I upgraded over the weekend from a Blue Yeti w/ popfilter and homemade soundproofing box to a Rode NT1 microphone with the Rode pop filter, as well as a desktop Aokeo isolation shield and the Focusrite Scarlett 2i2 audio interface. I’m seeking feedback on the sound mainly because I’m been fussing over this all day and am unsure if I’m going deaf to any errors. A few things of note:
I tried to use the Ableton Live Lite software that came with the Scarlett, but struggled and went back to Audacity
The microphone records REALLY quiet. In Audacity, recording volume is set to 1.00 and on the Scarlett the gain knob is probably 75% turned. I’m almost on top of the microphone when recording.
I have a swatch of moving blanket under the isolation shield and microphone, but am otherwise recording at my desktop
I do have foam panels to my side and on the wall behind the desktop, as well as a heavy quilt behind me for sound control
When listening over speakers, it sounds fine. With headphones, I hear a bit of…emptiness? Hard to explain it, and even harder to figure out a fix! I know it’s most likely something to do with my room setup, which I understand. I do still have my homemade soundproofing box (just a cardboard box lined with moving blankets), but I worry about using that and being too quiet. I’m teetering on the line of being “too quiet” for ACX now, even when not specifically removing noise.
It doesn’t sound terrible to me. I’m only listening on my laptop speakers but I think it’s fine.
I think it just sounds “dry” just because there’s no music or anything behind it.
Are you aware of the [u]Recommended ACX Mastering Procedure[/u]? It will nail the peak & RMS levels every time, and then you mostly just have to worry about noise (and a good-sounding recording of a good “performance” )
The levels are OK for the raw recording. The main thing you want to avoid is hitting 0dB because you’ll clip (distort) if you “try” to go over 0dB. It’s normal to leave headroom and then boost after recording.
Adjusting the knob on the interface has almost no effect on sound quality (as long as you avoid clipping) because the signal and noise are changed together. A stronger-louder signal-level into the mic (speaking with a strong-confident voice close to the mic) does help because you are increasing the acoustic signal without increasing the acoustic or electrical noise.
Your noise level is also OK (after filtering). When you boost the level to meet the ACX requirements the noise gets boosted and it’s around -50dB which is too much for ACX. But it’s low-frequency noise, so it’s OK after running the Low roll-off for speech effect (now it’s around -70dB). That’s very good for a “home studio”.
I think ACX will only “ding you” if your noise is around -90dB from excess processing.
I currently do use the “Recommended ACX Mastering Procedure” with a few additions:
Bass and Treble = Bass -6, Gain 1, Link Sliders 0, Treble 6
Notch Filter = frequency 120, q 10
Filter Curve = low roll-off for speech
Loudness Normalization = normalize RMS to -20dB
Limiter = Soft Limit, 0.00, 0.00, -3.50dB, 10.00, No
I don’t remember adding in the first two, so I’m unsure if they are actually needed still. I haven’t adjusted this since getting the nice, shiny new microphone haha. Glad to know everything sounds good on laptop speakers, though! Like I said, overall I think it sounds okay apart from a slight empty/dry sound over headphones. And my concern over that sound comes from figuring people will be listening with headphones and pick up on that
Running the macro I listed still turns out audio that passes the ACX Check, which is a good feeling. It’s just that darn “hollow” sound, and I can’t decide if I’m being too anal about it or not.
Sounds OK. What did it used to sound like with the other microphone and system? Do you have a sample of that?
Post a Catskill Farms where you don’t do anything to it. Announce, Press Stop, select the performance if needed to stay under 10 seconds, export as WAV and post it. If the performance is Mono (one blue wave) you can go all the way out to 20 seconds—so you don’t have to cut it off.
You’re applying three different filter/effects that can remove bass notes from your voice. That can make you sound gutless, squeaky, and wispy.
This is Molly Wood from a Marketplace broadcast. She has a killer presentation voice.
I have added a recording I did with the old microphone (Blue Yeti in sound dampening box), but it’s not the Catskill Farms script. It has had the macro run on it that I mentioned. I, unfortunately, do not have a sample of the old setup without editing.
