Fixed gain field recorder?

[MarcelvdG]

What's more refined about it?

 

[pos]

They claim a much wider dynamic range, and Microtech Gefell has a better reputation and track record than Zoom when it comes to actually backing up their claims.

I also read bad comments about zoom's approach in their field recorder which supposedly implement the same approach as their new interface.

Eg with the F3, a user reporting audible clipping on transients: https://www.thomann.de/fr/zoom_f3.htm#bewertung
Google translated from French to English:

It's the second one I own and despite all its qualities, it seems to have trouble with certain transients in the bass and clip sometimes in the middle of the attack. This only happens on certain types of sounds but with guitars it becomes very audible as you exceed -30 dBfs. On a bass guitar it's unacceptable and impossible to take advantage of the last 30 dB of dynamics.
I've already talked about it with Zoom support and I'm waiting for feedback from other users to get back to them, but it's very unlikely that I ended up with two faulty F3s in a row.
However, the background noise is very good. -130 dBfs (A) measured and up to +0.10 dBfs without distortion (0 dBfs = +4 dBu I assume). A light and practical format. For Field Recording it is perfect but for music production it is not, Dual ADC technology is not optimal in this case. It's also not too bad because it remains very affordable and performs well in the majority of cases. Proof of this is that I am the first to report this defect.

I don't know how much they improved their process with the newer interface though, but for now I would take a good "fixed" preamp/ADC over this any day of the week.

 

[TNT]

This just in...

Big baraboom...

The Kings birthday and Sunday - that means a 21 shots salute at 1300 hours.

Watch your gear...

I had to gain this only 2x6dB, ending up as:

Minimum sample value -0.6132 -0.7208
Maximum sample value 0.4991 0.5985
Peak amplitude -4.2 dB -2.8 dB
DC offset 0.0001 0.0003
Minimum RMS power -65.9 dB -62.4 dB
Average RMS power -31.4 dB -30.1 dB
Maximum RMS power -16.5 dB -15.2 dB
Clipped samples 0 0

No other edits than gain.

 

[MarcelvdG]

So you originally had 14.8 dB of headroom?

 

[TNT]

Yes. I was about where I took the photo - say 100m away

//

 

[Rick PA Stadel]

Brilliant -- love it! Thank you. What a gift to the rest of us!

Cheers

 

[TNT]

Improvements for UGFR... it's a constant development

Theme: Fs filtering and EMI susceptibility.

Two actions:

a) Add analog filter to attenuate energy above Fs/2
b) Shield whole recording system from outside interference.


a) Designed and built anolg 5th order LP filter:

Approx 100 dB down at Fs/2 if at 384 ksps:

Made me some coils with the help from a 3D printer owning friend

Flaskships...

b) Sheilding with the help of copper tejp. Worked really well...

The whole recording system is one single continuous cage - mic, cables, "box" and ADC. Hmm, well, the battery pack remains... but great effective as we shall see...

Measurements:

Analysis with continuous cage connected to earth vs. not connected to earth. 30 dB diff :

Same recording but processed to 24/44,1... comparing two sections of same recording - one I did no noise, the other I struck a glass with a metal knife - focus on "hf"... NB1 frequency axis... NB2 - the very low rise level at near FS/2 - NB3 - this is from the recording system as intended for field use - recorded at 24/384.

I know these absolute numers are probably not correct - but the difference is probably... Seems promising - now out on the filed ;-D

Data above is from my DAV, Amadeus Pro (Mac).

//

 

[TNT]

Here are two excerpts from yesterday recording with the revised gear and down-sampling process. A is from the silent beginning and B is from the roaring end

The list of improvements:

• Added analog anti-aliasing filter.
• Partly new cables.
• Farady cage created
• New resamplingsprocess - 24/384->24/44,1 by CamillaDSP
• Dedicated battery for the Cosmos ADC

Complete recording data:
L , R
Minimum sample value -1.0000, -0.9924
Maximum sample value 0.9867, 1.0000
Peak amplitude 0.0 dB, -0.0 dB
DC offset -0.0000, -0.0000
Minimum RMS power -59.6 dB, -59.4 dB
Average RMS power -22.8 dB, -23.1 dB
Maximum RMS power -11.4 dB, -11.7 dB
Clipped samples 9, 3

Improvements remaining:

• USB isolater for the ADC to get the phones earth out of the picture
• Include the batteries in the Faraday cage for a complete EMI isolation.
• Mechanical decoupling of the stand and Jecklin disc.
• Calibration of each microphones to a Earthwoks M30 - FR and phase, using the complete recording chain i.e. including also filters and AD.

Feedback welcome.

//

 

[TNT]

I did two things:

• Calibrated each mic towards a Line Audio OM-1 and convoluted FR compensation using CamillaDSP.
• Improved the downsampling process.

