I’m curious about MP2 and MP3 encoding (or decoding) with mAirList. I realize that during playout, there are codecs decoding these files for sound card output (provided the soundcard doesn’t use onboard converters). I’m curious whether these codecs are part of the Windows operating system or are specific to mAirList. More curiousity than anything else. Thank you…
All decoding is performed by mAirList. The decoders are part of BASS and its optional add-ons (check out www.un4seen.com for a list of available add-ons).
One special case is MP3 - for this format, BASS (more precisely, the “mp3-free” version of BASS that ships with mAirList) does not have its own decoder, but it uses the ACM codec installed in your Windows system to avoid patent infringement (the ACM codec from DirectX is already licensed and paid for by Microsoft).
In any case, only linear PCM data is passed to the sound card. Any hardware decoding mechanism that may be included in your hardware won’t be used by mAirList/BASS.
Regarding the encoding of MP3 files (for mixdown and streaming), things are even more complicated. You have to obtain a license to use MP3 encoding, but that may not be the case for all countries, depending on the local patent laws. mAirList does not ship with an MP3 encoder at all. The user is required to download and install the encoder (lame.exe) himself, having verified that it’s ok to use it in his country. mAirList does not depend on an MP3 decoder, but you will need to use a different format for streaming (Ogg Vorbis) and mixdown (WAV or Ogg Vorbis) if you cannot use an MP3 encoder in your country.
Thank you for that information. In theory, for 24/7 playout, would the system be more “stable” if only linear PCM files were used? My thought is that it would be one less resource mAirList had to process non-stop over a period of perhaps months.
With the cost of hard drive storage dropping, it seems that Linear PCM will eventually be the best choice regardless.
On a modern PC, the CPU load for decompressing an audio file can hardly be noticed. And smaller files mean that you need less hard disk access, or mAirList can even cache the file in RAM during playback (check out the File Management options in mAirListConfig).
Have you considered using a lossless codec, e.g. FLAC? That should give you a good trade-off between high quality (perfect quality actually, because it’s lossless) and mid-size files.
I’ll have look at the cache option. I’m aware of FLAC but have never used it. I’ll look at that as well. This question was originally mostly out of curiousity but now I want to do a little exploring
FLAC is pretty cool. It’s lossless, so when you deccompress the file again, you get the exact same WAV file as before, bit by bit. The compression is not as good as with MP3, of course, but as you said before - disk space doesn’t matter so much these days.
Okay, I had a look at BASS and FLAC. I think I understand now.
When FLAC is decoded, does the sound card “see”, receive, and then output linear PCM? In other words, the sound card will “see” a linear PCM file? I infer it does…
However, what does a sound card “see” and output after a lossy file, such as mp2 or mp3, is decoded by BASS?
Thanks for the teaching in this area. It’s an interesting conversation and hopefully an education to others who read this post…
As I mentioned in my post above, mAirList (or more precisely, BASS) does always pass linear PCM data to the sound card. All decoding is performed in software.
When using a lossless codec (e.g. FLAC), the decoded PCM data is exactly the data that was in the WAV file from which the encoded file was created.
With a lossy codec (e.g. MP3, MP2 or Vorbis), the decoded PCM data is only “similar” to the original file. That’s the concept of lossy codecs - they remove part of the information in the original file that is not so important when a human is listening to the audio. In particular, the codecs remove frequencies that the human auditory system cannot hear anyway.
Thank you… The third paragraph answered my question when you mentioned:
With a lossy codec (e.g. MP3, MP2 or Vorbis), the decoded PCM data is only “similar” to the original file.
I was wondering whether the card did, in fact, “see” linear PCM, although only a similiar form of it. Thanks again for the instruction. You’re a great teacher.
Alec: yes, your sound card always ‘sees’ linear PCM from mAirList (more specifically, from the BASS engine used by mAirList). That’s always, regardless of the source audio file storage/compression format.
As Torben says, if your audio files are stored in FLAC (or some similar genuinely ‘lossless’ format), the PCM generated by mAirList/BASS will be identical to the same audio stored as a WAV file.
If your audio is stored in (for example) MP3, there will be a SLIGHT loss in audio quality compared to the original, depending on things like the bit rate used when the file was encoded.
