Ultimate 4 channel audio recorder

There were a few rare recordings left in the apartment, so it was time to clean the 30 year old beast & fire it up.

Transferring tape to the Zoom replacement was an eye opener because the 30 year old tape deck's VU meter only showed but a tiny portion of the modern ADC's dynamic range.    0dB on the computer is around +6 on the tape deck.

Silence on a 30 year old type 1 tape with no noise reduction was around -40dB on the recorder but completely off scale on the tape deck's VU meter.  The unused AUX channels show how much farther the ADCs can go.  The self noise of the tape deck came in at -70dB.  Lions never had a way to fully measure the tape deck's noise until now.  The minidisc & Zoom H2 only went to -48dB.  

The original idea was for the Zoom replacement to dedicate 2 ADC's to each channel, but the dynamic range has been good enough to just let it record silence on 2 ADC's & record stereo on the other 2.  The mane problem is it uses twice as much storage as necessary.

The 30 year old tape deck had a cracked gear for the last 20 years. It only became obvious when the gear was turned just right. The lion kingdom also noticed the reverse capstan's pully grabs more of the belt than the forward capstan, hence why reverse always had less flutter than forward. After using the modern ADC, it's kind of amazing such a sloppy belt, pullys, & magnetic ribbon can make useful sound at all.

After a few tapes, the RS-BR465 finally died.  

Repaired the broken gear with a hot glue strap. It'll last long enough for the current batch of transfers but need a more permanent teflon strap. Also cleaned the capstan rollers & lubricated the plastic parts. Noticed the reverse capstan roller had much more aggressive ridges. There is now a tool fastened to the case which is required for replacing the belt.

There was an attempt to document more of the connections.

The 4 channel finally got the wifi power switch of the vlogging mic.  Connect GPIO 18 to UART TX to put it in low power mode so the wifi doesn't interfere as much with single ended signals.

Finally built the line input box for the recorder.  It was just 10uF 50V blocking caps connected in series with the line level.  The computer output was referenced to ground & the ADC input was referenced to 2.5V.  The computer output the 5V peak to peak voltage required to reach the recorder's full dynamic range.  

The recorder simultaneously handled the balanced microphone input & mixed the 2 inputs into the headphones, as designed. 

This feature was completely forgotten.  The original idea was to record I2S directly from the computer's optical output, to give purely digital copies of whatever it was playing.  The problem was the sample rate of the I2C & the microphone inputs would never match & the I2S signal would stop when the computer stopped playback or changed sample rates. 

A better way to record computer audio was believed to be pulseaudio, jack, Esound, or ALSA, but no sound driver was reliable enough to stay synchronized with the microphone track.  The final solution was recording the analog line output with the same clock signal as the microphone, as was already implemented.

The audio recorder still doesn't have the jumper for wifi adjustment that the vlogging microphone had.  The wifi has to be manually increased to 30dB in order to transfer a file & 5dB in order to record.

This gadget hasn't gotten the use hoped for.  The teenage lion of many decades ago dreamed of having an open reel tape teck & sounding like Van Cliburn.   The closest he ever got was the mighty Technics RS-BR465, but that Yamaha C3 still sounded terrible.  The biggest leap toward the Van Cliburn sound came from getting better piano teachers, years of practice, better instruments, & digital editing.

Now the lion kingdom made better gear than anything that existed decades ago.  Given what this circuit revealed about the noise introduced by just millimeters of wire & any single ended analog signal, any tape head or stylus is a gigantic radio receiver.  There's no way any open reel tape deck or record player could ever compete with even a 16 bit digital system at 48khz.  They can only color the sound & introduce a musical form of noise.

A good explanation of the noise encountered in this circuit.  Noise in very low frequencies grows because of 1/f noise rather than the nV/root Hz rating.  1/f noise is the reason you need blocking caps.  The nV/root HZ rating of an op amp is actually for a very small bandwidth.  It actually compounds as the bandwidth increases.  His explanation of RMS noise is a bit easier than TI's.

After upgrading the preamp, decided to make the VU meters show differential output instead of single ended levels.  Single ended  noise was too high for the meters to be any use, but

The differential output was in the weeds with the preamp off.

Maximum preamp gain with no microphone was still around -79db.  The gain is only 12x.

The aux channel has pops.

Single ended preamp with maximum gain was a bit noisier.  For each mixdown mode, the VU meters have the relevant titles & the relevant number of meters.

The room noise with the microphones connected & maximum gain, so obsessing over -79db or -60db is about as pointless as any game.  

When lions 1st saw digital audio on a computer 25 years ago, -40db was the lowest useful level for any VU meter.  All the consumer gear stopped at -40db, even after minidiscs came out.   -90db was a ludicrous option on DAWs for many years.  Who knew it would be necessary for testing home made gear.

Plotted some waveforms comparing the noise of various configurations in differential mode.  All were magnified to 1/262144 per Y pixel.  With all the junk beyond what was connected to the Zoom, it still outperformed the Zoom.  The Zoom doesn't have a gain control or phantom power, but does have some passives, so it's like the AK4524 with no preamp.

