Female vocals metallic overtone. The cause?

Discussion in 'Audio Hardware' started by pdxway, Feb 15, 2019.

Thread Status:
Not open for further replies.
  1. John Dyson

    John Dyson Forum Resident

    Location:
    Fishers, Indiana
    First -- let me preface -- I am not selling my SW here. I don't think that it is right to do as I am a pure techie who wrote the program. We have a project structure where my project partner is the official interface to the recording/audio community.

    You have the concept correct about using the DolbyA unit. Usually in the professional environment where the output of the tape is used directly, the DolbyA unit is intimately connected to the recorder. However, when material is subsequently to be used in studio or distribution (to the presssing facility), then the material is often NOT decoded until the endpoint distribution mechanism. (I think that the digital endpoint has no mechanism for decoding, unlike the vinyl endpoint.) This missing part of the digital process is probably why the DolbyA leaks occur (thereby causing the metallic/tinny sound on digital recordings.)

    So, normally, in a consumer application -- whicih would be uncommon, the DolbyA unit would be used as a loop on a pre-amp, not directly to the stereo amp. When using DolbyA HW you want level controls in and out (which are not provided on the unit), because the levels need to be set correctly by a tedious listening procedure (because there is no DolbyA tone on consumer material.)

    For using the software, your idea is correct about 100%. The software allows setting the DolbyA level as a command line option. When I have the GUI written, it will be able to be tuned during the decoding run (making it easier to both use the tone and/or listen for optimum decoding.) Right now, when testing, I have to re-issue the command line until I hear the correct sound, and the decoder can be used in real-time on a relatively fast, recent intel (or fast AMD) CPU -- 4 core or more computer. At the highest quality levels (better than a real DolbyA HW), the decoder barely runs 2X or even barely 1X faster than realtime for the very highest quality mode on a fast computer, but for sounding 'roughly the same' as DolbyA HW mode, it runs about 8-10X realtime in speed. (The decoder always uses careful shaping of the gain control signal, but in the higher quality modes actually scrubs the signal in a very secret/proprietary way to mitigate some of the gain control caused (specifically) IMD distortion components.) The 'scrubbing' takes huge amounts of CPU (4 cores of a recent Intel/AMD box is A LOT of computing power), and is really helpful in making the near-master-tape sound rather than the DolbyA characteristic sound.

    I normally do this when testing the program:
    sox infile.flac --type=wav - rate -v 96k | da-avx --tone=-13.50 --outgain=-3.0 --floatout --info=10 | sox - outfile.flac gain -n -1.0
    or, the same command line, but using 'play -q -' instead of sox for the last item in the pipeline. This command also provides for a running log on the terminal of the gains & levels for each band and each channel.

    The '--tone' switch allows specifying the magic calibration (instead of messing with the input and output signall levels as on a real HW unit), so it is more convienient for the SW, given that the --tone level sometimes needs to be determined by ear -- and it can be tricky, as I sometimes get fooled also. The good news is that leaked digital recordings often have the --tone= level in the -13.XX dB range. It must be the standard way of using the real DolbyA HW for distributing recordings. (I have sometimes been embarassed by using the wrong --tone= level -- and this is one reason why my program is not targeted to the consumer -- it is hideously difficult to use without DolbyA tones.)

    John
     
    The FRiNgE and pdxway like this.
  2. Brother_Rael

    Brother_Rael Senior Member

    Not on this album. Had it for years, lovely sounding disc. The recording isn't the problem here.
     
    SandAndGlass and pdxway like this.
  3. John Dyson

    John Dyson Forum Resident

    Location:
    Fishers, Indiana
    That recording just sounds like a bit of close miced sibilance -- probably either a mic or HF boost. I don't hear fast HF compression on that recording -- just normal processing.
     
