I have two books in production, they are my first projects.
I don’t have much issue getting my tracks to pass ACX check. My room tone meters in at -57db, so noise floor is OK. Sometimes I need to use the W1 limiter to get RMS in spec, but I usually get that after a couple of attempts, and peak db is easy.
Where my question comes in is the ACX requirement to submit mastered finals. I can follow the You tube videos on mastering in Audacity, but I don’t know if that will satisfy the requirement. In the distant past I owned part of a state of the art studio and we always had Sterling Sound do the mastering. Doing a book for royalty share doesn’t provide funding for mastering 4 or 8 hours of audio. So, I thought I’d try LANDR. Thus far my MP3 and WAV files error out when I try to upload them, LANDR support says they are on the problem, but not having a phone number I can’t talk to a real person. Hoping one of you have experience with this.
ACX Noise Floor is -60 and it can be a serious juggling act to get all three to hit at the same time. It can be really painful if your background noise isn’t plain pink noise FFFFFFFFFF.
Are you using ACX-Check, the Audacity audiobook verification tool?
LANDR support says they are on the problem, but not having a phone number I can’t talk to a real person.
Who? Them or you? Or both?
If you’ve been following the discussions, you know that anyone can brutally force a submission into ACX compliance, and it’s even possible to semi-automate it. What you can’t do is “clean up” a ratty submission without it sounding like a bad cellphone call (incidentally, why cellphone calls sound like they do). ACX has a failure they call “overprocessing.” That happens right after their automated robot looks at the simple quality numbers and passes the submission on to a human. Then, the jig is up.
If LANDR has any kind of success, it will immediately sink under the waves of billions of people trying to submit at once. We’ve tried this. It’s not easy. I just read a business magazine story about “fully automated and algorithmic personal digital assistants” which are actually a bunch of badly overworked humans in a barn somewhere working their butt off over killer hours. Eventually, they’ll get the actual algorithm working…
I see what they’re doing. They are a paid music mastering service. Not once in the documents I saw do they mention Noise Reduction, ACX, AudioBooks or Spoke Word recording.
I DO NOT work for LANDR, I AM NOT a shill for them. Your implication is wrong and offensive.
I told you in the first post, I;m not having any problem with noise floor, passes ACX check every time, RMS can be an issue, but one or two attempts with the W1 limiter takes care of that. Additionally, my tracks sound good, I have Klipsch Heresy speakers and a McIntosh Amp, good headphones, a top notch car stero that I use to cross check, also a set of AR powered partner speakers for cross checking, plus the lousy computer speakers. I do thorough checking. And I have excellent ears to boot.
BTW, I used to be one of the owners of Forum Studio which produced 2 #1 in the world records and I apprenticed under Bill Halverson CSN producer and engineer.
However, I am not an Audacity expert and just a beginning ACX producer. For those reasons I solicit your advice, you are the Audacity expert.
Sounds like you’ve got it licked. Have you tried sending a sample submission to ACX for them to review? I think that’s the next step. They will have full instructions on their site - read them carefully and follow the instructions to the letter.
The question is regarding mastering, in my world mastering is the last step after mixing, EQ, effects, etc. If you listen to a final mix and then compare it to the mastered version, you should hear a difference, often a rather large difference, more dynamic, greater presence etc. In years gone by I always had Sterling Sound master my final mix, unfortunately I can’t afford the excessive cost of that service as I am a retired guy on a fixed income, trying to establish myself in auidio books. Perhaps my understanding of what ACX means by mastering is off point, BUT
This is what the ACX blog says about mastering
Why do I need to master my audiobook productions?
A: Mastering is the the final step of post-production and the glue that brings the entire audiobook together. All chapters/sections are brought up to matching levels, which provides a smooth listening experience. Additionally, removing unwanted high and low frequencies can help reduce any hum or hiss that may be in a recording.
If it is simply passing the first level check, then yes, I’m good to go (yes KOZ it also sounds good).
ALEX’S MASTERING BREAKDOWN
Assess all audio files to ensure no peaks or clipping exist in the audio.
A good recording and careful editing are both necessary to achieve this.
Group all similar files together during the assessment so they can be processed at the same time.
Apply your “Mastering Chain” by using the following processes, in order:
Remove all unnecessary low and high frequencies by applying EQ to clean up the sound of your recordings and provide more headroom in order to boost your files levels effectively. This is a great way to minimize hum and hiss in an otherwise good recording!
Bring all files up to the proper dynamic levels as specified by the ACX Audio Submission Requirements page by using normalization, compression and/or limiting, and, if necessary, a final volume adjustment.
