Apply fxs 2 make podcast sound as good as possible (podcast)

Hello friends and thank you for all the support and valuable information that you all taught me.
Special thanks to: Trebor; kozikowski; billw58. They gave me very useful tips to remove the background noise.

https://forum.audacityteam.org/t/firebox-or-inspire/174/1

Now that I achieved a clean background noise, I am now once again in need of your kind expertise:
Please listen to this sample audio clip. It is after I applied all the steps in the link (please see below steps).
I wish my audio clips would sound like the audio talk shows.
What effects do you suggest I need to give my podcast the missing ingredients to sound as good as possible?


I. Sound Formatting exports
A. While Recording: Use 44100 Hz and 32 bit float.
B. Export a raw capture copy at 32-bit WAV (and back it up)
C. Do all editing in Audacity at 32-bit float (that gives you the headroom you need)
D. Export final production to WAV 44100 Hz and 16 bit float (and back it up) - this is your master and you can burn CDs from this.
E. Export to MP3: 128 kbps - for use on a portable player

II. Suggested work flow to reduce noise:

  1. Amplify on default settings, (i.e. to 0dB).
  2. select a region for the noise-profile which has no voice.
  3. Apply Audacity’s native Noise Reduction (with the settings).
  4. Select all of your track and put the code below in the Nyquist prompt to remove the faint 1KHz whine + 10 harmonics.
  5. Select any conspicuous unwanted noises between phrases, e.g. 11.40-11.60sec, press “Ctrl” +“L” to silence them.
    Note: Make sure you remove all the unwanted noise in the silence parts (conspicuous unwanted noises (spikes)) which occur between phrases. Silence spikes by selecting them and pressing ‘Ctrl’ +‘L’, then apply noise-gate.
    Then select the entire track and apply a noise-gate to squelch the noise between the phrases.
  6. Select the entire rack and apply a limiter to reduce the variation in volume, and make the track generally louder.
  7. Use Notch Filter at 171Hz Q=10.
  8. Steve’s pseudo-stereo plugin (“make stereo track”).
  9. and de-essing via “Spitfish” plugin.

If I listen on my laptop speakers, it sounds OK and you’re done, but if I listen on headphones or a very good quality sound system, you still have a little science fiction “martian voice” as a holdover from doing all those corrections.

You are trying to simulate in software having a very good quality microphone working correctly in a quiet room. Improvements may require you fix the studio.

The only other thing I would change is your “lecture voice.” Successful radio performers always sound like they are speaking without question to you and not to a room of 300 anonymous students. You also sound a little like you’re reading from a page and not extemporaneous. That works OK if you’re trying to put your kids to sleep…

One of the reasons free-form multi-voice podcasts work is the performers are clearing having fun with each other and the banter is unexpected, even if the sound quality isn’t the best. A local radio station lost its afternoon drivetime announcer and decided over weeks to try different combinations of staff announcers. Almost as an afterthought they tried a grizzled veteran announcer-presenter paired with a young squirt. They had the control room on the floor laughing and radio magic was born.

It’s hard to make a Nyquist filter do that.

Koz

The Martian effect may be due to the multi-notch filter on step 4 …

  1. Select all of your track and put the code below in the Nyquist prompt to remove the faint 1KHz whine + 10 harmonics.

You could try leaving that step out if it’s doing more harm than good.

Always just use the least amount of noise reduction possible [step 3] to minimize the computery artefacts which it can generate.

You can save disk space by saving the first non-modified original recording as 16 bit.
Audacity records always at 16 bit. the 32 bit quality applies only to afterwards applied effects (amplify for example).
Dither should of course set to none for this first backup.

The Martian effect may be due to the multi-notch filter on step 4 …

To be clear. The effect is almost invisible on good speakers, but it’s prominent on headphones in a quiet room.

So then all we have left is reading technique. Do you pronounce “a” like that when you’re ordering a chocolate milkshake?

And the need to apply all those filters and corrections every time a new recording is performed – and remembering what you did. The majority of making a show is not aiming a camera. It’s shot and clip bookkeeping.

Koz

Hi and thanks to all of your feedback! Sorry for not replying earlier, I have been trying (trial & errors) to minimize the alien sound as much as possible.
Here is what I came up with. I applied all the steps 1 to 9 (minus step 4).

Please Koz and friends, is the attached audio clip better? does it sound better? if yes, do I need to add any more effects? like bass boost? or something else?

