Fixed gain field recorder?

Here is a field recording in a sunken down square in central Stockholm. Concrete as far as the eyes can reach 🙂

Solo trumpet, mics 4 meter away low - I didn't rig the stand as this was made in a whim. I asked or permission to record. The man playing also held a speech / prayer - the level is adjusted as the trumpet was much stronger than the voice. Amazing what energy a single trumpet really has... I reflected while recording that the trumpet sounded quite small and thin.... he also gathered a lot of saliva which he had to pour out once in a while (hence some gargling/crackle.-.) - I edited these intermissions out 😉

//
 

Attachments

  • Like
Reactions: MarcelvdG
No change. But as I was thinking about the answer and became a bit unsure it dawned on me that I need to develop a version handling and file naming system to keep track which conversion process version a certain "release" was made form. I'm quite sure I'm not at the final variant so it will be needed.

//
 
Yet and excerpt from the latest organ recording session (No.9). It is the same HW as the last one. Mic positions quite radically changed. Mostly interested to hear a judgement about stage / height for this one but all impressions are valuable thanks!

//
 

Attachments

A short snippet of "jass-jam"... no processing other than gain by bit-shift and down-sampling by CDSP (Nyquist true).

Minimum sample value -0.6410 -0.8256
Maximum sample value 0.7237 0.8390
Peak amplitude -2.8 dB -1.5 dB
DC offset -0.0000 -0.0000
Minimum RMS power -78.5 dB -78.5 dB
Average RMS power -19.9 dB -19.3 dB
Maximum RMS power -9.1 dB -7.4 dB
Clipped samples 0 0

//
 

Attachments

Last edited:
New capture from the streets.... really awesome jazz here... 5 afternoon in a public square. Some wind...

Rig now with working USB-isolator to get the phone ground out of the picture...

Minimum sample value -0.7188 -1.0000
Maximum sample value 0.8036 1.0000
Peak amplitude -1.9 dB 0.0 dB
DC offset -0.0000 -0.0000
Minimum RMS power -54.5 dB -53.8 dB
Average RMS power -30.3 dB -27.1 dB
Maximum RMS power -17.8 dB -15.3 dB
Clipped samples 0 16

24_44 for your pleasure...

//
 

Attachments

Made an even less processed version... the audio data above was for the whole recording so not describing the excerpt only I'm afraid.

Here is that data:

Minimum sample value -0.3494 -0.4211
Maximum sample value 0.3079 0.4730
Peak amplitude -9.1 dB -6.5 dB
DC offset -0.0000 -0.0000
Minimum RMS power -58.6 dB -55.8 dB
Average RMS power -30.6 dB -27.9 dB
Maximum RMS power -22.8 dB -18.6 dB
Clipped samples 0 0


Here is data for the below part only:

Minimum sample value -0.3481 -0.4192
Maximum sample value 0.3092 0.4745
Peak amplitude -9.2 dB -6.5 dB
DC offset 0.0002 0.0006
Minimum RMS power -44.0 dB -42.0 dB
Average RMS power -30.6 dB -27.9 dB
Maximum RMS power -22.8 dB -18.6 dB
Clipped samples 0 0

Please compare and report finding if any!

//
 

Attachments

Revisited the same church as the other snippets but now with an USB isolated phone.

Info L R
Minimum sample value -1.0000 -1.0000
Maximum sample value 0.9778 1.0000
Peak amplitude 0.0 dB 0.0 dB
DC offset -0.0001 0.0047
Minimum RMS power -65.2 dB -63.0 dB
Average RMS power -14.9 dB -14.8 dB
Maximum RMS power -10.1 dB -10.0 dB
Clipped samples 1 13

//
 

Attachments

I have listened extensively on the above recording and find it considerable better than the previous ones. Phone apparently contaminate the ground etc in the recording system. It sounds to med so much more effortless and with more low level content. Really nice. I might also now have a chance on a much bigger and airy church and its organ... lets see what happens... 🙂

//
 
Want to my church for another recording session - this time to make recording to compare my Cosmos ADC to "another one"... The person usually playing the organ was not able to play this day so for a stand in, two persons singing and playing piano was present and approved on me setting the mics up.

This picture shows the setting. Mics say 6-7 meter away and I sat 1 meter behind them. I judge both recordings in general to describes very well what I heard. Hard acoustics of course in a huge building almost only made out of stone. Female singer had a strong voice that for my ears was a tiny bit challenging in the fortes.

I started with ADC A and changed to ADC B half way through - this meant changing a pair of signal in XLRs, a battery and a USB connection between the ADCs. Mic position was unaltered.

Recordings was in post gained to nearest 6 dB step to clip - this meant 7 times 6dB = 42 dB with only a dB or so left to clip for one ADC and about 5 for the other. The clips I attach below has been further level to be reasonably equal. Both ADC worked at 192Ksps. Excerpts below are 24/44,1.

Please have a listen and comment on any differences you might find. Also preferences are of interest.

AF.jpg


Note stand for pianist and singer are circled in orange. Mics pointing approx between them with a little bias on singer.

//
 

Attachments

I definitely prefer ADC A.

ADC B inverts the sound staging. The singer is behind you, not in front of you as with ADC A. I can also not hear as deep into the room with ADC B as with ADC A. In general the room is much better presented in A than in B. A is more detailed. This also makes A sound what some individuals describe as warmer, B colder.

The sound stage inversion can happen with LM4562/LME49720 (but these tend to be very detailed) or if resistors/capacitors are unluckily orientated (all in the same - well - apparently "wrong" direction).
 
  • Thank You
Reactions: TNT
Stats for the posted files just above. There is an other difference than bit depth compared to the first "release" and that is that the last ones dont have its DC component removed... sorry... and I see that they are not perfectly matched level wise... sorry again.. maybe you want to fix that yourselfs... but its also a bit hard considering different song in different part of the concert. Its probably quite OK as is 🙂

A:
A.jpg



B:
B.jpg
 
Last edited:
Just a minor note - both flacs are 44.1/16. Perhaps the original 192/24(32) recordings would be interesting too, thanks for considering.
192/24 14s snippets... 🙂

A:

Minimum sample value -0.8688 -0.8607
Maximum sample value 0.9464 0.8930
Peak amplitude -0.5 dB -1.0 dB
DC offset 0.0734 0.0336
Minimum RMS power -21.6 dB -26.4 dB
Average RMS power -16.6 dB -18.2 dB
Maximum RMS power -10.3 dB -10.7 dB
Clipped samples 0 0


B:

Minimum sample value -0.9031 -0.8284
Maximum sample value 0.9741 0.8914
Peak amplitude -0.2 dB -1.0 dB
DC offset 0.0002 0.0212
Minimum RMS power -32.3 dB -31.2 dB
Average RMS power -18.0 dB -18.7 dB
Maximum RMS power -10.9 dB -11.4 dB
Clipped samples 0 0


//
 

Attachments

Last edited:
Thanks for noticing!

Problem is I think: top is with DC removal applied, bot don t have this. Run it through a 1Hz HP filter and check it.

I might actually re-upload new versions with DC removal...

I also had a L/ R channel switch - one can tell by comparing... bottom one is correct.

Sorry for the confusion...

//