HunterZ wrote:

In any case, I am fairly confident that 0% line-in level is supposed to be functionally equivalent to mute.

You can easily test in Audacity whether the Line-In level control attenuates or boosts the signal.
A typical motherboard sound card should have noise floor of around -80dbFS with the line-in levels at zero.
Start with Levels slider at zero and slowly raise it till you see the noise floor volume change in the recording meters.
This point should be the center between cutting and boosting (unity) of the line-in, if it actually works as we assume.

In professional equipment like mixers, microphone preamps, and audio interfaces, the gain trims are only boosting (hence "gain") the inputs, so zero gain actually means maximum available headroom before clipping.
A hot input signal will have the benefit of lower recorded noise floor because we are not raising the input op-amp self noise by adding gain.

James-F wrote:

You can easily test in Audacity whether the Line-In level control attenuates or boosts the signal. A typical motherboard sound c […]
Show full quote

You can easily test in Audacity whether the Line-In level control attenuates or boosts the signal.
A typical motherboard sound card should have noise floor of around -80dbFS with the line-in levels at zero.
Start with Levels slider at zero and slowly raise it till you see the noise floor volume change in the recording meters.
This point should be the center between cutting and boosting (unity) of the line-in, if it actually works as we assume.

Here's what I see - I don't think it lines up with your expectations:
0%: No signal
1%: -129dB
2%: -126dB
3%: -123dB
4%: -120dB
5%: -117dB
6%: -115dB
7%: -114dB
8%: -112dB
9%: -111dB
10%: -109dB
15%: -103dB
20%: -99dB
30%: -93dB
40%: -88dB
50%: -84dB
75%: -77dB
90%: -75dB
100%: -70dB

Also, I can't hear my MT-32 at all at 100% synth and 100% PC speaker volume when line-in is at anything below 3%.

Hm... it appears that 50% is unity because -84db noise floor is typical for on-board sound cards.
Also from the last (Round 3) recordings I see proper waveform without distorted peaks, with 100% synth and lowered input levels (even though the MT-32 itself is distorting).

BTW, What is the noise floor in RMS (noise is measured only in RMS) of your line-in with the levels at max?
On my Focusrite audio interface with the preamp gains at zero (max voltage input) and line-in levels at 100, I get -102db (marked in green).

There is another test I want you to do (at your spare time), you need a 1/8" loop/jumper cable (male-male) for it.

1. Loop your PC output and Line-in input with this cable.
2. Play a 1kHz test tone at full volume from you favorite media player.
Here: test tone
3. Check your recording meters and adjust the line-in levels.

Does the rec meter peak with the max signal from the PC?
Do you need to lower the line-in levels down from 100 so they don't peak?

If the max signal from the PC doesn't reach the line-in peak, you can also boost this 1kHz test tone through the JVC and back into the line-in and check again.
What I want to confirm is whether the line-in can boost and cut to accept a wide range of signals.

I get about -85dB RMS at 100%.

The loopback test you describe will have to wait until tomorrow. Is there any danger in doing this? I don't want to fry my laptop.

Boosting via the JVC (should it prove necessary) would probably require using the headphone out, because as I said the recording out has no level adjustment. I should be able to manage all this with cables that I have on hand.

I should mention that the sound chip is a Creative Sound Blaster Recon3Di in a Dell Alienware 17 R3 laptop.

HunterZ wrote:

I get about -85dB RMS at 100%.

Thanks. Spot on with my previous knowledge.

HunterZ wrote:

Is there any danger in doing this? I don't want to fry my laptop.

🤣 🤣 🤣
No, it's a standard procedure to measure Frequency Response, Latency, Distortion and other parameters of a sound cards or audio interfaces.
You have to make sure that the "Listen" option is UN-ticked (default) or else you will get endless feedback, if it is ticked you will hear the signal, so you can play with your CM-64 and DOSBox for example.
Besides, the Line-Out should be much less-Hot than the MT-32, moreover, the Line-In has protection diodes on the input so it's not actually clipping at the op-amp rails.
Worry not.

@Falcosoft
Please pardon the non FSMP related talk.

Loopback test results: It looks like the Line-In peaks if either of the following are in effect:

• Line-In level at 83+.
• Headphone Out level at 69+.

