Measured JBL LSR310S output signal distortion with MOTU M4

[txbdan]

Hi all,

Preface: Take this all with a grain of salt. I don't know what I'm doing and I'm sure there will be things to tweak and improve with my measurement.

I have a JBL LSR310S along with a pair of Genelec 8030Cs. I've been curious about what the JBL LSR310S subwoofer is doing to my signal quality as I run my DAC to the sub and then sub's HPF outputs to my Genelecs. I have a MOTO M4 that I'll attempt to measure with. Of course it isn't perfect, but hopefully I can at least set a baseline looping the interface and then compare against running through the sub.

Baseline, connecting M4 output to input with balanced TRS cable:
Just by looking at the plot, I see a -115dB distortion (I don't understand why this isn't somewhere in the distortion measurement box)

Now for the sub. The first thing I realize is that with the sub's cross over bypassed, it doesn't output on the outputs at all and it plays full range through the subwoofer. So in order to get output on the output connectors, I have to engage the crossover. I'm using a 1khz test tone, so that should be ok and is more representative of the real configuration anyway. I then notice that the output signal level is quite low. On the M4's meters, I can see my -3dBFS output using most of the meter, but the return level is maybe a third of it. This is because the JBL seems to output "-10dBv" levels. This jives with the manual which says that you should set the monitors to "-10dBv" mode while using the sub. It's annoying that the sub can accept +4dBu, but can't output at that level.

But that's what we have and here is the result:
I see around -100dB distortion

Questions:
The distortion properties (in the upper left hand box) are correct and don't require my signal to be exactly at 0dBFS, right?
Is this apples to apples if the JBL case is outputting a much lower (-10dBv) signal?
I understand THD and THD+N and it seems the sub does add a decent bit of distortion and noise... right?
I must be misunderstanding N and N+D as those values don't jive with what I see in the plot for distortion. They are also inverse of the THD numbers.

What to make of this? Thanks

 

[Blumlein 88]

Are you thinking why is the noise floor so much lower than harmonics and lower than noise levels in the distortion read out panel? If so that is due to FFT gain as applies to noise.

 

[McFly]

My take: you have nothing to worry about. That output is clean.

Now the measurement bug has you.

 

[LTig]

I have a JBL LSR310S along with a pair of Genelec 8030Cs. I've been curious about what the JBL LSR310S subwoofer is doing to my signal quality as I run my DAC to the sub and then sub's HPF outputs to my Genelecs. I have a MOTO M4 that I'll attempt to measure with. Of course it isn't perfect, but hopefully I can at least set a baseline looping the interface and then compare against running through the sub.

Baseline, connecting M4 output to input with balanced TRS cable:
Just by looking at the plot, I see a -115dB distortion (I don't understand why this isn't somewhere in the distortion measurement box)

It's the 3rd harmonic, shown as 0.00022%. You can configure REW such that it displays dB values instead of %

Now for the sub. The first thing I realize is that with the sub's cross over bypassed, it doesn't output on the outputs at all and it plays full range through the subwoofer. So in order to get output on the output connectors, I have to engage the crossover. I'm using a 1khz test tone, so that should be ok and is more representative of the real configuration anyway. I then notice that the output signal level is quite low. On the M4's meters, I can see my -3dBFS output using most of the meter, but the return level is maybe a third of it. This is because the JBL seems to output "-10dBv" levels. This jives with the manual which says that you should set the monitors to "-10dBv" mode while using the sub. It's annoying that the sub can accept +4dBu, but can't output at that level.

Well it's according to the manual. My Genelec sub (7060B) outputs what's fed into, and you can disable the sub with a remote switch so the full signal is passed through to the sats. But you pay a lot of money for that.

But that's what we have and here is the result:
I see around -100dB distortion


Questions:
The distortion properties (in the upper left hand box) are correct and don't require my signal to be exactly at 0dBFS, right?

Yes. EDIT: I'm not sure now

Is this apples to apples if the JBL case is outputting a much lower (-10dBv) signal?

No. The lower output level just means that its prone to contain more noise.

I understand THD and THD+N and it seems the sub does add a decent bit of distortion and noise... right?

IMV noise and distortion are very low. I didn't expect this for such a cheap product, especially where the corresponding sats are known to hiss.

I must be misunderstanding N and N+D as those values don't jive with what I see in the plot for distortion. They are also inverse of the THD numbers.

What to make of this? Thanks

• N: the noise floor cannot be used to determine noise level directly, you have to add the FFT gain which depends on the number of bins: FFT gain = 10 x log(FFT size / 2). With 256K FFT size the FFT gain is 10 x log(256*1024/2) = 51 dB, meaning the real noise level is 51 dB higher than the noise floor.
• N+D: this is the sum of the noise level plus the level of all harmonics (2, 3, 4, ...) up to the ninth.
• Both values (N, N+D) are levels in dB hence negative as they are below the reference of 0 dB.
• SINAD is Signal / (N+D) which is the distance between the peak and the N+D level. SInce the signal has 0dB level SINAD = 0dB - (N+D) = -(N+D) .

