I had a look at the part around 37 seconds and beyond with the GoldWave audio editing program and its moving DFT plot. It looks like plain old harmonic distortion to me; all peaks appear to be at integer multiples of the test tone and all peaks move to the right when the test tone frequency goes up, the higher harmonics faster than the lower harmonics. At frequencies above 20 kHz, I sometimes see something move to the left for a second or less and disappear, which is to be expected when half band filters are used in the ADC decimation chain (and as they are cheap, all DAC and ADC chip manufacturers I know of use half band filters). The passband ripple of the decimation filter is clearly visible.
Even at frequencies just above 1 kHz, the third and fifth harmonic distortion products look stronger with than without the modified Xvive. Is this still with line levels?
Even at frequencies just above 1 kHz, the third and fifth harmonic distortion products look stronger with than without the modified Xvive. Is this still with line levels?
Dear Marcel,
„Sweep_direct-6dB“ is without the modified Xvive inserted.
„Sweep_Xvive-6dB“ with the Xvive inserted. Both are recorded at mic level, as this will be the real world
scenario when shooting handheld.
But i chose the incoming signal rather hot at -36dB, a deliberate worst case scenario. To get the best S/N ratio the camera input is damped at-12dB. The firmware does not allow to get lower than -12dB. As the mics signal is mono, there is also a voltage divider inserted, to have a -9dB backup on the right channel. What you got is the right channel. The left
channel would already clip by 3dB but has a limiter. Again, by purpose, this is the hottest what the camera would get. I can make a test tonight with a modest signal,
with -12 to -9dB headroom, which
would be a more average level.
„Sweep_direct-6dB“ is without the modified Xvive inserted.
„Sweep_Xvive-6dB“ with the Xvive inserted. Both are recorded at mic level, as this will be the real world
scenario when shooting handheld.
But i chose the incoming signal rather hot at -36dB, a deliberate worst case scenario. To get the best S/N ratio the camera input is damped at-12dB. The firmware does not allow to get lower than -12dB. As the mics signal is mono, there is also a voltage divider inserted, to have a -9dB backup on the right channel. What you got is the right channel. The left
channel would already clip by 3dB but has a limiter. Again, by purpose, this is the hottest what the camera would get. I can make a test tonight with a modest signal,
with -12 to -9dB headroom, which
would be a more average level.
Wow - revealing!
I did test the sweep with the camera's (Panasonic Lumix S5)
input set to Mic Level (-55dbV) and also set to Line Level (-10dBV)
I did test Mic Level twice: With the modded Xvive inserted and without.
I did test Line Level without the Xvive as we already found out
that the the audio transformer will distort.
Signal on all tests has -9dB headroom on the left channel, -18dB on the right channel.
It was dampened digitally by -45dB for testing it with the mic inputs.
I strongly assume with 24bits of resolution, ran through a 64bit system,
it is unlikely that the distortion was introduced by the source when reducing amplitude.
The DAC for playback is 32bit, Twisted pear Buffalo II with Ivy Output stage, OPA1632.
With the camera set to Line level, the signal audibly has less distortion and does not cut the bass.
https://filetransfer.io/data-package/MwUjrZur#link
I did test the sweep with the camera's (Panasonic Lumix S5)
input set to Mic Level (-55dbV) and also set to Line Level (-10dBV)
I did test Mic Level twice: With the modded Xvive inserted and without.
I did test Line Level without the Xvive as we already found out
that the the audio transformer will distort.
Signal on all tests has -9dB headroom on the left channel, -18dB on the right channel.
It was dampened digitally by -45dB for testing it with the mic inputs.
I strongly assume with 24bits of resolution, ran through a 64bit system,
it is unlikely that the distortion was introduced by the source when reducing amplitude.
The DAC for playback is 32bit, Twisted pear Buffalo II with Ivy Output stage, OPA1632.
With the camera set to Line level, the signal audibly has less distortion and does not cut the bass.
https://filetransfer.io/data-package/MwUjrZur#link
@MarcelvdG I am already contemplating to toss out the transformer and put in your OPA1612 based design.
https://www.diyaudio.com/community/attachments/micverstb-png.903067/
Voltage would be 5V. No need for a pot.
What resistor values would I need to amplify according to the cameras specs, from -55dBV to -10dBV?
I assume it is roughly about 45db? The transformer added +6dB, so roughly 50dB.
