUE sheets (with limitations)
  • Includes MD5 hashes for quick integrity checking as standard
  • Fits the Ogg, Matroska, MP4 and CAF containers

Monkey's Audio (APE)

https://www.monkeysaudio.com/

Monkey's Audio is a very efficient lossless compressor developed by Matt Ashland.

APE pros

  • Open source (from version 10.18, Windows)
  • High compression
  • Fast encoding at default setting
  • Good software support
  • Supports multichannel (Windows, since version 4.86, channel count later increased.) Channel mask in WAVEFORMATEXTENSIBLE is supported
  • Supports high resolutions
  • Simple and user friendly. Official GUI provided.
  • Multiplatform Java implementation available (3.99 version)
  • Tagging support (APEv2, optionally ID3v1)
  • Supports RIFF chunks
  • Pipe support (Windows version, other OSes only in a special 3.99-based version)

APE cons

  • Slow decoding. Higher compression levels are particularly CPU intensive.
  • Limited hardware support (Rockbox, some Cowon players); potentially poor battery life due to CPU-intensive decoding (see MP3 player benchmarks)
  • No hybrid/lossy mode (and not LossyWAV compatible)
  • Older versions (for non-Windows platforms) released under a problematic license

APE Other features

  • Includes MD5 hashes for quick integrity checking (without decoding)
  • Supports APL image link files (similar to CUE sheets)

OptimFROG (OFR)

http://www.losslessaudio.org/

OptimFROG is a lossless format developed by Florin Ghido to become the champion in audio compression.

OFR pros

  • Very high compression
  • Good software support
  • Error robustness
  • Streaming support
  • Supports high resolutions
  • Hybrid/lossy mode
  • Tagging support (APEv2, ID3)
  • Supports RIFF chunks

OFR cons

  • Closed source
  • No multichannel audio support
  • No hardware support
  • Very slow decoding
  • Slow encoding
  • More than one tagging method allowed (ambiguity possible)

OFR Other features

  • Supports 32bit float streams
  • Fast verification (without decoding) or full decoding verification
  • Optional MD5 hash

Tom's verlustfreier Audiokompressor (TAK)

http://www.thbeck.de/Tak/Tak.html

TAK is a lossless codec developed by Thomas Becker.

TAK pros

  • Very fast decoding
  • Very fast encoding
  • Very high efficiency
  • Error robust
  • Supports multichannel. Limited to 6 channels. Channel mask in WAVEFORMATEXTENSIBLE is supported
  • Supports high resolutions. Limited to 192 kHz sampling rate.
  • Tagging support (APEv2)
  • Supports RIFF chunks
  • Pipe support
  • Streamable

TAK cons

  • Closed source (but unofficial open source decoder is available as part of ffmpeg)
  • No hybrid/lossy mode (but is LossyWAV compatible)
  • No hardware support
  • Average software support
  • Unicode only in GUI

TAK Other features

  • Optional MD5 checksum

True Audio (TTA)

http://tta.tausoft.org/

TTA is a lossless codec developed by a international team of programmers.

TTA pros

  • Open source
  • Supports multichannel. Reference encoder/decoder is limited to 6 channels. ffmpeg's encoder/decoder is limited to 16 channels. Channel mask in WAVEFORMATEXTENSIBLE is not supported (but implemented in 3rd-party foobar2000 component).
  • Supports high resolutions
  • Tagging support (ID3v1, ID3v2 or APEv2)
  • Embedded CUE sheets support
  • Pipe support
  • Fast encoding/decoding

TTA cons

  • No error detection nor error handling (neither in reference implementation nor ffmpeg)
  • Tagging ambiguity: tagging applications differ on what tag format to choose, and might not read each others' tags
  • Reference implementation is picky on input files and rejects a range of TTA files as well.
  • No hybrid/lossy mode
  • Doesn't support RIFF chunks
  • Limited hardware support

TTA Other features

  • Fits the Matroska container
  • Optional password protection

WavPack (WV)

http://www.wavpack.com/

WavPack is a fast and featureful lossless codec developed by David Bryant.

WV pros

  • Open source
  • Fast decoding
  • Very fast encoding
  • Good efficiency
  • Error robustness
  • Streaming support
  • Supports multichannel. Limited to 255 channels. Channel mask in WAVEFORMATEXTENSIBLE is supported
  • Supports high resolutions
  • Hybrid/lossy mode
  • Tagging support (APEv2; ID3 is possible but not recommended)
  • Supports RIFF chunks
  • Pipe support
  • Good software support
  • Works with Android (Through third party software, such as VLC.)
  • Independent encoder implementation available. (FFmpeg WavPack)

WV cons

  • Limited hardware player support (RockBox)

WV Other features

  • Can compress the Direct-Stream Digital (DSD) audio recording format
  • Supports 32bit float streams
  • Supports embedded CUE sheets
  • Accept audio files bigger than 4GB
  • Includes MD5 hashes for quick integrity checking
  • Fits the Matroska container

Windows Media Audio Lossless (WMAL)

https://msdn.microsoft.com/en-us/library/ff819508(v=vs.85).aspx

WMA Lossless is the lossless codec developed by Microsoft to be featured in their Windows Media codec portfolio.

