Fledgling voice artist seeking counsel.

[MichloIW]

Wow, that must have been some cuppa.

[kozikowski]

All my cuppas are award-winning. I'll go up and look again when it's daytime.
Koz

[Robert J. H.]

Hi Ian,
I've modified my curve a bit to include the LF roll-off for speech.
It sounds quite nice on my system.

Cattle - Foreward (processed).mp3

(973.82 KiB) Downloaded 382 times

What I did:
- amplification to 0 dB
- equalization with this curve:

correction_for_speech.xml

(9.28 KiB) Downloaded 59 times

- amp to 0 dB
- Noise reduction 18 dB, sens 0, smoothing 300, attack 0.01.
- lowered the 10 longest/highest rms values and the 20 longest/highest peak values.
(and amp to -3 dB)

The curve should still be suitable for new recordings if you did not change anything since.

The statistics show:

Code: Select all

---------------------------
Nyquist
---------------------------
Length of selection: 42.760 seconds.
2052476 samples at 48000 Hz.
Analysis of first 42.760 seconds:
(2052476 samples)

Mono Track.
Peak Level: -3.0 dBFS
Peak Positive: -3.0 dBFS
Peak Negative: -3.1 dBFS
DC offset: 0.0 %
RMS: -22.3 dBFS
RMS (A-weighted): -26.4 dBFS

---------------------------
OK   
---------------------------

I did the last step (lowering the peaks) only to get over -23 dB Rms, although the Acx didn't bother about that for the previous submission.
The Rms-lowering affects words that are too loud (from my subjective perception that is).
In the forward for instance "Faish" or however the virus is called. This step is optional, particularly because it isn't so disturbing as the exclamations in the actual chapter narration.

[kozikowski]

Perfect. Installing a new curve is covered near the bottom of this section of the instruction book. Manage Curves.

http://manual.audacityteam.org/o/man/equalization.html

Koz

[MichloIW]

Robert!

What a generous gent you are. Thank you so very much.

Would you recommend that I start over in order to ensure consistency or just apply your tool to what I've done so far (foreword and first five chapters)?

And is there anything I can help you with in return? I've already offered to help Koz with real British tea.

Thanks for the manual link, Koz.

Cheers.

[kozikowski]

British tea.

From all those tea plantations in the Midlands?

Koz

[Robert J. H.]

British tea.

From all those tea plantations in the Midlands?

Koz

..."The riots of 2014, known as the LA Tea Party"...

Koz, have you already managed to spread your little finger away while you're nipping tea?

[Robert J. H.]

Robert!

What a generous gent you are. Thank you so very much.

Would you recommend that I start over in order to ensure consistency or just apply your tool to what I've done so far (foreword and first five chapters)?

It depends what you mean with "done so far".
The equalization can in principle be applied at any stage.
If you apply it on a track that is already treated with Steve's roll-off curve, you can set all below and including 100 Hz to 0.
I've just noticed that the forword has 48000 Hz, that was actually not intended.
There's no need to re-record all chapters (unless you have an artistic reason).

What we now need to do is to ensure that the average Rms values are about the same for all chapters.
We have so far:
- equalization
- noise reduction

You can then apply Paul's de-clicker if necessary (it crashes on my system for longer audio; I haven't the newest version though).
After that, it is time to measure the Rms and to decide what we do with each chapter.
Interestingly enough, my favourite audio books are all at about -28 dB, not even near to the range -18 to -23 dB, more so as the peak is at -1 dB.
And the noise floor is at -100 dB. Since this are all professional german productions, I suspect that the philosophy is quite different from Acx.

And is there anything I can help you with in return? I've already offered to help Koz with real British tea.

A quite dangerous offer, I may fall back on it.
I might someday have the need for an English voice, either for a Audacity tutorial or another English text.
Here's a poem of mine that has been read by a German speaking actor:
https://www.dropbox.com/s/wkq1vdmvxs7s1bi/ballade2.WAV
Incidently, it fullfills the Acx requirements (after normalization to -3 dB) with Rms at -20 dB and noise floor at -70 dB.

[MichloIW]

Koz,

how did you know the best tea comes from the Midlands fields? That was supposed to be secret!

Robert,

I believe I still have the raw projects saved so that is good to know, it will save me some time if I don't have to re-do them.

I've not yet had a crash from the de-clicker so yes, I'd recommend you get the latest version. Depending on my warm-up and hydration, however, I'm needing it less.

The offer is there, dangerous or not. I'm listening to your poem right now but my German is far too rusty. I'm liking his flow, however.

Cheers.

[kozikowski]

I believe I still have the raw projects saved

Excellent idea. I do them as WAV files, but the principle is the same. I have original captures from professional work I did years ago. I've had too many producers call me later and want a separate copy of the presentation session they paid for.

"Koz, did you keep a copy of the John-Mark interview?"

Of course.

Koz