Thanks to all of you for your continuous support!

Hi again, I thought maybe to make things easier for you friend.
I think it is better to attache the raw file of my recording, instead of the mp3 (please check out the attachment).
This way you can easily tell me what I am doing wrong.

Here are the detailed steps that I used: https://forum.audacityteam.org/t/does-noise-removing-twice-badly-impact-the-actual-rec/30078/1

and here are the bullet points of the actual steps that I normally use (step 4 is removed):

Your kind help and advices are truly appreciated, and I will always be grateful!
All the best!


I. Sound Formatting exports
A. While Recording: Use 44100 Hz and 32 bit float.
B. Export a raw capture copy at 32-bit WAV (and back it up)
C. Do all editing in Audacity at 32-bit float (that gives you the headroom you need)
D. Export final production to WAV 44100 Hz and 16 bit float (and back it up) - this is your master and you can burn CDs from this.
E. Export to MP3: 128 kbps - for use on a portable player

II. Suggested work flow to reduce noise:

  1. Amplify on default settings, (i.e. to 0dB).
  2. select a region for the noise-profile which has no voice.
  3. Apply Audacity’s native Noise Reduction (with the settings).
  4. Select any conspicuous unwanted noises between phrases, e.g. 11.40-11.60sec, press “Ctrl” +“L” to silence them.
    Note: Make sure you remove all the unwanted noise in the silence parts (conspicuous unwanted noises (spikes)) which occur between phrases. Silence spikes by selecting them and pressing ‘Ctrl’ +‘L’, then apply noise-gate.
    Then select the entire track and apply a noise-gate to squelch the noise between the phrases.
  5. Select the entire rack and apply a limiter to reduce the variation in volume, and make the track generally louder.
  6. Use Notch Filter at 171Hz Q=10.
  7. Steve’s pseudo-stereo plugin (“make stereo track”).
  8. and de-essing via “Spitfish” plugin.
    1- A Journey.zip (999 KB)

It sounds good. IMO it’s slightly staccato which is fixable with more dynamic-range-compression (via limiter again or a compressor like “blockfish” or “Chris’s Compressor” , or Audacity’s native compressor ) …

If you do apply more dynamic-range-compression you will probably have to apply more de-essing.

PS I’d reduce the strength of pseudo-stereo effect a bit , but that’s just my personal taste.

Hi Trebor and thanks again for all of your contributions!

  • pardon me for the request here. I have attached the step by steps of what you suggested that I should do with the raw clips.
    could you please take a look at the steps, and please make the adjustments that you made in the last post?

*And do you think I might need a bass boost?

Please note: I want to ensure that I follow your tips and instructions to the t. I have 30 audio podcasts, and I want to ensure that I am following
the steps correctly with the 1st one, before I proceed with the rest (29, and future podcasts).
It took me 3 years, 3 book editors to go thru the script, and 14 other friends that helped me collect and compile the info.
I am grateful for your understanding, and your help will give the podcast that ingredient that will give it a lasting flavor.
The suggested steps.zip (201 KB)

These were just guidelines , not written in stone, feel free to experiment …

Unless you’re trying to compromise mp3 file-size I’d use a higher kbps say 192 or 256Kbps for a (pseudo)stereo recording with a sample rate of 44100Hz, (the larger the kbps the larger the mp3 file will be).

Using 8dB rather than 10dB may be worth a try, (i.e. use as low a value as is acceptable to you : it’s a trade-off : more noise-reduction means less hiss but more computery artifacts , 10dB is a typical value).

The -3dB value shown on the limiter is the default setting , using say -6dB would avoid the need for further dynamic-range-compression later.

That step will be specific to the recording apparatus you used : so not a general rule.

A lower value say 10 would give a more subtle pseudo-stereo effect, it’s a matter of taste.
If you are posting the audio on the internet and are obliged to minimize file size then abandon pseudo-stereo and stick with mono.

The 7KHz value I selected is based on looking at the spectrogram of your recording, on other recordings of other people or using other apparatus the optimum value may be different, (it’s in the range 4KHz - 8KHz )

BTW I recently found another free de-esser plugin … https://forum.audacityteam.org/t/just-tried-this-free-de-eser-vst-called-lisp/31088/1

IMO no : the equalization sounds just right. More dynamic-range-compression will make it sound beefier if you want, (via limiter or a compressor).