You need to get rid of the buzz.
If you post a short sample, we can take a look to see if your recordings are recoverable. (see: https://forum.audacityteam.org/t/how-to-post-an-audio-sample/29851/1) Please post a short sample in WAV format that has not yet been encoded to MP3.
Thanks so much for getting back to me so quickly! And for your offer to help
Problem is I actually used Reaper and so I only have finished files in MP3 format???
I’ve attched a WAV file as well as an MP3 - maybe you can work with one of them??
I don’t hear (or see) any “buzz” in that sample, though I can hear the Noise Reduction effect
(particularly noticeable on the “Cursed Country …plink…”)
It is below -60 dB, which is what ACX specify, but it is “suspiciously” low. Nobody gets that low without processing the noise out.
Why did you use so much noise reduction? Were you trying to get rid of a buzz? If so, perhaps some of that buzz is still audible in some of the “quieter but not silent” parts of your recording.
I have written back to ACX to ask if they can tell me what I have to do
You may not get a reply. Their job is to insure high quality work for their platform, not do detailed analysis and servicing. Most of the world is trying to read for audiobooks and their quality control offices are overloaded.
Problem is I actually used Reaper and so I only have finished files in MP3 format???
Do you have the unfinished files in WAV? Reading, processing and submitting one single sound file per chapter is a very serious new user error. Every time you make, change or correct an MP3, the sound quality gets a little worse and you can’t stop it.
re-record a 15 hour book
ACX has provision for you to submit a brief sound test rather than arrive one day with a complete book.
This isn’t the whole story. They don’t actually publish specs for this. The spells are actually those for publishing a book summery. Between one and five minutes of typical work. I need to drop for a while.
It’s notes one and three of these publication instructions.
OK, Steve, I don’t know if you’re familiar with Reaper but I have always simply used their ReaFir (FFT EQ+Dynamics Processor) plugin on the first few seconds of my recording.
In over 100 audiobooks I’ve only ever had an occasional problem with volume levels - usually when doing multiple characters!
This whole “noise” thing has only started since I changed to a SHURE 37B mic with Cloudlifter and I have the feeling that I’m still too high on the preamp gain setting.
I will tweak it some more for the next project (& get ACX to approve a sample beforehand LOL)
But my immediate concern is being able to resolve this situation WITHOUT having to re-record the whole **** book!!!
Problem is I actually used Reaper and so I only have finished files in MP3 format???
Do you have the unfinished files in WAV? Reading, processing and submitting one single sound file per chapter is a very serious new user error. Every time you make, change or correct an MP3, the sound quality gets a little worse and you can’t stop it.
re-record a 15 hour book
Hi Koz,
Thanks for your input - it was actually one of your old posts that prompted me to try here on Audacity’s forum!!
To answer your comment above:
I fully understand (& appreciate) how quality is affected every time one plays around with an MP3 file…BUT…my overall performance has improved so much over time since I recorded my first audiobook that, (without sounding presumptious - I hope!) I’ve always managed to record & edit on Reaper and then render as an MP3 file ready to upload to ACX.
I have only just started to use Audacity’s “ACX Check” plugin - it didn’t exist when I first started with Audacity!!!
Are you, in fact, suggesting I should render as a WAV first then run the checks & then render again as MP3 if all OK?
Problem is, seems to me, that ACX only gets to do their “quality control” when MP3 files have been uploaded and…obviously from my current situation, Audacity’s ACX Check doesn’t seem to be “strict” enough to get files past ACX’s Quality Control???
You can render to WAV or MP3, but don’t apply the noise reduction effects. That should be “raw” enough. Or, you can render the truly-raw original recording with absolutely no adjustments or processing/effects.
There is too much processing. There is the “Blackness Of Space” silence between each word and in some cases I can hear odd tails on the words.
We know they test for background noise and then start getting critical for sound quality if they find “odd” readings—such as noise suspiciously quiet.
This is a tiny segment of “silence” at 11-1/2 seconds. I artificially boosted it to illustrate the point.
That’s supposed to be clean rain-in-the-trees ffffffff sound, not talking-into-a-milk-jug echoes.
This odd effect is actually audible at normal volume after the last word in the test clip. You have a tongue click with an odd tail after it at about four seconds.
If you clearly pass technical considerations (ACX Check, etc) then the only thing left is to fail Human Quality Control. In my opinion, that’s what happened.
Your production pathway normally works perfectly (as you found) until Something Goes Wrong. Then you may need to fall back on the safety copies of the raw readings and WAV versions of the submitted export.
What does Reaper have to say about .rpp files? There is an unusual Apple sound file format whose quality exceeds WAV and AIFF. It’s possible the .rpp files are just fine. Look for words such as “uncompressed” and “lossless.”
And you can get a hint from the file sizes. A proper 20 second mono uncompressed WAV voice file comes in at about 1.8MB to 1.9MB.
We know everybody with a pulse is trying to read for audiobooks. I can well imagine Human Quality Control after 8 hours of listening to badly behaved microphones in noisy, echo-filled rooms hyperventilating and losing it after a while. Since everybody always fails noise, all the failures start to run together.
“Buzzy sound. Yeah. Yeah. That’s the ticket. Buzzy Sound. Wait. Which clip is this? Doesn’t matter. Buzzy Sound!!”