However, on an FM radio broadcast signal, MP3 files encoded at 128 kbps will be almost impossible to tell from the same audio stored as a WAV file when it gets to the listener’s receiver (which may well be a crappy portable radio or a cheap car stereo: not exactly ‘audiophile!’), due to the limitations of the FM system, and moreso if (like every other FM station) you ARE using a signal processor of some sort to improve the audio delivered to your FM exciter (aka ‘transmitter’).
For audio transmitted by Internet, again you’d be surprised if you checked the actual bitrate being used to serve ‘radio stations.’ Even the BBC’s iPlayer rarely goes beyond 96 kbps.
So to sum up, I wouldn’t get hung up on this. Some ‘experts’ will try to tell you that your audio ‘must’ be stored in WAV or some other lossless format instead of ‘lossy’ formats like MP3, but the truth is that both will sound the same (to 99+% of listeners) after you’ve broadcast them on FM, or on the Internet. They WILL sound a little different in the studio, but what your listeners are hearing off-air is NOT the sound you get in the studio!
With the price of hard-drives now, it’s really a no-brainer to use PCM WAV as a music storage format. Whilst MP3 (at a shocking 128k/bit) may sound OK on an FM broadcast, the MPEG Layer-3 format doesn’t take too kindly to the encode/decode cycle (the professional variant, MPEG Layer-2, ie: MP2) is a bit more resilliant not to mention the encoding is done in a different way. A 128k MP3 shoved through one of Mr Orbans breeze-blocks isn’t going to sound great!
…and on a related note, anybody who thinks a 128k/bit DAB station sounds better than FM needs a poke in the ear with a cotton-bud ;D
The BBC are linear for their music delivery - 44.1KHz at most sites (although some regionals are using 48KHz). Whilst FLAC is a good format for domestic use, many broadcast vendors wouldn’t go anywhere near it. MP2 was the standard for many years - with pro sound cards offering hardware decoding (which was good because a Pentium 133 was considered “top notch”) and the price of hard-drives meant that only rich people could have 100’s of GBs to play with.
With a 1TB SATA drive at under £40 inc VAT, why not just use WAV ?
I’m not saying that you should use MP3, Charlie, what I’m saying is that you don’t have to re-rip your entire library of MP3 files to WAV ‘because it’s a better format to use.’
In other words, if you already have 10,000-plus fully-tagged MP3s, I don’t feel you should suddenly think ‘Argh! I must re-rip all of them to WAV immediately!’ By all means start using WAV from now on, and maybe re-rip some of the more frequently-played items, or any where the quality difference is really noticeable compared to your ‘new’ WAV files.
Unless, of course, you do have staff/interns/work experience droids available to do that for the next month or three … ? ;D
…and on a related note, anybody who thinks a 128k/bit DAB station sounds better than FM needs a poke in the ear with a cotton-bud
That quote made me laugh!
Cad and Charlie-
Thank you both for the additional info… It’s been an interesting conversation.
Interestingly, just yesterday, I spoke with an engineer who echoed Charlie’s comments: the trend is slowly toward WAV but that many stations are still using mp2@384kbps (or 256) as well as sound cards with onboard converters. The onboard converters are probably a carry over from the days of slow CPUs and limited system resources.
[quote=“Charlie, post:12, topic:7083”]With the price of hard-drives now, it’s really a no-brainer to use PCM WAV as a music storage format. Whilst MP3 (at a shocking 128k/bit) may sound OK on an FM broadcast, the MPEG Layer-3 format doesn’t take too kindly to the encode/decode cycle (the professional variant, MPEG Layer-2, ie: MP2) is a bit more resilliant not to mention the encoding is done in a different way. A 128k MP3 shoved through one of Mr Orbans breeze-blocks isn’t going to sound great!
…and on a related note, anybody who thinks a 128k/bit DAB station sounds better than FM needs a poke in the ear with a cotton-bud ;D
The BBC are linear for their music delivery - 44.1KHz at most sites (although some regionals are using 48KHz). Whilst FLAC is a good format for domestic use, many broadcast vendors wouldn’t go anywhere near it. MP2 was the standard for many years - with pro sound cards offering hardware decoding (which was good because a Pentium 133 was considered “top notch”) and the price of hard-drives meant that only rich people could have 100’s of GBs to play with.
With a 1TB SATA drive at under £40 inc VAT, why not just use WAV ?[/quote]
Couldn’t agree more. Had this discussion about three years ago IIRC!