Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
James May: 'Hold The Line' kapu mix
#11
Hah. I'm just writing this stuff here, so that people having second thoughts about using EBUR128 in music production or distribution might take some confort in the fact that there are other people sharing the view. In my opinion, it's just too ... coarse or 'rough' for normalizing masses of music. And this of course has nothing to do with the pure master smashing and squashing, and implementing the EBUR128, or any other standard, in 24-bit playback or production requires both normalizing the material and the calibration of the speaker levels. And there isn't anything revolutionary in the EBUR128 recommendation. Before the EBUR128, the SMPTE reference and calibration levels was widely (the only) in use, which actually provides more headroom and has enabled production of very dynamic material in the digital domain for ages. Probably because the industry standard Pro Tools HD I/O converters are set to this calibration at the factory, although this might no longer be the case. The problems came with the small independent studio scene, where after setting the gain structure 'by the book' the system seems overly loud when playing 16-bit masters at unity level. The other 'false' intuition is that when starting mixing with a properly calibrated system, the channel levels mostly go down, if the tracks are initially recorded at good signal-to-noise ratio, and this would lead to degraded audio quality, and in theory it does, but in reality even the best converters aren't capable of the reproducing the full dynamic range of 24-bit audio, because it's practically impossible to build such equipment in the conditions of planet Earth due to the behavior of copper molecules used in the electronics, or something like that. This notion of the degrading audio quality at lower levels then leads to an idea of building the levels upwards against the 0 dBFS limiter and taking down the speaker level accordingly, resulting in a more quiet calibration with less headroom. And many people 'abuse' their systems in this manner still making great sound, because it's practical when having to produce 16-bit 'pseudo masters' straight out from the main mix bus. It's just that the benefits of the expensive gear and high headroom are mostly wasted, although in reality there are very rare occasions where the end user can benefit of such things. People also have a hard time fitting the two facts together, that properly set 24-bit 'dynamic' systems are louder by definiton, or in a technical sense capable of producing those ear destroying peak levels (the headroom) than -10ish reference level matched 16-bit systems. Probably because of the 'falsely' intuitive trail of thought; 'I want make dynamic material > Loudness needs to be avoided at all cost > My overall system and peak levels seem loud > Speaker levels need to go down' and the engineer starts to tweak the system too much by ear based on 16-bit dynamic reference, and if the reference material is detailed, subtle and dynamic, the ending result is still significantly more quiet with less headroom than the 'textbook' calibration. Then the final 'error' is start following the -23 LUFS or RMS, or what ever 24-bit recommendation, with this new personal calibration level, where now the ear tells the engineer to go louder, but the meters tell the opposite; the 16-bit yet dynamic Al Schmitt engineered reference CD was probably around -15 LUFS, and this what the ears (correctly) are guiding the engineer towards, but the conscious mind orders him to follow -23 LUFS. Then the engineer tries to make the mix sound like the -15 LUFS reference material to his ears yet still somehow it should measure integrated -23 LUFS in the meters, and this results in just bad sound at any listening level, because the whole workflow, setup and calibration is now ... distorted between senses and conscious mind, and the overall level is just too low to make right judgement and decisions. The better practice of course would have been to start working by ear with the 'textbook' calibration and conform the reference material levels for this setup. As we have correctly done, and can now feel good about ourselves. ^_^

I guess the VU meter can be viewed as just another type of RMS meter, where the 0 VU is calibrated at the gear nominal operating level, or 'sweet spot', usually +4 dBu. Then there is of course some headroom above this for peaks, as VU meter is rather slow. About 'breaking the rules.' If I remember correctly Geoff Emerick almost got fired from Abbey Road or something when he experimented with close micing the drumset and setting the gain levels by the slow VU meter, resulting in the distortion or clipping of the peaks in the circuitry or tape input. This was viewed as unnatural approach and abuse of the mics and equipment, although Emerick probably did this because the end results just sounded closer to his perception of the banging backbeat. I believe he then applied these methods and ended up engineering some really bad sounding financial failure albums like The Sgt. Peppers Lonely Hearts Club Band from The Beatles. Internet, this was a joke.

In Finland, the national public broadcast company seems to want the material at -23 LUFS with -1 dBFS True Peak levels, and recommend that speech or dialoque is mixed -23 LUFS, and the max permitted loudness range is 20 LUs for surround and 15 LUs for stereo, and range should be greater than 5 LUs for program material longer than 2 minutes. This is for 24 bit linear PCM. Basically adopting the EBUR128 'such as' for TV and recommending the film industry mixing strategy. For music they seem to want 'high quality CD or MP3'. They seem to be streaming their radio channels at around -12 LUFS max with approximately -2 dBFS peak levels, at least for the spoken word channel Puhe and contemporary pop channel YleX. ^_^
Reply