The Catskill Farm file I linked in my original post is without any editing done to it. It is Mono and I trimmed room tone from the beginning to be under 10sec, but that was it editing wise. Nothing has been applied to it.
I really like the Molly Wood sample you linked and I love how “present” her voice sounds. I’ve had audio approved through ACX and have a few books out there I have narrated recently, so thankfully I don’t think my presentation is an issue. Hopefully. I do have a bad habit of editing out breaths, and it was refreshing to hear a few in hers.
She used to be a regular, then she was here and there and finally, I haven’t heard her in a long time.
Marketplace comes from Los Angeles and last I checked she’s in Oakland so that may have something to do with it. Running a live, current-events show isn’t fun when everybody is in the same room.
It is Mono and I trimmed room tone from the beginning to be under 10sec, but that was it editing wise. Nothing has been applied to it.
How did you get the mono? The 2i2 is a natural stereo interface system and many of the methods of producing mono also reduce the volume by half (6dB). So that could be where some of your voice volume is going.
the Scarlett the gain knob is probably 75% turned.
The 2I2 has built-in volume indicators. You don’t have to guess at the percentage. Assume your microphone is plugged into the left channel, turn the right channel knobs all the way down. De-select INST and AIR. Turn the Left knob all the way up. De-select INST and AIR. Speak louder and louder and get closer until the 2i2 knob starts turning green. (the second knob in the illustration). If you get too loud, the knob will turn red. Never blow into a microphone.
Can you get the green knob like, ever? It’s not unusual for the volume knobs to work all or almost all the way up on home systems.
There are ways to boost your speaking volume without having overload distortion or picking up lip smacks and tongue ticks.
Got that running, so far?
More to come.
Did you record your previous books with the Yeti? Do you still have it and all the studio setups? There’s a New User problem where somebody gets a new microphone and then immediately throws all the old stuff out. I wouldn’t do that.
I does sound like a raw file, with the exception of echo/reverb-reduction, which you can hear kicking-in after “Catskill”.
i.e. “Catskill” has stronger reverb after it than “farms”, (echo/reverb-reduction takes a second to work out what to do).
If you do turn off echo/reverb-reduction it’s going to sound more roomy.
In mild cases a De-reverb plugin can reduce the room reverb, but they are expensive (>$99), and generate digital artefacts.
That is something else that can get in the way. Do you use Skype, Zoom, Meetings, Group Games, or other chat programs? Their voice processing can get “stuck” on and affect your voice even if the program is napping or off. Do a clean Windows shutdown.
Shift+Shutdown > OK > Wait > Start. Don’t let anything automatically start. Pay attention if your voice suddenly sounds different and that’s the only thing you changed.
In regards to mono, it is how I recorded everything prior to the new microphone, so it never occurred to me to change it. Audacity has the drop-down menu for mono or stereo and somewhere in the beginning of me learning to record, I set it to mono and never considered it again. I didn’t realize this could affect how speaking volume is picked up, though. Huh, the more you know.
As for the gain, my microphone is plugged into the first port (i.e. the left one) and the knob is now per your suggestion turned all the way to the right (or, up). The right port knob is turned all the way to the left, or down. Neither have INST or AIR selected. 48V is activated, naturally. Direct Monitor is not activated since I don’t have a headphone jack that fits the port (and can’t get my USB headphones to place nice). So, the two right-most knobs I have ignored.
With the gain knob turned all the way up, if I’m sitting still it is dark. The microphone is about a foot to the left of my mechanical keyboard, and while typing this the knob is lighting up, varying from green to some yellow mixed in. When I speak aloud in a normal tone, it varies from green to some yellow, but blips red when I pronounce “CATskill farms” emphasis on the “c” in cat. When I pull the microphone in front of me (distance from my mouth to pop filter is from thumb to pinkie) and speak the sentence, it is a small amount of green, but more yellow and red than anything.
