Audio converters, FFmpeg, Opus tools, what tools do you use?

Discussion in 'Software' started by GabsIT, Jun 13, 2021.

  1. GabsIT

    GabsIT Producer

    Joined:
    Oct 29, 2020
    Messages:
    204
    Likes Received:
    133
    Location:
    Asia
    I had been updating my knowledge in the latest months and I found many useful things.

    Again the idea of this thread is to share what tools we use and some pro and cons, now Internet is full of content about this but not really at a pro level so there is a lot of misinformation about it.

    Now starting from basics:

    The best audio compression taking into consideration quality,space,speed of conversion and streaming buffering is simply Opus, that usually can use .ogg and .opus file extension and there is no reason to have no compatibility as it's actually pretty old already 5-10 years.

    There is no magic coding that will make compression any better, and why is better than any other codec? is simply because it's open source and use two of the best audio codec technology one use for a voice range of audio and another to compress the higher frequencies until 20khz (maybe beyond, not clear at the moment of write this post) more info at the bottom.

    Now that is in a nutshell but to complement what I am saying and if you want to read or compare by yourself keep reading
    Tl;dr

    Codec comparison, quality vs bitrate, quality vs streaming latency https://opus-codec.org/comparison/

    Lossy formats public test https://listening-test.coresv.net/results.htm

    The next tab give a very good idea of Bitrate vs use
    acreenshot.8.jpg
    Source: https://wiki.xiph.org/Opus_Recommended_Settings

    Also it's said that opus 64 KB/s is equal or better in quality to Mp3 128 KB/s
    I personally prefer the Opus at 128 KB/s as is preserve all the higher frequencies, could be compared to a Mp3 256KB/s (while most people can't tell the difference after Mp3 160 KB/s)

    In the next table is shown the frequency chop down of 96, 128, 160, 192, 256, 320 KB/s of Mp3 (or other formats but not tested at the moment of write this post) also not VBR encoding

    fakinthefunk-03-spectrum.png
    Source: https://fakinthefunk.net/en/#download useful app to determine the quality based in where the chop down of higher frequencies were made. (search fakin the funk in the sister site or download directly from this link)

    Why Opus is so special?

    Opus is more than just two independent codecs with a switch.

    In addition to a Linear Prediction SILK mode and an MDCT CELT mode it has a hybrid mode, where speech frequencies up to 8 kHz are encoded with LP while those between 8 and 20 kHz are encoded with MDCT. This is what allows Opus to have such high speech quality around 32 kbps.

    Another advantage of the integration is the ability to switch between these 3 modes seamlessly, without any audible "glitches" and without any out-of-band signalling.

    Source: https://wiki.xiph.org/OpusFAQ

    I will edit the content to keep all the apps, tools and some scripts, useful stuff.

    At the moment of writing this I am using command line tools, FFmpeg and opusenc, maybe also include other tools or cases as android freezer, AI tools to separate audio, etc.
     
    • Interesting Interesting x 2
    • Like Like x 1
    • List
  2.  
  3. phumb-reh

    phumb-reh Guest

    To me, the issue of audio compression has pretty much become moot these days. I keep my stuff lossless and can deliver that in any format necessary.

    If I need to deliver, say, WIP raw mixes/renders or somesuch, I've used AAC for a long time since that's works for anyone and has a good quality/size ratio.

    The choice of codec for streaming platforms is pretty much out of my hands, and since Opus became an IETF standard it's already used under the hood already in many places, so again, can't change that.

    If I were setting a VoIP/streaming service Opus would be codec of my choice, with perhaps AAC fallback.
     
    • Like Like x 1
    • Useful Useful x 1
    • List
  4. BEAT16

    BEAT16 Audiosexual

    Joined:
    May 24, 2012
    Messages:
    9,082
    Likes Received:
    6,990
    [​IMG]

    MediaInfo v21.03 - Graphical user interface, with installation routine, for Windows.
    Other versions (operating system, user interface ...) are also available (Microsoft Windows macOS Debian Ubuntu Linux Mint Raspbian RedHat Entreprise Linux CentOS Fedora openSUSE ArchLinux Android iOS)

    MediaInfo provides technical and additional tag information about video and audio files.

    What information does MediaInfo provide?
    - Basics: title, author, director, album, track number, date, playing time ...
    - Video: codec, aspect ratio, frame rate, bit rate ...
    - Audio: codec, sampling rate, channels, language, bit rate ...
    - Text: language of the subtitles
    - Chapters: number of chapters, chapter overview

    Which formats (containers) does MediaInfo support?
    - Video: MKV, OGM, AVI, DivX, WMV, QuickTime, Real, MPEG-1, MPEG-2, MPEG-4, DVD (VOB) ...
    - (Codecs: DivX, XviD, MSMPEG4, ASP, H.264 / AVC, H.265 / HEVC, FFV1 ...)
    - Audio: OGG, MP3, WAV, RA, AC3, DTS, AAC, M4A, AU, AIFF ...
    - Subtitles: SRT, SSA, ASS, SAMI ...