Will 5V be sufficient for this task?
https://www.diyaudio.com/community/attachments/micverstb-png.903067/
Voltage would be 5V. No need for a pot.
What resistor values would I need to amplify according to the cameras specs, from -55dBV to -10dBV?
I assume it is roughly about 45db? The transformer added +6dB, so roughly 50dB.
Will 5V be sufficient for this task?
It would need some modifications to survive phantom supply, I would have to look into that.
Just checking:
Are you aware that the camera has different wind suppression settings for internal and external microphones?
The microphone amplifier has an input resistance of about 2 kohm, the line input presumably has a higher resistance. Could that affect the DAC output? That is, is there a relatively small DC blocking capacitor in the DAC at the output?
Just checking:
Are you aware that the camera has different wind suppression settings for internal and external microphones?
The microphone amplifier has an input resistance of about 2 kohm, the line input presumably has a higher resistance. Could that affect the DAC output? That is, is there a relatively small DC blocking capacitor in the DAC at the output?
Dear Marcel,
many thanks!
See "OUT_R.1 / OUT_R.3"
and "OUT_L.3 / OUT_L.5"
http://www.twistedpearaudio.com/docs/linestages/ivy3_schematic_1_0.pdf
BTW, as far as I remember Twistedpear uses the OPA1612 as well.
many thanks!
No, it has no different settings for each input. And wind filter is always off.
As fa as I see, no:
See "OUT_R.1 / OUT_R.3"
and "OUT_L.3 / OUT_L.5"
http://www.twistedpearaudio.com/docs/linestages/ivy3_schematic_1_0.pdf
BTW, as far as I remember Twistedpear uses the OPA1612 as well.
Hmm... because of the low supply voltage, (very likely only 5V) the the OPA1692 might be better suited?
https://www.ti.com/lit/ds/symlink/o...72532&ref_url=https%3A%2F%2Fwww.google.com%2F
https://www.ti.com/lit/ds/symlink/o...72532&ref_url=https%3A%2F%2Fwww.google.com%2F
No. It will always be the Sennheiser MKH-416.
Technical Data can be found here:
https://www.bhphotovideo.com/lit_files/74414.pdf
I would like to sqeeze the circuit into the Xvive, with a fixed resistor for about 45dB amplification. Circuit should not be bigger than an SD-Card, except for the 220uF Cap. Possible?
Technical Data can be found here:
https://www.bhphotovideo.com/lit_files/74414.pdf
I would like to sqeeze the circuit into the Xvive, with a fixed resistor for about 45dB amplification. Circuit should not be bigger than an SD-Card, except for the 220uF Cap. Possible?
I had downloaded a manual, one in English, and saw on page 295 in chapter 11:
"Reduction of Wind Noise
This reduces wind noise when an external microphone is connected.
Select [Wind Cut], settings [high]/[standard]/[low]/[off]"
In chapter 10 page 263, there is an audio setting [Wind Noise Canceller] described as "This reduces the wind noise coming into the built-in microphone while maintaining sound quality". Hence my remark about different settings for internal and external microphones.
OK, clear!
Don't worry. Rule #1 when filming professionally: Turn all little helpers off.
Even Autofocus. Fix it in post.
I had my camera modified to full sprectrum, so that you can also film in Infrared.
Every CMOS Sensor can do this, the Sensor of the S5, Sony IMX410, is also used in astro-cameras.
I asked the kind guy, Markus Meel from Astro-Modifikationen, to make a photo from the audio Input
jacks while disassembling the S5, as they do not have enough grip for the 3.5mm plugs, so that I might
replace them at a later point.
Over the "DV1634" marking might be the resistors/caps that could cause inaccuracies
when switching the preamp from mic to line? Don't know if the resistance of the cam's preamp / input IC rises
when switching to line?
Unfortunately the preamp IC is underneath the board, it feeds the "DAC3101", a TLV320DAC3101 from Texas Instruments
BTW, this is how the image looks like in Infrared 630nm,
first half of video "out of cam", second half blue and red channels are swapped:
About powering the planned preamp:
Those LiPo batteries protect themselves from deep discharge?
Voltage is 4V. Then a Mic Preamp could be attached there?
(2 Pads can be seen on the photo below)
I assume an attached preamp will not influence charging the Lipo battery
as 220µF (the virtual ground) would be between - and +?
But maybe the 48V could also be tapped?
