ACX-Check

[Synapse001]

What was the delay submission to finish?

The robot seems to check each file during upload

Koz

Just the normal Uploading 0 to 100%, then Processing Audio for a couple of seconds. There was no additional discernable delay.

[kozikowski]

then Processing Audio for a couple of seconds.

No, not that delay. The one between your first pass at submission with a successful product and they sent you a check. Ten days?

Koz

[Synapse001]

then Processing Audio for a couple of seconds.

No, not that delay. The one between your first pass at submission with a successful product and they sent you a check. Ten days?

Koz

Oh.... The author and I were in contact a couple of times the day I completed the book. It was RS+, so before the book would go to review, the website had a button asking me if the author had paid me yet, and let me know the amount due based on the running time it calculated.
The author asked for a PayPal address, I sent it and he paid me. He actually paid me more than was recommended by the PFH agreement, and I emailed him to let him know and he said it was intentional and he wished he could pay me more. It was nice.

I clicked the [Yep, He Paid Me] button and it went into this Awaiting Audio Review status and said it could take up to 30 days.

I did that book because foolishly I choose an 18 hour RS book as my first project and it is beating the crap out of me.

Hope that helps
-Synapse001

[Chadillac]

Win 2010, audacity 2.3.3

my first project and it is beating the crap out of me.

I am just starting audio book narration. Working on two shorter projects. And also new to Audacity. I am trying to find time to explore this forum, and youtube to try and learn as much as possible.

My primary problem right now that I am struggling to find the answer to is:
After amplifying to get to acceptable RMS levels, I have peaks that I have manually select each one and negative amplify to prevent distortion. This is very time consuming. What would be the best way to do this without having to INDIVIDUALLY select each peak and negative amplify them, without affecting the lower volume portions of my audio?

Thank you!!

[Trebor]

... I have peaks that I have manually select each one and negative amplify to prevent distortion. This is very time consuming. What would be the best way to do this without having to INDIVIDUALLY select each peak and negative amplify them, without affecting the lower volume portions of my audio? ...

Audcaity's (soft) limiter will automatically knock down ALL loud peaks in one fell swoop ...



(don't choose "make-up gain")

[kozikowski]

You may find that just adding Limiter solves all your problems, and you win. We have an Audiobook Mastering Suite of three tools that can turn any announcing style into a passing submission. If you announce well in a quiet, echo-free room, you may get a passing recording with no further effort.

The third number, noise, is really hard.

MasteringSuite.png (41.82 KiB) Viewed 789 times

That's lifted from the more extensive wiki post.

https://wiki.audacityteam.org/wiki/Audiobook_Mastering

Also see ACX Check, a handy way to see how you did without submitting and waiting.



Koz

[Chadillac]

Thank you Trebor... I will try it out.

And thank you Koz. It took a bit of digging and time to find the ACX Check plug-in, but I did the leg work since you did not post the link or where to find the plug-in. No worries because without your advice I would've never found it. Here is the link for others if they want to add the plug-in.

https://wiki.audacityteam.org/wiki/Nyqu ... #ACX_Check

[kozikowski]

Thanks for the post.

you did not post the link or where to find the plug-in.

Actually I did. It's in the Audiobook Mastering instructions.

Koz

[Richken71]

Hi

I applied the ACX Suite to a 60 minute 1st chapter of a book I'm recording and everything passed with the ACX Checker (though I had to run the recording in two halves as the checker said the whole thing was too long) No other processing was applied.

However, I saw something on Youtube which seem to be saying that although the whole file may pass the ACX Check, shorter sections of the same recording may fail. No explanation was given. With this in mind I selected random bits between a few seconds and a few minutes and applied the checker to each. Although most passed on all three counts there were two extracts which failed, one for noise (-58.1) and another for RMS (-17.8). I didn't cover the whole recording in this way so there are probably other failures lurking elsewhere. Listening to these extracts didn't reveal any extra noise beyond those that passed.

If the above are typical of the failures they don't fail by much but I guess enough to be returned by ACX QC? Or can we be confident that if the overall recording is given a pass then it should be ok on the technicals, leaving aside the performance? I realise of course, that a Checker pass does not give a guarantee but wonder if anyone else has come across this issue and what the outcome was.

Apologies if this has been covered before, but I could not find any reference to it in the forum.

Many thanks

Rich
Audacity 2.3.3, Macbook pro OS 10.13.6, Rode NT2 mic, Audient Id14

[kozikowski]

Some notes about that:

We're not a service of Audible or ACX. We're working backwards from their published specifications and have a reasonably good track record of getting people published.

one for noise (-58.1)

It's good to know how the noise measurement works. It needs at least 3/4 second of pure, clean background noise to work right. Not you gasping for breath, shuffling in your seat, rearranging your paper script, or slurping coffee. If you happened to select a portion of script without a chunk of clean noise, you may get an astronomically high noise value which is completely bogus.

another for RMS (-17.8).

RMS is a little magic. Root Mean Square is a way to measure complex waves from anywhere, not just sound. There is an RMS value for the electrical power delivered to your house. It's a close cousin to Average. Someone noticed that RMS measurements for sound happened to work out approximately to loudness. Many other methods of measuring loudness are complicated and hard to use, so this was handy.

Because of these limits, it can be messy to measure really short selections. It's like trying to measure your lawn by sampling one blade of grass. It doesn't work that way.

Audiobook Mastering is not a gift from the angels. If you insist on announcing really quiet for the first few minutes and then really loudly for the next few, On Average you will have a perfect performance. Nobody is going to pass that damage for publication.

There was a recent posting from someone with several errors in the reading. It was possible to rescue the performance, but the corrections ran to eight or nine tools... for each segment... of each chapter... for the whole book. Reading one book that way would be a career move.

Yes, you can add compressors to your work to even out minute-by-minute volume variations, but those screw up noise measurements.

This is...sssSSSS A test...sssSSS of my voice...sssSSS.

Much better to listen to yourself on headphones while you read. Volume variations tend to even themselves out.



Mastering is the fewest number of tools you can use to turn a good performance with a quiet, echo-free room into a successful submission.

If you violate any of those three requirements, you're in damage mitigation and that can be painful.

We also note all that will only get you past the automated robot at the start of ACX evaluation. You still have to read well in a nice voice.

Announce one of these and post it.

https://www.kozco.com/tech/audacity/Tes ... _Clip.html

Koz