    Rockinrob, Brother_Rael and pdxway like this.
  4. avanti1960

    avanti1960 Forum Resident

    Location:
    Chicago metro, USA
    I hear it with (very warm sounding) headphones- the metallic overtones are in the trailing edges of her voice. Sounds like "air" that came along with the recording process. In this case it isn't "objectionable" to my ears though.
    Different recording and playback components can minimize this or make it sound worse. If you really want to hear "metallic" try a set of KEF R series speakers (for example).
     
    pdxway likes this.
  5. harby

    harby Forum Resident

    Location:
    Portland, OR, USA
    Close-mic only really affects the lower bass, the "announcer" deep voice you get from being close, and a higher chance of wind noise plosives.

    Looking at videos of the live Diana Krall Paris performance, it looks like a Neumann KMS150 small-diaphragm hypercardioid condenser mic was used, which has an even sharper bump in the high frequencies.

    [​IMG]

    The mic was probably run into a board with a sound reinforcement pedigree. If a singer uses a chest voice and nasally mouth vowel shapes that diminish the natural harmonics, the engineer may pump up the treble more which emphasizes a siblant rattly tone to the voice.
     
    Kyhl, Rockinrob, AaronW and 4 others like this.
  6. The Pinhead

    The Pinhead KING OF BOOM AND SIZZLE IN HELL

    No metallic overtones on any of my 5 Enya albums. Maybe on my Joan Jett's and Lunachicks, but that's expected:laugh:
     
    bever70, pdxway and The FRiNgE like this.
  7. The FRiNgE

    The FRiNgE Forum Resident

    I have no clue on Richard Carpenter's recording techniques, nor any other personnel at their sessions. By whatever means Karen's vocals were processed, Dolby A pre-emphasis in a sidechain, or simple eq, the effect shouldn't be overdone. I have noticed Karen's vocals on some of her songs are a bit "over the top" so much above 10kHz that the harmonics sound less elevated, and more into harshness. I have found in mastering my own music projects, the right amount of eq is like a fine recipe, too much or too little quickly spoils it. Voice has very little harmonic energy beyond about 12kHz, so the small amount of harmonics that normally roll off sharply must never be boosted too much. I have found that slight attenuation above 10kHz, can sometimes sound more fine and silky.... which means there was perhaps too much.
     
    Dave and pdxway like this.
  8. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    Haha, yeah, thanks for checking! It is more acceptable with my pretty warm headset. But with my Paradigm and playing louder than 85db average, that metallic overtone of female vocals can be painfully obvious.
     
  9. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    I think post #55 would explain what I heard?
     
  10. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    Thank you for checking!

    Really not trying to be negative, but simply curious why some people hear the metallic tone and some don't. May be due to different hearing sensitivity to certain frequency? May be hearing range is not that high due to age?

    I mentioned in my other thread that I recently checked my hearing with tone generator in my phone with headset. I didn't like certain frequencies (around 6 to 8K, found it more irritating) and I started to slowly lost sensitivity after 13K and pretty much not hearing much after ~16.5KHz.

    If you don't mind I asking, how is your hearing range?

    Thanks!
     
    Last edited: Feb 16, 2019
  11. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    Wow, thanks for the detail explanation! Based on your post #53, it means your decoder would not solve certain issues like overboosted mic in the 8 to 15K range, nor extra high frequency boost during signal processing, etc, am I correct?
     
  12. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    Your explanation makes sense. But I wonder those are enough to explain why many female vocals have metallic overtone issue. May be overprocessing using software program during the mastering process could also add to the metallic tones?
     
    Last edited: Feb 16, 2019
  13. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    Hmm, you reminded me about my brief observation in a music composer place. He was basically listening to every second of vocal with his headset and fine tune the vocal with his computer! I guess many vocal album can be produced like this? All highly tuned and highly processed?

    Anyway, as I recall, I could not listen to his eventual production album. Could not stand the very obvious vocal metallic overtone!

    I wonder may be it was the computerized tone processing that was partially adding the metallic tones.
     
