Friday, 3rd of August

02007-08-03 | Uncategorized | 9 comments

Checked the new DRM-less iTunes Plus (256kbps mp4 files) out for the first time. Noticed that the top two albums in the New Age genre are mine, and 4 out of the top five singles. Nice.

After Jon and I compared the quality of AIFF, FLAC and mp3 (at 320kbps) a couple of weeks ago and found no obviously perceptible difference between them, I decided to discontinue AIFF files in our ListeningLounge. They will be available to those people who bought songs or albums in the past for another month. I see the LL getting quite a bit larger over the coming months and this will save a lot of space – almost 800MB per album – and bandwidth. We will still serve better quality than iTunes Plus for less money: we will continue to give you the choice of 128, 192 and 320kbps LAME encoded mp3s for a buck per song, while iTunes sells the 128kbps file for 99 cents and the 256kbps file for $1.29. Looking at the LL download stats confirms my decision as well – downloads of the 128, 192 and 320kbps mp3s outweigh the 256kbps mp3 and AIFF files by a lot.

9 Comments

  1. steve

    Will FLAC files be available?

    Reply
  2. ottmar

    Steve: I don’t think so. The difference between FLAC and a 320 kbps mp3 is really hard if not impossible to notice and we found that meta-data can be a problem with FLAC. The other problem is that iTunes is the 800 pound gorilla. iTunes is the 3rd largest music retailer in the U.S. – larger than amazon.com. That means most people will want files that are compatible with iTunes. If I download a FLAC file I have to first decode it to an AIFF file and then re-encode it as Apple Lossless or a high-quality mp3.

    Reply
  3. Curt

    This is good news!

    Reply
  4. steve

    Yeah … that was my idea: download the FLAC and encode it using MAX or xACT to Apple Lossless, for the iPod.

    But primarily, for my home system (a Mac mini w/ 500Gb external HD connected to the sound system) I’d just leave it in FLAC and play it back as FLAC file using COG.

    As an audiophile/open source geek, I tend to prefer lossless formats like FLAC, etc…

    But, hey, not everyone is a geek, and I understand you are trying to balance being pragmatic, running a business, and offering high quality audio.

    -s

    Reply
  5. Panj

    Kudos on the Top 2 and the 4 out of 5, Ottmar. I can’t think of anyone on the planet who can top you in quality…cream ‘does’ rise to the top! Hopefully I will be blessed to attend more of your live concerts, your Music is so healing. God Speed on your Journey!

    Reply
  6. ottmar

    Steve: I took an AIFF (exhibit A) and used xACT to encode the FLAC and LAME to encode the mp3. Then I used xACT to turn that FLAC into an AIFF (exhibit B) and Peak to turn the mp3 into an AIFF (exhibit C). Then I loaded those three AIFF files into a new ProTools session, which allowed me to select a section of the song and switch between the three files seamlessly.
    1. listening to one song at a time one would not be able to hear a difference – one had to conduct the test as I described, allowing instant switching between the files
    2. exhibit B (FLAC) did not appear to be lossless. It didn’t lose any highs, but there was an ever so slight loss of bass. Exhibit B and C were more similar than B and A.

    Reply
  7. steve

    Regarding item 2, what I find most interesting about this, is that the high bit rate mp3 sounded more like the source file than a “lossless” transform: that’s FASCINATING! Or do you mean, LOOKING at the waveforms? (i.e., you mentioned “appear” – did you inspect these with a ‘scope or a spectrum analyzer? I suspect I’m reading too much into the word “appear.”) Could this be an artifact of the transform? (i.e., the algorithm used in xACT?) Wow. So many questions, and ideas come to mind!

    My second reaction is, “I wonder how a high bit rate mp3 (320kbps) and a high bit rate Ogg Vorbis (320kbps) would compare to the source file?” which is to say, how does the commercial format and the open source format compare?

    Reply
  8. ottmar

    Maybe I was not clear. I meant to say that the FLAC-AIFF sounded more similar to the mp3-AIFF than it did to the AIFF-AIFF. In other words, the diff between the FLAC and the 320mp3 seemed smaller than the diff between the FLAC and the AIFF.

    Yes, I would love to compare OGG and mp3 and maybe I will do that. I personally would love to see iTunes support both FLAC and OGG.

    Reply
  9. steve

    Ottmar:

    Apparently there is empirical evidence demonstrating a near perfect overlay between a PCM linear file, and a 256kbps LAME encoded mp3 file up to 17kHz. I don’t recall the test setup, or the version of LAME that was used.

    I suspect that if the bit rate is increased another 25% that the high end response might approach 20k, though I am speculating. I have no empirical evidence to back up that suspicion, but the math involved would seem to intimate that such a response would be possible in theory.

    LAME is a pretty impressive encoder, and, since it’s open source, COUNT ME IN!

    Reply

Leave a Reply to ottmar Cancel reply

Your email address will not be published. Required fields are marked *

Archives

Images

Social

@Mastodon (the Un-Twitter)