Xvive claims to provide 12mA (@45.6V, not 48V by the way, measured without load)
https://xvive.com/audio/wp-content/uploads/sites/2/2022/01/Xvive_P1_MANUAL_EN_V1_2022-01-25.pdf
Sennheiser needs 2mA
https://www.bhphotovideo.com/lit_files/74414.pdf
Those LiPo batteries protect themselves from deep discharge?
Voltage is 4V. Then a Mic Preamp could be attached there?
(2 Pads can be seen on the photo below)
I assume an attached preamp will not influence charging the Lipo battery
as 220µF (the virtual ground) would be between - and +?
But maybe the 48V could also be tapped?
Xvive claims to provide 12mA (@45.6V, not 48V by the way, measured without load)
https://xvive.com/audio/wp-content/uploads/sites/2/2022/01/Xvive_P1_MANUAL_EN_V1_2022-01-25.pdf
Sennheiser needs 2mA
https://www.bhphotovideo.com/lit_files/74414.pdf
I would expect there to be an undervoltage lock-out circuit somewhere on the main PCB. Are there any ICs with readable type numbers that turn out to be step-up converters or battery management ICs when you enter the number in a search engine?
Dear Marcel,
dank u zeer!
Nope - only the microcontroller "ABOV8216" shows up.
Fron the test points, the most I get is 3.3V, except the lipo battery
that delivers 4v.
Assuming that a circuit can be safely run from the LiPo battery:
Could an LM358 based circuit be a solution?
(The gain factor for 45db is 180, correct?)
EDIT-Now I get it - you assume the undervoltage safety circuit is not on the battery.
Could be this K9Kj428 - thing, googling returns nothing useful
dank u zeer!
Nope - only the microcontroller "ABOV8216" shows up.
Fron the test points, the most I get is 3.3V, except the lipo battery
that delivers 4v.
Assuming that a circuit can be safely run from the LiPo battery:
Could an LM358 based circuit be a solution?
(The gain factor for 45db is 180, correct?)
EDIT-Now I get it - you assume the undervoltage safety circuit is not on the battery.
Could be this K9Kj428 - thing, googling returns nothing useful
Last edited:
An LM358 is about the worst op-amp there is for audio: class C output stage that generates lots of crossover distortion (although there are workarounds for that), relatively high voltage noise, low gain-bandwidth product.
I'll try to come up with something that can work on anything from 3 V to 5.5 V, doesn't take too much space and has low enough noise for your microphone (low enough not to degrade the noise floor by more than 1 dB(A)). An LME49721 should do fine.
I'll try to come up with something that can work on anything from 3 V to 5.5 V, doesn't take too much space and has low enough noise for your microphone (low enough not to degrade the noise floor by more than 1 dB(A)). An LME49721 should do fine.
You could try something like this - at your own risk, as it is neither tried nor tested nor simulated. It may need up to 10 seconds to settle after power-on. If that is annoying for your application, reduce the 94 uF capacitors in the phantom supply to 47 uF or 22 uF. The noise level would be rather high for dynamic microphones, but with the sensitive condenser microphone you use, it should contribute less than 1 dB(A) to the total noise.
I forgot the input protection network. R11 is only needed if there is a chance that the supply can't sink 12 mA for a short time when the phantom supply is turned on, otherwise it just wastes current.
Dear Marcel,
many, many thanks for your efforts!
As I am a layman, some questions about the revised version:
-Leakage current from the 94µ capacitors is sunk by R1 and R7?
-Output seems to be single-ended. Any disadvanteges from a balanced output
that could be made single-ended by the usual "Minus to GND"?
-How much is the output impedance?
Frankly, it will be tough to squeeze this design into the small box, as the battery uses the space.
There is only room left and right of the XLR connector. I will very likely have to put
the opamp with it'ts feedback resistor on one board and the input circuitry on a second board.
All the best,
Salar
many, many thanks for your efforts!
As I am a layman, some questions about the revised version:
-Leakage current from the 94µ capacitors is sunk by R1 and R7?
-Output seems to be single-ended. Any disadvanteges from a balanced output
that could be made single-ended by the usual "Minus to GND"?
-How much is the output impedance?
Frankly, it will be tough to squeeze this design into the small box, as the battery uses the space.
There is only room left and right of the XLR connector. I will very likely have to put
the opamp with it'ts feedback resistor on one board and the input circuitry on a second board.
All the best,
Salar