I still have the Yeti and little box I made for it. Trust me, I don’t trust my audio skills enough to just chuck the thing when it was working for me for the longest time. I only upgraded my setup cause I want to continue increasing my quality and ability to take better jobs.
I don’t use any chat programs, although I occasionally will record through User Testing, and their recorded will access the microphone. However, when this happens I go into the system settings and will turn off “Allow apps to access your microphone” for a few moments then turn it back on. This has seemed to return control to me/Audacity in the past. As for startup, none of my startup programs control the mic to my knowledge. I will do a clean shutdown, though.
EDIT: I did the hard shut-down as suggested and I don’t think it sounds any different. I triple checked to make sure Skype didn’t try to steal control/ensured it is denied mic access for future instances.
There’s a reason many studios insist you read from paper.
Do you use wired headphones plugged into the 2i2? That can help you keep level, even volumes over long readings. It’s hard to see, but those headphones in the home illustration are plugged into the little black Behringer interface.
Currently, this is my current setup with the new microphone. I have the swatch under the microphone, foam panels behind the monitor, as well as on the wall to the left of me. And behind me on a large screen stand is a heavy quilt. With this new setup, I am reading from one of my monitors, but previously would read from a phone on a stand next to the microphone. I never noticed it pickup up any sort of electrical whine except if I used a poor quality charging cable for my phone.
I want to use wired headphones for the 2i2, but the ones I have do not fit the jack port. I plan to get an adapter soon for it.
It’s not unusual to play tag with forum posts. I have to leave in a bit, but I’ll be back to see how things went and if you ran into other problems.
It’s not good when someone appears on the forum with a laundry list of effects, filters, and corrections without seeing how a simple recording runs without all that. Just think how pleasant this will be if you can get a good recording without half your corrections.
It’s not the best idea to slide additional effects in the middle of Mastering. The three mastering tools hold hands tightly and clean up after each other.
There is even a one-step Audacity macro that automatically applies all three mastering tools. And I’ll post it as soon as I find it.
Just a parting shot. Are you speaking into the front of your microphone? The side grill just above the yellow dot? The early Yeti microphone instructions had an error and you can still find videos with people using it wrong.
I would love to cut down on the things running in my macro. If you can find the one-step one you are thinking of, let me know! I am no audio engineer by any means, so I welcome advice from someone who knows better than I do. I’ve been content with my sound for a little while, but I want to be better and produce audio I am proud of.
As for the microphone, I am speaking into the front with the gold dot facing me. The microphone itself is upright with the Rode popfilter in front of it and the isolation shield around it. I posted a link to a photo of it in a previous post. I think we might have missed some replies to each other, so I’ll double check. I’ll be back in the morning.
I posted a link to a photo of it in a previous post.
Yeah. I went back and found it.
The quilt on the desk does two things. It keeps rumble and floor noises (street noise, the people downstairs) from reaching the microphone through the mount, and it keeps out sound reflections bouncing from the desk. That can sound like talking into a wine glass.
Mastering guarantees RMS (Loudness) and Peak. If you did everything just right, Noise will fall in and you’re done.
This is a forum posting that explains things.
I think this particular posting was to someone having overload damage problems, so you can ignore that part.
Home performers traditionally have trouble with noise. Background noise. Room Tone. The sound when you stop talking.
If you make specifications after Mastering but the noise is a little too close to failing (-62dB out of -60dB specification), you can apply Noise Reduction of the Beast. That will give you an additional 6dB of quiet. In that example, -68dB noise which is good to go. This Noise Reduction application doesn’t create audible distortion while it’s working. People get into trouble with Noise Reduction when they try to convert complete trash into an ACX submission. That doesn’t work.
Drag-select some clean background sound. This would be the two seconds of silence at the beginning of Catskill Farms for example. Room Tone is built into the ACX submission standard, so you should have some there somewhere.
Effect > Noise Reduction > Profile > OK.
Then select the whole chapter > Effect > Noise Reduction > 6, 6, 6 > OK.
That’s why I call it Noise Reduction of the Beast.
If you’re listening on headphones, you can hear the background noise get quieter. There’s a common misconception that the goal is to make the noise vanish. That’s dangerous because that can damage the voice.