    What can the program do?
    - Reads many video and audio formats
    - Various display options for the information (text, table, tree structure, HTML ...)
    - You can adapt these representations yourself
    - Information can be output as text, CSV, HTML ...
    - Graphical user interface, called up via the command line or via the DLL
    - Integration in MS Windows (drag'n'drop, context menu)
    - Internationalization: Every language is supported under every operating system

    License
    This is an open source software - with the exception of the GUI for Mac - that is, the possibility of free distribution of the source code and derivative works (BSD-like license).

    Code:
    https://mediaarea.net/en/MediaInfo
     
    • Like Like x 1
    • Agree Agree x 1
    • List
  5. GabsIT

    GabsIT Producer

    Joined:
    Oct 29, 2020
    Messages:
    204
    Likes Received:
    133
    Location:
    Asia
    To me space is still a issue, as with libraries with 3D surround or HQ ambisonic 4-5 channels 24-32bit 96KHz-192KHz few minutes can turn into GB, I had no problem converting from WAV to Flac 16bit 48KHz but I had issues with 24bit also many libraries probably by marketing ad very high values but analyzing the files sometimes are filled with pink just noise or chopped down to X KHz... and well I still don't understand if Flac converter in FFmpeg chose the better value based on the real quality, but in some test some 32bit audio after converting from Wav to Flac and back to Wav was converted as 16bit, I reversed the signal of the converted audio and played along with the original one and 0 differences, no idea why.

    Stuff like this is what is hard to find on internet. Also I wasted a lot of time to try to insert the album art to opus files, it's hard lol.

    Flac is compatible with almost any DAW so no problem to use it as storage or for production.

    The latest 6 years I've being recording monks chantings and also natural atmospheres, and maybe I will go pro and get HQ ambisonic microphone and equipment, before I start to travel again (hopefully Covid will end) so I can't carry much stuff with me, actually I used a very low opus setting to convert dhamma talks into opus file with an amazing quality vs space.

    So I am between Wav, Flac and Opus no more Mp3 or M4a
     
    • Interesting Interesting x 1
    • List
  6. BEAT16

    BEAT16 Audiosexual

    Joined:
    May 24, 2012
    Messages:
    9,082
    Likes Received:
    6,990
    Best Way is: Don't rip wav files use Flac or Apple losssless --> Ripping Program: dbPowerAmp --> Tag the files: mp3tag

    Formats for Internet Distribution
    Audio files can get very large in their original created state, so some method must be used to make them smaller and more manageable. This is the only way they can be sent over the Internet. This method is known as data compression. At some point in the distant future, when everyone is hooked up to a fast broadband internet provider, large files will no longer be an issue. But to the present Time is data compression in the only way for a successful online transmission.

    Data compression
    Data compression is a process that uses psychoacoustic principles to reduce the number of bits required to represent an audio signal. Data compression is currently used because regular LPCM files are so large that they cannot be transferred or stored online quickly and easily enough. Data compression reduces the physical storage space required to store a sound. Therefore it also reduces the timewhich is required to transfer a file.Data compression can be lossy. This means that the sound quality - as already discussed - is reduced or impaired by the compression. But it can also be lossless, which means that there is no loss of quality when the file is decompressed again. A number of methods or codecs (derived from the words encoder / decoder) are used for data compression, all of which have a different sound and a different purpose.

    1.) Lossy compression method: MP3, AAC, MPEG-4, Windows Media Audio, Ogg Vorbis (.ogg) u-law
    2.) Lossless Codecs: Apple Losless, FLAC,

    Source: Bobby Owsinski - The Mastering Engineer's Handbook: The Audio Mastering Handbook
     
Loading...
Similar Threads - Audio converters FFmpeg Forum Date
SSL 12, new 12-in/8-out USB audio interface with 32-bit converters Soundgear Dec 2, 2022
Checking Out: Hearth and Hollow Folk Voices by Spitfire Audio Software Reviews and Tutorials Today at 3:03 PM
Audio Bro LADD $99 when purchase their free Silver - worked for me! Software News Today at 1:15 PM
NEW: Fluffy Audio - The Zen Garden (Kontakt ) Software News Today at 10:03 AM
Acustica Audio Bundle Selling / Buying Today at 1:56 AM
Loading...