  14. Dave

    Dave Esoteric Audio Research Specialistâ„¢

    Location:
    B.C.
    I'm going with the digital mastering. All too often the 10khz is boosted, even subtly, and it causes the effect you are describing with female vocalists especially. Another sign of this is over accentuated tape hiss on analog to digital recordings.

    To eliminate your equipment as the problem take any CD pressing you know for certain is a good non-boosted mastering and listen carefully to see if it has this effect. If not, well you've solved the problem. You also can't blame the media source itself because there's enough evidence these days to show it doesn't hold up.
     
    pdxway likes this.
  15. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    Thanks! I am thinking the issue is due to combination of the microphone, the sound board, the digital processing, etc.

    I was given quite a few YouTube examples that actually are pleasing to my ears. I have hopes for better vocal experience now. :D
     
    Dave likes this.
  16. Dave

    Dave Esoteric Audio Research Specialistâ„¢

    Location:
    B.C.
    I guess you've figured it out then. That's always a good feeling. Have fun and enjoy! :cheers:
     
    pdxway likes this.
  17. Nick Brook

    Nick Brook Forum Resident

    Location:
    Yorkshire, UK.
    Very familiar with Diana Krall's Live In Paris on cd and vinyl , through solid state and tubes. I'm not hearing any metallic tones , though I've not sort of tried to tune in to them . It'll just be my luck that now I'm aware of this possibility , that's all I will hear from now on.:rolleyes: :)
     
    The FRiNgE and pdxway like this.
  18. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    Haha, I like loud. May be at loud level like 85 dB, you might hear something?
     
  19. The FRiNgE

    The FRiNgE Forum Resident

    Yes! I have found it impossible to mix or master through headphones. The ears become fatigued (for some reason, even at lower levels) and the mix or mastering ends up too bright.
     
    pdxway likes this.
  20. The FRiNgE

    The FRiNgE Forum Resident

    I found this selection to be enjoyable to listen to. I could hear what the op described, but not distracting at all. Krall's vocal has beautiful presence and dynamics, and set against a warmer, subdued piano tone.. the piano somewhat enhanced in the 400Hz to 500Hz range.
     
    pdxway likes this.
  21. Nick Brook

    Nick Brook Forum Resident

    Location:
    Yorkshire, UK.
    I'm listening to it now through Marantz amp and Audio Monitor speakers , it sounds lovely . Maybe I need to crank it up.
     
    pdxway likes this.
  22. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    I just listened to the youtube version again with my headset and it is OK, not distracting. But with my Paradigm setup and loud, yeah, kind of distracting.
     
    The FRiNgE likes this.
  23. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    I actually like the song and have listened to it many time. But, I can't really go too loud with my Paradigm setup. As one example, can you hear some low level metallic overtone from 4:55 to 5:20? I think most people would not even notice the slight metallic "noise" unless really paying attention to it.
     
    Last edited: Feb 16, 2019
  24. The FRiNgE

    The FRiNgE Forum Resident

    Just to mention, a soloist with piano accompaniment isn't meant to be played loud, not too loud. Imagine yourself seated at a club venue with the singer before you, and set the volume to a realistic concert level for the style. (how loud is a grand piano in a live setting?) I am estimating my listening level was at about 80dB, during the louder passages. (she sounds wonderful there) As we turn the volume up, our hearing response increases in the treble, according to the Fletcher Munson loudness curve.
     
    pdxway likes this.
  25. pdxway

    pdxway Forum Resident Thread Starter

    Location:
    Oregon, USA
    I see! That is why I can't listen too loud to the song with my Paradigm setup.

    I actually want to hear and feel the piano like I am sitting next to her. : )

    I have acoustic Piano for my kids. Piano can reach 80db in my living room easily. I was trying to go louder than that with the song to "feel" the piano. :D
     
    The FRiNgE likes this.
Thread Status:
Not open for further replies.

Share This Page

molar-endocrine