I put together a collection of processed files with varying degrees of correction. You can listen to each of them until you find one just about right for you.
Be sure to take out your hearing aids, if you wear them, before these listening tests.
You can download the collection (100 MB) here. This zip file contains two folders – one for most people (0 dB MPEG) and another for the most serious degrees of impairment (-9 dB MPEG). The audio files in the first folder were recorded at -23 LUFS, while the -9 dB files were recorded 9 dB even lower because of the extra headroom needed for those serious degrees of impairment.
When you start up, there is a calibration file labeled EBU-reference_listening_signal_pinknoise_500Hz_2kHz_R128.m4a that contains a monaural file with pink noise in a band from 500 Hz to 2 kHz. This is the file you should play to set up your listening levels. The audio tracks are all 6 dB softer than this reference level, so you should set your amplifier to hear it as loud but comfortable. Then the audio tracks should come through at a not-so-loud and pleasant level for you.
[ if you normally replay with auto-leveling, you should disable that for these tests. All the files should sound the same level to you, despite the more heavily processed files producing a higher loudness scale reading. ]
For those -9 dB files there is a separate version of the noise reference level file. The same instruction apply, but since it was recorded 9 dB softer that just means that you need to crank up your gain by an additional 9 dB to hear these tracks at the same consistent level.
The tracks are copies of the song named Cobalt by the musical group Les Jumeaux. I chose this track out of a collection because it illustrates several hearing challenges for you. The tracks were all pre-processed with HdphX to keep headphone listening comfortable for you.
We start with the unprocessed reference Cobalt Ref vT00.m4a. The digits after the vT designation in the file names tell you what level of vTuning was applied. So the next level up is Cobalt vT10.m4a, which means that vTuning was set to 10 dB of processing.
This naming convention continues until we reach a level of vTuning = 50 dB at which point it becomes sensible to process through stacked Crescendos, with the actual corrective Crescendo running at the vT level, and the p designation indicating what vTuning was used in the pre-Crescendo. So Cobalt vT50p38.m4a specifies that the corrective vTuning was 50 dB, while the preprocessing Crescendo had vTuning = 38 dB.
In the 0 dB MPEG folder:
- EBU-reference_listening_signal_pinknoise_500Hz_2kHz_R128.m4a – the -23 LUFS reference level calibration track
- Cobalt Ref vT00.m4a – the unprocessed original track
- Cobalt vT10.m4a – 10 dB of vTuning, very light processing, often preferred by people with normal hearing
- Cobalt vT20.m4a – 20 dB of vTuning
- Cobalt vT30.m4a – 30 dB
- Cobalt vT40.m4a – 40 dB
- Cobalt vT50p38.m4a – 50 dB with 38 dB of pre-Crescendo
- Cobalt vT60p43.m4a – 60 dB with 43 dB of pre
In the -9 dB MPEG folder:
- EBU-reference_listening_signal_pinknoise_500Hz_2kHz_R128_-9dB.m4a – the -31 LUFS reference level calibration track
- Cobalt Ref vT00 -9dB.m4a – the unprocessed track again, but with 9 dB of additional attenuation
- Cobalt vT70p49 10kLPF -9dB.m4a – 70 dB of vTuning with 49 dB of pre-Crescendo, 10 kHz lowpass filter to remove the un-hearable and spiking transients that would have needed even more headroom, and 9 dB of extra headroom
- Cobalt vT80p55 10kLPF -9dB.m4a – 80 dB with 55 dB of pre
We stop at 80 dB of vTuning, because the next increment takes us to 90 dB which is the defining line for profound hearing loss. Nothing anyone can do to help you there.
Once you have your monitoring level set, have a listen to the unprocessed reference track. Then try each incremental increase in vTuning processing till you find one that’s closest to sounding right for you.
Admittedly, these 10 dB increments are a bit coarse and won’t exactly fit your hearing, but you ought to be able to find a best one in the pile.
If your hearing is so bad that even the 60 dB processing isn’t quite enough, then you’ll need to go over to the -9dB MPEG folder, set up your reference gain anew with the noise file, then have a listen to the reference unprocessed track, and try out the other two extreme processing tracks.
If you find yourself needing either the 70 dB or 80 dB processing, then you probably can’t hear much above 4 kHz either. But, of what you can hear, below 4 kHz, you should notice a dramatic improvement.
Now, on to the musical content. Here are some listening challenges for you…
At, 0:36 there is a lady whispering and a snare drum starts up with an exaggerated reverb tail on each beat.
At 1:13 the snare takes on a more sane reverb, but you should still be able to discern the faint reverb tail even there.
In the interval from 1:13 to 2:08 there are whisper voices mixed in with the honking synth.
At 2:10 a female voice begins talking audibly. I think it is in French, but the talking is intentionally indistinct.
Starting again at 2:30 there is a repeating whisper, with occasional female voice speech atop it and the music, to the end of the track. You should still be hearing the faint reverb tails on the snare drum. The track grows gradually louder, but only by about 2 dB by the end.
All through the track you should notice the brightness of the honking synth – the higher harmonics are strong. The sound has a bite to it.
At the end of the track, the reference (vT00) track shows an integrated loudness of -29 LUFS. At stronger levels of processing, the effective integrated loudness grows in response to your hearing needs.
But unlike someone shouting at you, and raising all their frequencies by a huge amount, only the higher frequencies are being lifted, and only just as much as you need to hear the sound properly – harmonics properly balanced with the fundamental tone, instruments should retain their character at all loudness levels.