Steve's Transcode Test

Update2: The test end has been rescheduled to indefinitely. I have not received enough results to form a meaningful conclusion. If you're interested in the results, please take the test and I will tell you what you heard and how it compared to others. If I have received enough results, I will post a public results page.

With the rise in popularity of hardware mp3 players, mp3 has become the standard by which people store their CDs onto their computers. Unfortunately, the convenience of hardware players has discouraged the adoptment of other potentially better codecs, such as Ogg Vorbis or Musepack. Those who have a hardware mp3 player must choose between storing their music with either lossless codecs or in mp3. A third choice is to transcode from the other codecs. This is considered blasphemy in audiophile circles. The objective of this test is to help determine whether that notion is valid.

This test's specific purpose is to determine how poorly various codecs transcode to a Lame --preset-medium encoding

The encoders and parameters tested are:



  1. This test is REALLY hard. I can't hear a difference between most of the samples. Do you still want my results?

    YES! Part of this test is to see IF one can hear a difference between a transcode and an original. It's perfectly alright if you can't hear the difference, and your results file is still important.

    It is also more important that you TRY to hear the difference. The results files show ABX attempts, and that's more important than the actual score

  2. Who should take this test?

    Anyone interested in audio for hardware players, or people that have no interest but would like to help making this test better. You don't need excellent hearing, but some good gear is welcome (headphones are must-have).

  3. How do I take this test?

    Download ABC/HR + the readme file + decode scripts from one of the following locations:

    Further instructions, and links to the sample packages, are inside the readme.

    The average size of each sample package is 3.8MB. It sums to 34.1MB, all the 9 packages.

  4. I'm on Linux/MacOS/something other than Windows. Can I take the test?

    Yes, but you need to have Java MADPlay and maybe Wavegain installed. Included in start.zip is an ABC/HR program written in Java.

    If you DON'T have MADPlay or Wavegain you'll have to decode the mp3 files yourself. If you do this, please mention it when you mail me the results.

  5. When will the test finish?

    The test is scheduled to end on November 12th, 2003. However, it may be extended as needed.
    Update: Due to popular demand, the test has been rescheduled to end at Nov 30th 11:59

    The test has no scheduled end. It will continue until I have enough results to publish meaningful results online. However, I will explain test results to all test participants. So if you're interested in the results, please take the test.

If you have any questions, please don't hesitate to contact me.

Thank you very much,
Steve Hayashi

To contact me type in: "sthayashi" followed by the 'at' symbol (Shift+2 on my keyboard) with "alumni" 'dot' "cmu" 'dot' "edu". It has been explained this way in order to avoid getting put onto SPAM mailing lists.


I would like to thank the following people: