Bottom Fan 1.0 vs 2.0 after post Noise Reduction?

[Kaan A]

Hi all,

Wondering which fan might impact recordings of the human voice the least, or which might be most transparently removable in post.
Has anyone run noise reduction (RX or Audition) specifically on dialogue, both with Bottom Fan 1.0 and 2.0?

I've heard that fan 1.0 has broad-spectrum white/brown style noise. And fan 2.0 has noise that's more distinct and more localized to certain frequency bands.
Anyone compare the results after removing either type of noise? Is one cleaner to remove than the other?

Thanks!

 

[Stephen Pruitt]

I'd absolutely love to know the answer to this, as well.

Stephen

 

[Patrick Tresch]

I personnaly do prefer V1 (with the FanPlate) but never did the test to find out wich has a best noise to clean in post. Jwsound is your best friend for this question.

 

[Kaan A]

Thanks, Patrick!

Actually, if anyone happens to have a WAV file of a boomed actor with the 2.0 fan noticeably impeding on the quality of the track, would love to take a listen. If you wouldn't mind posting/sending the file, I can try some noise reduction on it in RX5 to see just what can be done with it.

Thanks!

 

[Patrick Tresch]

Hello Kaan,

I only have this vimeo but no voice on it...

https://vimeo.com/75245337

Pat

EDIT: I found this with my voice indentification. Hope it's not too lost in the room...

http://we.tl/hBnCrRSuIg

 

[Kaan A]

Merci pour les enregistrements, Patrick! I'll have a listen and see what I can learn from these three files and get back to you here.

 

[Daniel Kelly Brown]

I have a V1 fan for sale if anyone is interested. Make an offer.

 

[Kaan A]

Okay, here's an attempt at fan 2.0 noise reduction, using one of the files Patrick posted. Thanks, Patrick!

Epic-X with Bottom Fan 2.0 and Top Fan
This example shows a before and after of manually notching out steady fan noise frequency bands in Izotope RX5 using the spectral display to manually identify them. I notched out approximately 30 or so bands, each from approx 3Hz to 15Hz wide, across most of the recorded spectrum. It's tedious, but doable if you have the time in post. Note, this isn't using a "Noise Reduction" process, only manually deleting small bands of offending frequencies which are clearly visible in a spectrographic view such as the one RX or Audition provide.

If anyone has other real-world recordings with Fans 2.0, would be nice to play with to see if this kind of notching becomes too noticeable in certain situations. In this recording, I notice some weirdness on the word "quarante"

 

[Patrick Tresch]

Thanks for your test. Impressive results.
Do you think that the broader and lower frequency noise of the fan V1 would be more tricky to delete?

Pat

 

[Kaan A]

Yes, I think so, Patrick, mainly because it's so random that notching wouldn't work well. That's why I'm most curious about trying to clean a "new bottom fan, fan plate" combo. Would you happen to have your voice on that take too?

Thanks!

 

[Kaan A]

Here are some more comparisons, using the files Patrick posted. Thanks again, Patrick.

Epic-X with Bottom Fan 1.0 only (12dB Noise Reduction)
A before and after of a 12dB noise reduction process in Izotope RX5. Using Advanced + Extreme algorithm based on an 8-second noise sample. More extreme levels of noise reduction were possible, but with the trade off of affecting voice quality.

Epic-X with Bottom Fan 2.0 and Top Fan (Notching and 12dB Noise Reduction)
A before and after of a combination of manually notching out steady fan noise frequency bands in Izotope RX5 using spectral display and ALSO doing a 12dB noise reduction process (Advanced + Extreme algorithms).

Comparing Noise Reduction Results: Bottom Fan 1.0 only versus Bottom Fan 2.0 + Top Fans
This is an attempt to answer the questions posed by this thread. Of course, this is only with a few seconds of one person speaking off axis. It would be way better to be able to compare both male and female voices, under regular shooting conditions, boomed on axis, etc.