WMAL pros

WMAL cons

  • Abandonware? To the extent that certain Windows preview versions have been unable to decode losslessly.
  • Limited hardware support (Microsoft Zune, Toshiba Gigabeat S and V. Both discontinued and obsolete. Rockbox, for 16-bit stereo files only.)
  • Limited software support outside of the Microsoft Windows operating system.
  • Low efficiency
  • Closed source
  • No hybrid/lossy mode (but is LossyWAV compatible)
  • Doesn't support RIFF chunks

WMAL Other features

  • Fits the ASF container

Other Formats

Aside from the formats mentioned above, there are in fact quite a lot of other lossless formats. To keep the table and list brief and readable, a few formats have not been mentioned.

DTS-HD Master Audio

Similar to the MPEG-4 SLS format, this format has a core track in an older, more widely supported format, DTS. This core lossy track is made lossless by a secondary track with correction data. It is an optional codec in Blu-ray implementations. Its main use is surround sound encoding, and as is the case with MLP, the price of the encoder ensures it is only used in mastering of Blu-ray discs.

LA

http://www.lossless-audio.com/

LA features an extremely high compression (on par with OptimFrog highest modes), but never made it out of beta stage and hasn't been updated for more than 20 years. Furthermore, backward compatibility is not guaranteed, so using it for archiving might pose a few problems. It isn't able to cope with file corruption either, software support is very limited and isn't open source.

MLP/Dolby TrueHD

The MLP codec (of which the mathematical basis was used in Dolby TrueHD) is the codec used for DVD-Audio. It was mandatory in any HD-DVD implementation and optional for Blu-Ray in its Dolby TrueHD form. It is known to support the 'wasted bits' scheme used in LossyWAV. As encoders are very expensive, its use outside DVD/Blu-ray mastering environments is non-existent. Its main use is encoding surround sound data.

MPEG-4 ALS

MPEG-4 ALS is the successor to LPAC, which it was based on. It has been as a ISO standard and there is a reference encoder/decoder, which also supports floating-point signals. It never gained much popularity. FFmpeg can decode some ALS files.

MPEG-4 SLS

MPEG-4 SLS is a special codec, having a AAC core track and a 'correction track'. Also known as HD-AAC, SLS stands for Scalable to Lossless. However, there is to date still no affordable software to play, encode or decode (the lossless part of) SLS files.

Shorten

http://www.etree.org/shncom.html

Shorten was one of the first widely-used lossless formats, and it still occasionally found on the internet, especially in archives, for example etree.org. It is quite fast in both encoding and decoding, but doesn't compress very much. Furthermore, seeking has a troubled past as well as tagging. It is considered obsolete.

Real Lossless

Just like WMA Lossless and Apple Lossless, it was created to fit in a codec suite, but unlike WMA Lossless and Apple Lossless, there is no hardware support and software support is limited. Compression is on par with most other codecs, but it is rather slow to encode. It is considered obsolete, and Real contributed to the creation of MPEG-4 ALS.

Oddball formats

There are a few archaic formats of which encoders and decoders are hard to come by. Most of those would have disappeared by now, but some of them are being preserved for posterity at rjamorim's ReallyRareWares.

See also

External links

Other lossless compressions comparisons Sorted based on last update date.

  • Martijn van Beurden's comparison - tries to compare all codecs and settings with a balanced pool of music, with separate sections on hi-res and multichannel material (last version 2023-08-29)
  • Squeezechart audio - tests as much codecs as possible, but not all their settings and with a limited test corpus (last updated 2013-10-31)
  • Synthetic Soul's comparison (last update 2007-07-28)
  • Johan De Bock's speed oriented comparison - best choices speedwise are indicated in green, mostly electronic music (last updated 2006-07-22)
  • Hans Heijden's -- used as reference to build the table (last updated 2006-07-07)
  • Josef Pohm's comparison, hosted by Synthetic Soul (last update 2006-05-29)
  • Bobulous' lossless audio comparison — a look at six lossless formats in terms of speed and file size (last updated 2006-05-22)
  • Jhan De Bock's size oriented comparison - aimed only at the maximum compression setting for each codec (based on a somewhat limited set of samples, however) (last updated 2006-05-19)
  • Gruboolez' -- comparing only classical music (last updated 2005-02-27)
  • Speek's (last updated 2005-02-07)
  • Lossless Compression of Audio Much information about oddball formats including comparison of them. (last updated 2005-10-21)

More on lossless compressions

References