Listen to your audio after mastering to ensure the operation did not over-process or under-process the recordings. If the resultant audio is at one consistent volume with no change in dynamic level, you’ve likely over- If your audio has sudden spikes and drop offs (indicating it is too dynamic), you’ve under-processed.
And warranted given at first pass, the product doesn’t appear to have anything to do with AudioBook Mastering. We have a sensitive trigger for people advocating paid services on the forum, and that is a paid service. That and you haven’t succeeded yet. Note that all our recommendations and suggestions are from forum elves who actually own and happily use the product or service.
I don’t completely agree with McElroy’s final video where he processes his test sample so the waveforms all but fill the timeline. Everything up to that seems to be fine; the soundbooth walk-through, etc.
You’re in a fuzzy zone. I’m with Steve. I would forget about advanced mastering and submit. Given you are competent in the recording and I don’t see any reason to doubt you. The performance model is being fascinated by someone telling you a story over cups of hot tea in a quiet kitchen. My opinion is the suggestions for advanced mastering are overkill, particularly because several forum posters are ACX published authors without it.
I have two books in production, they are my first projects.
That may be awkward.
There is one condition you may run into. Many people read through their first book, go back to the beginning and read it all again. The experience, smoothness, competence, and quality go way up through the course of your first work as you get used to reading. More than one poster claimed to have gone back to the beginning to listen to a full run-through and were shocked how amateurish it sounded.
That because you were an amateur.
The only people who avoid this trap are existing actors, presenters and performers.
I’ve gave LANDR a test drive a few months ago. While it could be useful for bands / musicians that want some quick and inexpensive polish to a clean and well balanced recording without needing to know about the intricacies of compression, equalization and so on, I wouldn’t personally describe it as “mastering” and I don’t think it would be beneficial for audiobook production. They do offer a free trial, though you only get “low res MP3” for free, which would not be good enough for ACX submission.
So that takes care of some of the mastering steps. You are unlikely to wander in volume or change character or voice quality (unless the story calls for it).
If you feel like it, you can apply their frequency filtering steps. Steve wrote a vocal “rumble” filter that takes out most non-valuable low frequency sound, but that’s overkill for this project.
Effect > High Pass Filter: 6dB, 30hz
Listen for your voice “tightening-up” and losing it’s broadcast rumble. Move the 30Hz up or down. Nobody can hear 20Hz. 40Hz and 50Hz are good trials. Steve’s stiff rumble filter starts at 100Hz and that’s audible on some voices.
Effect > Low Pass Filter: 6dB, 7500Hz
Listen for the crisp sparkle and clarity to come and go. The performance shouldn’t be too crisp. Move 7500 up and down. AM radio only goes to 5000, FM poops out at 15,000. Human Hearing can go to 20,000. I knew one human who could hear to 19,500, but he couldn’t go any higher. We are assured some young girls can do that.
Do you have a monitor/speaker system or good headphones to judge the work? My good rule of thumb is the inability to hold your sound system in one hand. I don’t care who made it, that’s not going to work.
You will find that adding these two filter steps will help with your ACX noise test value. The noise measurement intentionally doesn’t care whether noise is audible or not. It measures everything. If your system accidentally creates rumble down below 20Hz (earthquake, thunder, big lorries/trucks, strong wind), you’ll never hear it, but it can cause you to fail ACX Noise “for some reason.”
I gave LANDR a try and was actually fairly satisfied.
One problem that I had: My submitted Flac files were all to hot after processing and converting to mp3 (peak about 0.46).
I’ve contacted the support and they were very helpful. They processed the file with the beta engine and let me download the result as Flac file (a service you have normally to pay for).
The peak of submitted files has to be relatively low (-6 dB or less).
Have you tried the desktop App (to simplify uploading)?
You can record the web-preview with Audacity’s Wasapi host and speaker loop back and check the result within Audacity (peak,RMS etc.).
LANDR has an automatic genre recognition but the result is not shown and so you cannot be sure that the narration file isn’t treated as Rap…
I use also AAMS, a desktop application for automated mastering.
It has far more options that can be tweaked. It works with reference files (*.aam) and you can make them yourself by just analyzing a file you like. The result will be shown as *.doc file, if you wish so and you can take the found values (compression, limiting, multiband and so on) for processing in Audacity.
However, AAMS produces all the file types you’ll probably need–Mp3, wav-16, wav-32-bit float and so on.
The main disadvantage is that the program is very slow (audio books are long…), that’s why I would just use it for analysis and not for processing.
Having said all this, I would still do the mastering within Audacity and let ACX-Check and your ears do the judgement…