Let’s see. What’s left?
If you use oblique placement with the microphone opposite your cheek instead of your lips, you can get closer and louder without picking up mouth noises and you may not need the tennis racket pop filter. Try that once.
You said something dangerous up the conversation. You didn’t have a clean WAV files of your earlier performances.
There are three recommended files in a reading.
– Raw performance WAV file, mistakes and all. This can be insanely handy if Audacity does something naughty and trashes the chapter edit forever. That does still happen even with Audacity 3.1.3. You don’t have to read it again.
– Edit Master/Archive. Your personal archive master is a WAV file, not the submitted MP3. You can improve and change a WAV file. You can’t edit the publisher’s MP3 without creating damage.
– The submitted MP3. ACX doesn’t accept WAV or any other file types because of strict business reasons. They’re expensive to store and the quality difference isn’t that great. This is also why they recommend Constant 192 Quality MP3—minimum and highly recommend mono and not stereo. They can perform their merchandising tricks with little or no loss in sound quality.
And way down at the bottom. Can you read Catskill Farms without rushing? As above, you don’t have to hit ten seconds if you post a mono track. If you try to post a large file on the forum it will just refuse to accept it. Slow it down slightly. Can you lower your vocal pitch just slightly without fry? I know that “Valley Girl” gritty sound at the ends of words drives some people nuts. This one is my opinion only. Technically, this is between you and ACX theatrical quality control.
You don’t have to be a contralto. One of my favorite readers is Sarah Vowell in just a natural speaking style.
If you use oblique placement with the microphone opposite your cheek instead of your lips, you can get closer and louder without picking up mouth noises and you may not need the tennis racket pop filter. Try that once.
There is some thump, thud, and pop filtration built into Mastering. That’s what the first step, Filter Curve does. It’s based on the wind noise filter built into pro outside broadcast equipment. It also makes it so the next two tools have an easier time meeting RMS (Loudness) and Peak.
The tools clean up after each other, so messing with them is not recommended.
I think I missed more postings than I thought. Is there a second microphone connected to your system? Is that a webcam on top of your monitor?
If you consistently have flat, featureless recordings, you could be recording the wrong microphone. You can do the scratch test. If you (gently) scratch a working microphone, the sound will be loud and immediate.
Nobody wrote you can’t have sound from more than one microphone. It’s unusual, but it can happen.
As a fuzzy rule, good compromise presence can be had with that Hawaiian Shaka spacing.
As you get further away, the sound gets flatter with less richness and emphasis. OK, now get really close and the bass tones will increase. The rich quality and theatrical emphasis gets more obvious. I once played two different people, one man and one woman by doing that. I’m a man with a bass voice. The producer was so impressed with the interchange, he left it in the show. That may have been the high point of my voice career.
Voice volume is much more critical when you do that. Voice control is a big deal.
The recording is a juggling act between the setting of the volume control, distance and positioning of the microphone, and voice volume. I would go with occasional green flashes on the volume knob. The yellow colors signify impending danger but no damage yet. If you get a red flash, record that part over.
I think without question I would try oblique positioning and ditch the tennis racket and see how it sounds. That should allow a good quality recording and a good shot at reading the screen.
These adjustments will go enormously easier and faster if you’re live monitoring on 2i2’s headphones. You can adjust everything in real time.
How do you handle mistakes? Most New Users try to stop and fix the error right there. That’s usually not a good idea because you lose the rhythm and tonal quality of the original sentence. Much better pause briefly and read that whole sentence over again correctly and go back later in post production and cut out the trash. Some people ring a bell or clap so they can easily find the error later.
I looked at the Macro you have linked, and it is the same as what I started with before adding extra stuff, so I’ve adjusted mine accordingly. I also noticed my Limiter was set to “hard limit” and not soft as you suggested, so I changed that.
