The Trebor piece above is quality patching in addition to mastering.
There is a test ACX applies after all that. I called it Human Quality Control. An actual person listens to the show. This is where people who can’t read or have vocal defects die. There aren’t particularly good automated tests for that.
There is an even more depressing note. If you have multiple problems, it’s not unusual for ACX to only complain about the worst one. It’s not their job to troubleshoot your studio—although they do post comments if they bounce you.
It’s about 2/3 down chapter two of this message thread. He applied some voice quality corrections and noise reduction. According to specifications, you don’t need noise reduction, but that’s a production call, not a technical one.
For example, “baby screaming on a jet” doesn’t have to be loud to be annoying. Also see: dentist drill. There is an irritating USB microphone error sound that has the same effect. It passes technical specifications, but not Human Quality Control.
This kind of problem also causes no end of consternation because it seems to conflict with ACX Check. You have to pass everything to publish and if I haven’t stressed this enough, you’re competing with professional sound studios and broadcast, not other kitchen recordings.
Let us know how it goes. If you get bounced, please post the exact message.
I don’t own one. I borrowed one (and the studio) for this shoot. I told the owner if he didn’t keep and eye on it, I was going to take it home and he would never see it again. He banged out commercial CD music with this thing. “Earth at Night” (US) is the band. I don’t know if they’re still performing.
First, thanks for your help – I got the job! Next, I upgraded to 2.3.3 Audacity. I first tested a sample using just the steps from the Audacity mastering tools. Failed RMS. Then I ran another using the strategy you gave me for amplification, but then I failed noise (but passed RMS). Nothing has changed with my fort set-up.
I think that’s a mistake. Once you start a job, Do Not Make Any Changes until you finish. Now you’re back at square one getting the tools to work right—and they’re different tools. I hope you didn’t tell them a completion date.
Failed RMS.
That’s one example. If you used the current published mastering tools, it’s almost impossible to fail RMS. The Mastering Suite that runs in 2.3.3 is very different from the tool collection I posted for you to use on the earlier Audacity.
There is a free***** expander plugin called couture which can squish down the noise so you pass ACX …
It also reduces your slight room-reverb, (noticeable on “accept that”).
The couture plugin is available for PC & MacOS,
I use it in Audacity on a PC, (I have no idea how to install plug-ins on a mac).
[ ***** No need to buy the full version, the free version is sufficient for this task ].
I am doing each step, in order, separately. No chain.
This job is not for an audiobook (it’s narration for an Ed company). I decided I needed to learn more before I did an audiobook. The Ed company does not require ACx specifically. Ultimately, however, I want to get to audiobooks.
Good call. The ACX technical specifications are a mirror image of broadcast. If you can pass that, you should be able to get acceptance anywhere.
I am doing each step, in order, separately. No chain.
Something Is Wrong. It’s all but impossible to miss ACX with the correction suite. You may not always pass noise, but the process was split right there for a reason. Noise can be a career move.
Looking…
I know this is such old news by now, but can you post another raw test? Making assumptions when troubleshooting is deadly.
192 - Constant is the minimum allowable MP3 quality for ACX. If you’re presenting in mono (one blue wave) the channel mode doesn’t make any difference.
There is a split here. Some will insist that you select Force to Mono. That can get you into trouble. If you have to deliver to a client in stereo, you have to remember to de-select Force To mono or the client will not get the stereo show they paid for.
ACX recommends strongly that you post in mono. So your decisions in post production editing (Stereo or Mono) should determine the delivery, not an unusual MP3 export setting. I should leave it in Joint.
When I shot voice tracks at work, I always delivered in stereo (two blue waves) even though there was only one person speaking. I knew the production was being edited in stereo and my stereo voice track would drop right into their timelines with no adaptation, conversion or surprises.
The help and instructions are always going to be out of step unless we’re in the same Audacity version. 2.3.3 has easy, convenient MP3 export, and you can use the published, three-step ACX Mastering tools.
I can’t use the processed track. AudacityTest.wav sounds funny because I think it’s been too heavily processed. I need the raw reading. Announce it, cut it down to length if needed, export it as WAV and post it. Do Not apply any corrections, filters or effects.
