Zoom sound recorder and/or DSLR mic input

[BeyondBeatenTracks]

I currently use a Zoom H2n to record the sound.
To avoid having to manage and synchronize one video and one sound file for every shot I make, I would like to record the sound from the H2n into the DSLR through the mic input.
I got a 25dB attenuator cable, and when I made some tests, I realized that what comes out through the headset/line out of the H2N is not just the same audio than the recorded WAV file; at least it differs by the volume, which for the cable is the headset output volume... By default, this level was 100%, which makes something inaudible. I read that you can get something better by setting the line out level to 50 or 60%, but now I am hesitant...
Has anybody here an opinion about the difference of quality between the audio directly recorded by external recorders such as Zoom, and the audio that can be injected in the DSLR using a cable? Do some people here use the Zoom this way (cable), eventually discarding the directly recorded files?

 

[BeyondBeatenTracks]

I will reply to myself...

I did somme testing, and I realized that the video recorded in the DSLR (actually in my case, a mirrorless, Canon M50, which I use with a Zoom H2n) depends on three factors:
- DSLR input gain
- Zoom input gain
- and Zoom line/headphones output level.
So, yes, what you hear in your headphones if you plug them in the Zoom jack output depends on both the headphone output level and the gain level, which according to me makes it a bit hard to have a real idea of the actual recording level. Anyway...

I did a range of tests, adjusting one parameter and keeping the two other steady, and to make it short, here is what I found:
- if you set the Zoom line/headphone level to 50 or 60 (on a scale from 0 to 100), and the M50's input gain to 1/2 (on a scale from 0 to 1), the audio levels are very similar between what will be recorded in the M50's MP4 file, and the Zoom's WAV file, when you make the Zoom's input gain vary between 1 and 10.
- for normal ambiant noise level, setting the Zoom input gain to 8 gives a correct level, the VU-meters doing no more than flirting with -12dB on the Zoom as well as the M50. If the ambiant noise is louder (just in front of a waterfall, trafic, loud music, ...), better set the Zoom input gain to 6 or 7 (adjusting using the VU-meters).

There doesn't seem to be any significant signal loss "through the cable". The M50's recorded audio sounds very similar to the Zoom's WAV file, and is very recognizable from the M50 internal mic's audio which is much poorer in low bass and high trebles. Visualizing the frequency spectrum with ffmpeg's showcqt filter is consistant with the listening experience: no obvious visible difference between the M50's cabled recording and the Zoom; obviously truncated spectrum for the M50's internal mic.

I stay interested if audio gurus here have experience or opinions about this kind of things (do you use an attenuator cable? In post-production, do you keep the file from the Zoom recorder, or do you just use the audio from the camera's video file?)

 

[mirrorlessNY(youtuber)]

man sound can be so complicated,,,,,,, put that on top of video editing, video recording and thumbnail design and you get a picture of how insainly complex youtubing can get

 

[BeyondBeatenTracks]

man sound can be so complicated,,,,,,, put that on top of video editing, video recording and thumbnail design and you get a picture of how insainly complex youtubing can get

And to make it worse, once you get your sound file(s), you can normalize.
I just tried that with ffmpeg-normalize. Audio in some Zoom files was really low, so I used normalization, and now the level is more "normal" and it fits nicely into the edited video. I think this is possible because Zoom WAV files contain some high definition video, maybe the result would not be so nice with MP3.

Again, if some audio guru here, ...