Disconnected successively more components, starting with the microphones & ending with just the op-amps to isolate the noise sources.  

The noise is over the air, not over the power rails.

There is 60Hz line noise from the bench & 10Hz wifi noise from the pi.

In differential mode, turning off wifi offers no improvement beyond using txpower 5.  In single ended mode, turning it off knocks out significant noise.

Powering off a battery, LM317, 78L05, RC filter made no difference.  

Unplugging the phantom power circuit from the op-amp causes the largest reduction in noise.  

Moving the PI farther away from the preamp, reducing txpower to 5, & using differential mode are going to be the standard use case & the keys to acceptable noise levels.  This gives the following waveforms at 1/262144 per Y:

With these measures, the system is at least as good as the Zoom was in its best days.  It's important to remember no other audio interface in the world uses wifi.  Though the noise is horrendous in single ended mode, it's acceptable because the preamp is the only subject of concern, it's the same circuit that was used with the Zoom, & it's never going to be used in single ended mode.

Would only complain that it's nowhere close to the ASMR videos of others & the headroom is reduced when combining the differential signals digitally.  You wouldn't notice the headroom except when listening to fully saturated levels.  The next steps would involve a new phantom power circuit, shrinking the wires in the preamp, using digital pots, disabling wifi, moving the entire ADC chain inside the microphone, optical outputs, adding pilot lights to the JFETs.

Kept what was there from 7 years ago.  Made the amplifier section output differential signals.  Of course, there are sections where the pairs aren't close together & this is where the RF gets in.  At least it worked as designed.  Considering the preamp, a Zoom H6 would be smaller & easier to use.

A few more software bug fixes & it finally captured microphones with the monitoring working perfectly in both channels.  Another change made the MANE volume control all the mane/I2S3 levels & the AUX volume control all the aux/I2S2 levels.  It now requires both faders to adjust a stereo volume, but it's more intuitive.  Also, the 2 channel mode using ADCs as differential pairs (L+/-  R+/-) now has to write a separate file for each channel.  A bit more effort could make every mode write a single file, but it's hardly ever going to be used.

Like the Zoom H2, the ADC board has no input protection.  To protect it from 48V pops, the preamp has spudger diodes.  The TLV320AIC32 in the Zoom simply blew up when plugged into phantom power.

The output of the ADC is a bit quieter than it was on the Zoom, because of the larger voltage range.  The preamp already does 100x.  The ADC can provide additional gain, but it has to be a GUI option.  

The real problems show up in the L/L  R/R mode.  Here, the differential pairs subtract away the microphone signal & sum the RF.  The RF is highly suppressed in the +/- modes, but still there.  The RF getting into the preamp is awful.  It's probably why others put thousands of parts into their preamps & pilot lights in their JFETs.

The mane feature of the GUI is the setting for mixing the 4 ADC's.  Mixing down before writing was deemed necessary, since the storage goes from 10 hours for a single channel to 2 hours for 4 channels & multiple files with differential pairs are a pain.  The mixing for storage is a completely different set of equations from the mixing for monitoring.  

The VU meters still show the independent levels of the 4 ADCs.  The top 2 are the AUX I2S & the bottom 2 are the mane I2S.  The board layout led to the mane I2S always being the higher number on every list.

The 2 monitoring volumes are applied differently, depending on the mixing formula.  The easiest way to figure out how they're mixed is by listening & tweeking the faders.

The mane trend is the use of the ADCs in the AK4524 as either a differential pair, an averaged pair, or 2 discrete channels.  It should be obvious from the GUI.  Each I2S needs to write to its own file because of synchronization problems.  The Zoom H2 had the same problem.

After a heroic programming session, the file writer was upgraded to a 16 second buffer which writes as soon as it has data instead of in 1 second chunks.  The SPI buffer was increased to 1/18 second, which was the maximum the STM32 could easily fit.  More draconian measures could get it to 1/11.  Unfortunately, its core coupled memory isn't accessible to DMA, which would have doubled the SPI buffer. 

With all that going for it, it still overran once when recording 4 channels for 40 minutes.  Don't expect to ever record 4 channels, except for testing the other modes.  The next step would be recording to USB storage or moving the whole PI to a faster SD card.

It still only does 48khz & only uses the AK4524 as the source for I2S2.  Have never ever used 96khz, despite it being on the Zoom H2.  The biggest value was 24 bits.  The problem is moving the ultimate stereo amplifier to a commercial 96khz DAC would give up the external tap for I2S.  It would be a matter of swapping the TOSLINK between 1 board purely for capturing it & a board for listening to it or somehow splitting the TOSLINK.  A USB soundcard would be the hardest to sniff.

The next matter is the preamp.  

The thought occurred of dumping 48V phantom power & replacing the microphone electronics with op-amps boosting directly to line level.  These mics are virtually free, have been replaced by USB versions & it would reduce the noise.  It would require exposing some kind of gain adjustment, but the metal enclosure is absolutely required to stop interference.  It's not for stopping singers from accessing the bass rolloff switch.  A preamp conversion would be a lot of work splitting mane hairs for a truly meaningless difference.