Like this only we assume you’ll be a lot better at it than I am.
And yes, it passes ACX Check nicely and I didn’t go to a lot of craziness to get it there.
Eventually I need to do one of these that doesn’t sound like I’m learning to read for the first time.
I have one from a restaurant experience buried here somewhere. “A couple sat down beside me at the counter…”
Koz
Here it is. This one’s much more graceful. It’s an MP3 because it’s too long for a WAV. It goes off the rails at about a minute.
That’s more the story-teller flow you have to generate.
Koz
Thanks for those two samples. And advice taken on the no-music sweetening, we’ll kill that idea.
Here is a more performance oriented sample. The only technical thing we changed was to raise the PAV input level to 135%, or approx. +7.4dB. Other than that, the file is raw and unaltered. Running ACX Check shows that Peak and Noise Floor pass but RMS still fails. How can we fix that? Would this sample be able to be finished with Audiobook Mastering effects?
Thanks again for all your input and expertise, koz.
Excellent. That’s what I was looking for. That’s a terrific presenting voice.
I need to spend some quality time with good headphones and see what turns up.
I’ll publish what I did to it and present examples, but you need to be able to listen to it with high quality speakers or headphones. Did we ever get that far in the discussion?
Koz
Aaaaaand. We’re done. Starting with your post, I applied normal Audiobook Mastering and then added noise reduction at 866 (slightly stiffer than 666). This is expected because the SM57 has “gentle” volume and needs the X2U to work a little harder. Nobody but us will know.
The whole correction suite including Noise Reduction took less than a minute.
This is the writeup on Audiobook Mastering. If your eyes glaze over at the narration, you can skip to the actual application of the tools at Process. If you get lost, it’s back to the top and read the whole thing.
https://wiki.audacityteam.org/wiki/Audiobook_Mastering
This is the writeup on Noise Reduction.
https://manual.audacityteam.org/man/noise_reduction.html
If that’s the voice on the presentation, I predict the thing will sail straight through.
You do need to pay attention to the ACX requirements. They need a special MP3 format for submission (not WAV) and they need Room Tone before and after each chapter.
https://www.acx.com/help/acx-audio-submission-requirements/201456300
Export a WAV of each raw reading (mistakes and all) and keep that in a safe place. Edit a copy of that for submission.
Write back if you get stuck.
Koz
If you’re still eager to mess with it more, you could apply the DeEsser. This helps tame the harsh SS sounds at the beginnings and endings of words.
This is a copy of that clip with gentle DeEssing applied after Mastering and Noise Reduction.
These are the DeEsser settings I used on that clip. Please note all these tools depend on each other. Don’t leave any out, add any, or rearrange the process.
This is where to get the DeEsser I used.
https://forum.audacityteam.org/t/updated-de-clicker-and-new-de-esser-for-speech/34283/1
Koz
And yes, this is where you listen to both clips on your excellent headphones or speakers to determine if the DeEsser is needed. It’s a theatrical quality call. Both clips pass.
Do you like the overall presentation? I do. If that’s the voice reading, this should go through no problem.
It’s possible to submit a short test to ACX before you start reading for weeks. Highly recommended.
https://forum.audacityteam.org/t/how-to-send-a-test-to-acx/49588/1
Koz
Yeah. I like the DeEssed version. Add that to the protocol.
I think you can still get through the whole correction suite in jig time. You get used to applying the tools and where everything is. Most of the tool settings stick, so you can jump from one tool to the other. The tools will take longer if you apply them to a chapter.
There is one actual bug—program error. Pay attention and note whether the Low Rolloff for Speech setting on Effect > Equalization sticks time after time or not. It should.
Koz
I’m trying to think if I missed anything.
You should cut the show before mastering. Take out all the gaps and mistakes and that time you had to take the dog out. Then, when the rhythm and flow is exactly right, apply Mastering, Noise Reduction and DeEsser.
It’s not unusual for someone to record 10 or 20 seconds of pure, clean Room Tone. Copy and paste pieces of that to the beginning and endings of each chapter according to their instructions. You can also use that if the presenter coughs in the middle of a piece but it sounds funny if you just cut out the cough and it will sound even funnier if you try to Generate > Silence in there. Paste in a snippet of Room Tone.
Koz
No, it’s not simple. You are the recording engineer.
Koz
I think I could listen to that person read the phone book.
Koz
I updated my web page to condense the mastering instructions.
Scroll down to the AudioBook Mastering Quick Notes box
http://www.kozco.com/tech/audacity/audacity.html
Koz
Thanks koz, it’s good to hear that we are finally getting there. She says thanks for the encouragement, by the way. I’ve always told her she had a quality voice, now she may believe it.