Test was performed using the 30-second 1kHz test tone WAV and VLC with the following configuration:

• VLC mixer volume at 100%
• No audio filters enabled
• Peak protection disabled

More data points:

• Playing the test tone through the loopback cable with Line-In at 50% (which is best guess for unity) and Headphone Out at 100%, I get a level of about -8.5dB and an RMS value of slightly below -11dB.
• Monitoring the "What U Hear" recording input of the sound card (which is like an internal loopback that appears to be unaffected by the speaker/headphone output volume level), I get a peak level of about -3.5dB at What U Hear at 100% and about -14dB at What U Hear at 50%.

Audacity says the raw test tone data is -3dB. If I try to hit that target instead of 0dB, I get these values:

• Headphone Out @ 100%, Line-In at 71%, OR
• Headphone Out @ 57%, Line-In at 100%

If 50% is unity on Line-In, then I guess this means that the Headphone Out of the PC is attenuating the signal a bit? I always thought that headphone amplifiers would put out a hot signal compared to line-level when at 100% volume? Would connecting headphones directly to a line-out really overdrive them in the typical case?

After all the work with the JVC, I found a passive 4-input S-Video/composite/RCA A/V switchbox in the garage. Doh! At least it will be more convenient for switching synthesizers because I will be able to keep something that small closer to where I actually sit to use my laptop. Now I just need an extra MIDI male-male cable and a pair of mono phono-RCA cables/adapters so that I can have everything connected at once.

Shame it didn't have AT or PS/2 keyboard switching as well! You'd have a complete solution!

gdjacobs wrote:

Shame it didn't have AT or PS/2 keyboard switching as well! You'd have a complete solution!

Yeah. If I was more of a hardware modder I might re-purpose the internal composite+S-video lines for 5-pin MIDI DIN.

Mentioned this project to my wife, and she said she would expect unity to be at a line-in level of 80-something (which may correspond to the 83 I found in my initial loopback test?) rather than 50 or 100. She also confirmed that line level signals are usually hotter than headphone signals.

Thanks for testing HunterZ.

I believe there is no such thing as 'unity' for the recording line-in, there is maximum voltage it can receive without clipping.
Line-Out will (should) output consumer level voltage of 0.316v AC RMS at maximum.
https://en.wikipedia.org/wiki/Line_level

Do you have a digital voltage meter DVM (every man should have one)?
If so please measure the AC voltage output on the tip and sleeve of the 1/8" plug when playing a normazlied 100Hz test tone from your PC at full volume.
You can import (simply drag into audacity) the 100Hz test tone into Audacity and Normalize it (Effect-Amplify), and play from there.

I assume you mean tip (left) versus sleeve (ground), as tip (left) versus ring (right) would be 0VAC when playing a monophonic recording.

Unfortunately the AC voltage mode of my digital multimeter has a resolution of only 0.1VAC. It shows 0.7VAC with headphone out volume at 100% while the normalized 100Hz tone is playing in Audacity.

Yeah ring-sleeve sorry.
0.7v seems to be twice the voltage of standard line-level output, but it is not unusual, most are higher; I assume the standard is the required minimum.

So the output voltage is not enough to test the line-in maximum voltage input, we need hotter undistorted sine wave at the input.
Do you have a way to boost the line-out and send it back to the line-in?

I don't own a signal generator or mic amp or anything, but I probably have the components to breadboard something. I'll take a look around the garage tomorrow if I have time.

What is the expected range between where line-in simply clips and where my PC starts frying?

Edit: Actually, how much amplification would the phono pre-amp circuit on the JVC perform?

So I found a bunch of LM324 and a few LM386 ICs in the garage, but this is really too big a project to try to do in one day, so I'm not sure when I'd actually get around to it (plus I don't currently have a good work space for this kind of thing).

HunterZ wrote:

What is the expected range between where line-in simply clips and where my PC starts frying?

Electrostatic voltage is many thousands of volts... its the current that fries electronic not voltage.
I think below 50v generated signal you'll be safe.

Edit: Actually, how much amplification would the phono pre-amp circuit on the JVC perform?

I think it'll do.

James-F wrote:

Electrostatic voltage is many thousands of volts... its the current that fries electronic not voltage.
I think below 50v generated signal you'll be safe.

ESD events are different. They last only about 20 microseconds so they are short and ESD protection devices can clamp 8kV voltages down to safe levels by shorting the overvoltage to ground. The shorting current can be in the order of 5A and for a brief moment the ESD protection device can dissipate over 100 Watts.

But it would be insane to feed 50V into line input. That should fry any protection and op-amps.

For instance Asus Xonar DG soundcard says full scale input and output voltages are 1Vrms or 3Vpp.
Their USB sound card can have levels of 2Vrms or 5.65Vpp.