EDIT: see this web page for online calculations.

 

[txbdan]

Hrmm, so SINAD (what I'm used to looking at in Amir's DACs/amps measurements, is -(N+D). So my loopback cable has a SINAD of 109.7 and the sub has a SINAD of 112.8? I think the sub must have a differential receiver, do the filtering, then have a differential driver for the balanced signal. Maybe that in a sense starts the signal from a clean slate? I would think not though, as anything coming in, noise or not, should be passed on through.

A: loop back cable
SINAD of M4 DAC + SINAD of M4 ADC

B: sub
SINAD of sub's output circuit + SINAD of M4 ADC? This doesn't seem quite right as the input signal and its noise must pass through to some degree
SINAD of M4 DAC + SINAD of sub's circuit + SINAD of M4 ADC? if so, how can it be lower than just a cable?

So the sub's output has a lower noise floor than the M4s?

But at the same time the sub's total THD is 3x the loopback cable... I think this is because THD considers all of the harmonic spikes.. so while they may be lower, maybe there's more of them? but that doesn't seem to jive as the second plot "looks" like it has fewer and lower harmonics.

 

[LTig]

Hrmm, so SINAD (what I'm used to looking at in Amir's DACs/amps measurements, is -(N+D). So my loopback cable has a SINAD of 109.7 and the sub has a SINAD of 112.8? I think the sub must have a differential receiver, do the filtering, then have a differential driver for the balanced signal. Maybe that in a sense starts the signal from a clean slate? I would think not though, as anything coming in, noise or not, should be passed on through.

A: loop back cable
SINAD of M4 DAC + SINAD of M4 ADC

B: sub
SINAD of sub's output circuit + SINAD of M4 ADC? This doesn't seem quite right as the input signal and its noise must pass through to some degree
SINAD of M4 DAC + SINAD of sub's circuit + SINAD of M4 ADC? if so, how can it be lower than just a cable?

So the sub's output has a lower noise floor than the M4s?

But at the same time the sub's total THD is 3x the loopback cable... I think this is because THD considers all of the harmonic spikes.. so while they may be lower, maybe there's more of them? but that doesn't seem to jive as the second plot "looks" like it has fewer and lower harmonics.

I'm no longer sure whether the THD and Noise values displayed are independent of the peak level. Loopback must be better than through the sub. And I think it is if you check the peak in both measurements. Loop back: -5 dB, sub: -18 dB. So all values for the sub should be 13 dB worse.

 

[mdsimon2]

Anything shown as a percent (THD, THD+N, individual harmonics) is relative to your fundamental. As mentioned previously you can change the settings to show these values in dB as well. Anything shown as dBFS is relative to full scale input of the ADC and anything shown as dBFS A is relative to full scale input of the ADC but A weighted.

THD+N is equivalent to SINAD, so on the loopback you are looking at 20 x log (0.00071/100) = -103 dB and with the LSR310S you have 20 x log (0.002 / 100) = -94 dB. Looks like you are getting an increase in distortion (-110 dB THD to -99 dB THD) but considering the 12.5 dB level reduction in the sub output the noise and THD+N/SINAD are really quite good. Although noise performance with the LS310S is actually a bit worse relative to your fundamental (I estimate it at -96 dB unweighted compared to -104 dB unweighted for the loopback) as you note it is at a lower absolute level implying that analog attenuation is being used.

Michael

 

[txbdan]

So, because the output of the sub is a lower signal level, I have to make up for that with the Genelec's input gain/sensitivity adjustment to get the same volume. Assuming that gain is "perfect", is that why you can add the 12.5dB to the sub's measurements to get an equivalent distortion measurement "at the speaker" so to speak? I want to boil this down to the actual difference in distortion/noise coming at the speaker.

Part of what's giving me heartburn is that I just got a RME ADI-2 DAC FS and if I'm running it through the sub maybe its all for moot signalwise. Granted the M4 is already clean beyond hearing, but it does feel good to know that things are optimal. There are other features on the RME so its not all for not, but it does bum me out a little. hah

 

[txbdan]

@amirm, what are your thoughts on this?

 

[howard416]

So, because the output of the sub is a lower signal level, I have to make up for that with the Genelec's input gain/sensitivity adjustment to get the same volume. Assuming that gain is "perfect", is that why you can add the 12.5dB to the sub's measurements to get an equivalent distortion measurement "at the speaker" so to speak? I want to boil this down to the actual difference in distortion/noise coming at the speaker.

Part of what's giving me heartburn is that I just got a RME ADI-2 DAC FS and if I'm running it through the sub maybe its all for moot signalwise. Granted the M4 is already clean beyond hearing, but it does feel good to know that things are optimal. There are other features on the RME so its not all for not, but it does bum me out a little. hah

hey, off-topic, but any chance you could confirm what kind of crossover this sub uses? 2nd order or maybe 4th order electrical?