I think that made a substantial improvement. Please compare...

//

 

[TNT]

Embarrassment again... the attachment in post above don't contain any impovements as my processing chain didn't do what I expected. And the expectation effect was apparently so strong that I thought I heard what I anticipated... well, well... Yes, I used the word "substantial"... gulp.. to my defence, there where some actual differences made but not at all what was intended... so...

I have now learnt my lesson and would like to present a verified version V4 of the process... it works now and I like the result.

Also, on may way home form the church, I saw a man playing trumpet on s big square in central Stockholm - here is snippet of him too.

Please feedback - bad or good - I need it for progressing UGFR many persons and systems is needed to get a view...

//

 

[Philbo King]

The new 32 bit FP recorders can do this.

An example:
https://zoomcorp.com/en/us/handheld-recorders/handheld-recorders/h4essential/

 

[TNT]

Okey, then I trash my gear - thanks for the tip!

How did my recording sound in your system?

This is a DIY forum, not a consumer of the shelf product recommendation site - right?

//

 

[Markw4]

Please feedback - bad or good -

Sounds better than the last time I listened. It might help if you recorded two people singing in harmony, since we tend to notice a lot about the sound of human voices. Also maybe a piano. Does it catch the sound of note where there are supposed to be 3-strings in tune with each other, but one is a little out. How about the textures of chords, the sound of hammers on the strings. Maybe something with some percussion too, not just droning. It might also help if you recorded with same source at the some time with another system so there is something to compare specific differences in how the sound is captured. IOW, the material you are recording is without certain types of sound stimuli, and without a whole lot of reference to things that are very familiar to most people and or things that are easier to tell what's right and what's wrong.

EDIT: it can also be helpful to record someone hitting a wood block and or claves and then listen percussive attack sound and to the room decay.

 

[TNT]

Well, it's an organ. In a church. But I would like to say that this particular one is a bit screechy in its top registers. At the final go, my ears ring... and it doesn't have 32" pipes so the lowest tone its about 58 Hz if I look in the FFT... so not that super impressive as other churches, but still should sound quite massive/potent in a "final push"..

https://sv.wikipedia.org/wiki/Adolf_Fredriks_kyrka#Orgel - it says that there is a "Subbas 16'" - the should be about 30 Hz - perhaps it broke nor not played.

This is from week before which had a more grand ending actually: (lot of fans and traffic etc infra garbage...)

So Mark, in your room, in your system, did you seem to be transfered to the location and got the feeling of a somewhat bigger church?

If not, what was the experiance?

I'm searching for other sets/instruments to record - not so easy - this place happens to allow recorders... I add a more "atmospheric" passage from the recording below - see what you think about it.

//

 

[Markw4]

Sounded a bit muddy in the midrange frequencies and rolled off in the highs. The room sound echo/reverb decay kind of cut off early; should have continued down to a very low level. I'm guessing some IMD and limited dynamic range?

EDIT: Are you using electret mics? If so, what bias voltage are you using?

 

[TNT]

The mics don't use the normal 48v phantom. The schema can be seen here:

https://www.diyaudio.com/community/threads/fixed-gain-field-recorder.373352/page-25#post-6845244

//

PS. I see now an error in the drawing - U8 is also a LT3045...

 

[Markw4]

Looks like 8v through about 5k? Maybe that's what the manufacturer recommends for those? I used to use some Shure MX184 (actually the R184 capsules) with 2.2k and a 9v. They are super cardioids:

Just wondering, since it all starts with the mics and what they are capable of.


Also being the experimentalist I am, if I were looking into why the sound of a recorder is the way it is, I would probably skip the AD/DA steps and just listen to the mics through an HPA and some cans to get an idea of the sound with an analog-only signal path. Then I would include the AD/DA in to see what that's doing. Just thinking its easy to measure the electronics; transducers are a bit different animal. My preference would be to listen to the mics along with any measurements of them.

 

[TNT]

Primo Em273

//

 

[Markw4]

Okay. 8v with a dropping resistor is not the same as regulated 5v?


Also, some electrolytics in the sound path area. Is their value high enough to keep the AC voltage on them less than 60mV at the lowest pipe organ frequency? Also sometimes it does help HF if electrolytics are bypassed with a decent film, such as Wima MKP, .01uf. (MKP may sound better for that application than FKP; sometimes it helps to get HF sounding a little better).

Looking at the mic specs some more, 80dB SNR @1kHz could be an issue. Maybe not going to get to CD quality (especially not like a well dithered CD).

 

[MarcelvdG]

At 1 Pa.

That's 14 dB(A) equivalent acoustical noise level, 3.5 dB lower than my AKG C900 stage microphones, but worse than typical large-diaphragm condenser microphones.