The noise that seems to be getting picked up in the background for me is white noise of the computer tower, but I don’t think it’s that offense when slightly removed with Noise Reduction of the Beast (love the name, and I’ve ensured my numbers are set to such). I do have to remind myself that some room tone/noise is to be expected. The perfectionist in me wants it dead silent, which I realize is bad. I’ve just listened to too many podcasts where it’s absolutely quiet when they aren’t speaking, so I try to emulate that.
For the admission of not having clean WAV files, I tend to delete files after holding onto them for a month or two, just to keep my computer clear of unnecessary files. When recording a project, though, I tend to keep an unedited WAV file and the finished MP3 file. I suppose I should also start keeping an edited, finished WAV file as well since you make a good point about editing down the line.
As for Sarah Vowell, I’ve always loved her voice. It’s unique and keeps me engaged in the way it sounds. Glad to see some love for her.
I’ve re-recorded Catskill Farms twice and attached. One is with the gain knob all the way up, no pop filter, and the microphone oblique to my mouth. Two is the gain turned down slightly, but everything else the same. No editing to either. I noticed when recording the first one, the gain knob would flash yellow on words. With it turned down slightly, it only ever flashed green. Both of these are the Shaka spacing from my mouth.
There is no second microphone to my system, and scratch test shows it is the NK1 that is picking everything up.
When it comes to mistakes, I will use a keyboard shortcut to “add a label” in Audacity, which creates a label track with a marker at the point of creation. It goes like this: make mistake, shortcut to create a marker, pause, re-read sentence. The marker track being parallel to the main track makes it really easy to find errors and edit afterwards. I used to use a dog toy squeaker to mark things, but find this to be much easier. I try my best to avoid making changes after the fact since, as you stated, the rhythm and tonal matching is much harder. If anything I end up having to record many more surrounding sentences to ensure the edit flows well.
This is another recording of the Catskill, with the gain still turned slightly down, no pop filter, microphone oblique Shaka distance from me. The difference is I put the microphone in my little sound dampening homemade box I was using prior with the Blue Yeti. I think this sounds better? Again, no editing is done to it.
To be honest, I was trying to cut down on the need for the box (since it is bulky) by having the isolation shield, but it doesn’t seem to be doing as much for the mic as I wish.
I think you have the vocal tone fixed. I would not have any trouble listening to a story in that voice. 2 and 3 sound more or less the same and master well.
soft as you suggested
Both Normalize and Limiter were designed to do their job gently without being obvious. In addition, if they’re not needed, the tools don’t do anything.
However. I’m beginning to understand where some of your original effects and filters came from.
I do have to remind myself that some room tone/noise is to be expected.
No, it’s not.
white noise of the computer tower
It’s not “clean” white noise (ffffff) gentle spring rain in the trees. It’s messy motor, fan, and ventilation noise. One of the loudest tones is 120 Hz (low growling) which explains that 120Hz filter in your original processing. That plus Noise Reduction to get rid of everything else.
This is the decision tree.
– I can try out filters to get rid of that trash. That’s not recommended because those tools all affect your vocal tones, and I like it the way it is.
– You can move the tower out of the room. You can’t separate the 2i2 and the computer by more than about 6 feet (USB cable) but you can have a good quality XLR microphone cable almost as long as you want. It’s common to buy XLR cables in 20 foot lengths and you can plug them into each other.
If you can’t manage the screens, read from paper.
– Try to “hide” the tower from the performance area. This is another juggling act. Make a quilted room for the tower, or put quilts between the microphone and the tower. You can’t ever block the tower air flow, so there’s lots of decisions. Do you have the tower on the floor? Put a quilt under it on the floor.
If the tower is on your desk, put a quilt under it and hang a quilt between it and you. You can hear all these tricks working in real time. It’s just normally you don’t pay any attention to the noise. I have a silly line that if you can tell your computer is on by listening, it shouldn’t be in there with you.
There’s another, less obvious decision. If it’s a career move to tame the computer problems, stop recording on the computer. I have produced acceptable quality voice recordings on my Zoom H4, my Zoom H1n sound recorders and my phone.
That’s a little Hollywood license. That wasn’t tall enough, so I had to add a roll of toilet paper. That was the script I read.