I do like your mods a lot, and I also feel that the DeEsser does add to the mix. It makes a difference and gives a cleaner track. Thanks for all those links, I’d seen some but didn’t know where to get the DeEsser. There is still a lot to experiment with mastering and noise reduction in the future, but your settings have gotten us a much cleaner take that passes ACX comfortably – thanks so much for all the guidance.
So I just went to test the effects series that you used and applied them in sequence as you recommended to my copy of that raw file. I got the exact results you did on the ACX Check except the Noise Floor was 0.1dB lower. All other readings were just the same. So that sequence replicates for me here too, and as you said, it really does not take much time at all. I’m sure other files and other vocal requirements will call for different settings, and that’s when I’ll learn just what effects really do on their own. But the synergy of the ones you chose really makes a difference. Hey, there’s hope for this SM57 yet, I was beginning to wonder….
I’ve got some old budget Sennheiser open-back phones around here somewhere but I do need to get a better pair, thanks for reminding me, I can see how that’s a weak link right now. And thanks also for the workflow plan: backup, edit, then filter. Makes perfect sense. Room tone can be saved as a separate reference file I guess, one for each project, right? Just grab a piece of it when it’s needed? (Now it hits home why all the sound engineers on film sets want those 30sec samples of quiet at the end of a scene’s work…)
Koz, thank you again for all your help, patience, and detailed attention – it has really gotten us heading down the road with a sense that we’re moving in the right direction at last.
I’m sure other files and other vocal requirements will call for different settings,
Maybe not so different. Mastering Suite is designed to get you past ACX technical requirements. The tools are integrated, automatic and if they’re not needed, they don’t do anything. You can change the suite when ACX changes the requirements.
The DeEsser is automatic at those settings and good correction depends on the Mastering Suite coming first
The only variable is noise, and in this specific case, the limit is the SM57/X2U hiss. I don’t hear any room trash at all.
And this brings us back to finding a process that works and don’t let go. The longest post on the forum is Ian. All he wanted to do is read audiobooks from his apartment in Hollywood. The message thread is over a year long.
“Still too much traffic noise from trucks on La Brea. Maybe you can record later at night.”
But he succeeded.
Then there’s the rule that you not change anything in the middle of a book. ACX wants everything to match.
Little Ray Of Sunshine time. You never found your hum, did you? That’s going to be a little time bomb until you figure out what happened.
Koz
I got the exact results you did on the ACX Check except the Noise Floor was 0.1dB lower.
Noise Reduction success depends on where you got the profile from and how much. One example is getting a little bit of voice in the profile by accident. Noise Reduction will try to remove your voice.
Koz
Since you have a stable platform now, there’s a theatrical note. Many new readers get to the end of their first book, listen to the first few chapters, make a face and read them over again. The beginning of the book was read by a rank amateur. The book was finished by a seasoned professional.
Although that may not happen this time. That voice has a natural tone to it and should carry through a book. Does she do answering machines/voice mail?
“Hi. You’ve reached Koz’s voice mail. If you want to leave a message…”
There’s an acting thing nobody thinks about. Yes, the actor can emote, sympathize with the premise, meld with the character and live the environment. Now do it exactly the same way fifteen times while they get the lighting right. Then come back from lunch and do it again.
That’s why some actors don’t break character and eat their Craft Services garbonzo salad as Jack Richter, Private Eye. I marvel at people who can switch back and forth.
Koz
We made another test run of a 1-minute clip, and applied your full 3-step Audiobook Mastering, 8-6-6 NR, and DeEsser to it and have submitted it to ACX through their survey form in the thread you linked to.
It’s possible to submit a short test to ACX before you start reading for weeks. Highly recommended.
viewtopic.php?p=349993#p349993
We are waiting to hear from them and I will report back when they do.
We made another test run of a 1-minute clip
I hope a little longer.
I hope you submitted to their MP3 standards.
https://www.acx.com/help/acx-audio-submission-requirements/201456300
I can’t prove this, but I wouldn’t be surprised if they do this to find out if you can follow instructions or not.
Yes, do post back. I’m on the edge of my seat.
Koz
We finally heard back from one of the ACX analysts, and – we passed! I was concerned that a one minute clip might have been too short, but the reviewer said that the file met all requirements. He also sent along a zipped attachment of good & bad samples, as well as ideal examples from each stage in the production process. They are WAV files to import into Audacity to see on the meters and hear what they are after. It’s a set of clips that will be a useful learning tool.
So I guess we are finally over this first hurdle – thank you koz, for all your support and input through this long thread; you’ve been a great help.
we passed!
I’ll put another notch on my mouse.
He also sent along a zipped attachment of good & bad samples
Apparently, they’ve fallen in love with those samples and they get sent out whether you passed or not.
Koz
