What causes a DAC to have bad directional imaging?

[Sokel]

How to do a null test in REW ... I am embarrassed to say that I do not know how. I know how to do it in Acourate (which is what I normally use). In REW I thought you do the inversion with Trace-Arithmetic 1/A, and then the convolution with Trace-Arithmetic A*B. I tried it with some spare measurements that I have. But for some reason it gives me a different result to Acourate. In Acourate, it nulls out. In REW, it doesn't. I'll have to ask a REW expert for help on this, very sorry.

In any case, the cable connection you made is fine.

The THD+N measurements are not great. We don't know if it's the output or the input since it was measured with a loopback cable, but it should still be below audible limits. I am starting to think that we won't be able to get to the bottom of why you are hearing a difference with this testing. You might have to send the device to Amir for testing.

You have to consider levels and the distortion structure looking at it.
I guess that if the mic input could stand higher levels (the missing 15-20dB or so) the limiting factor would only be noise,so a -90dB THD+N seems perfectly plausible.
We also have to consider impedance mismatching I suspect, interface's input are usually low.

 

[Keith_W]

You have to consider levels and the distortion structure looking at it.
I guess that if the mic input could stand higher levels (the missing 15-20dB or so) the limiting factor would only be noise,so a -90dB THD+N seems perfectly plausible.
We also have to consider impedance mismatching I suspect, interface's input are usually low.

I am out of ideas, so i'll have to wave the white flag. I am sorry. I don't know how he is going to measure those other things you are suggesting without special test equipment. I tried to keep it within possibility of what he already owns. So for now, it's going to be either "accept what your ears are telling you" or "send the equipment to someone who knows how to test them".

 

[Sokel]

I am out of ideas, so i'll have to wave the white flag. I am sorry. I don't know how he is going to measure those other things you are suggesting without special test equipment. I tried to keep it within possibility of what he already owns. So for now, it's going to be either "accept what your ears are telling you" or "send the equipment to someone who knows how to test them".

I don't think he has to measure, that's what I'm saying.
Connecting the dots seems there's no problem to it other than the limiting noise (-95dB ballpark,good enough already in my book), probably inaudible as the test fixture can add some (a lot) too or even digital attenuation if he used such.
So...

 

[Keith_W]

I don't think he has to measure, that's what I'm saying.
Connecting the dots seems there's no problem to it other than the limiting noise (-95dB ballpark,good enough already in my book), probably inaudible as the test fixture can add some (a lot) too or even digital attenuation if he used such.
So...

I would love for the OP's issue to be resolved. It seems so ... unsatisfying. If you can help him, I am sure he would be very grateful. And i'll be watching to see what I can learn

 

[Sokel]

I would love for the OP's issue to be resolved. It seems so ... unsatisfying. If you can help him, I am sure he would be very grateful. And i'll be watching to see what I can learn

I would love so too, only I don't see an issue electrically as he measured it.
Maybe real world use changes something, foreign to the loopback? I can't tell.

 

[RandomEar]

Maybe this is much simpler: @toasty what headphones do you use to listen to the DACs? According to the Bridge Cast specs, the output impedance of the phone/headset output is 10 Ohms. That's a lot and will change the frequency response of low impedance headphones and IEMs.

The phone/headset connector is also a combined TRS/TRRS out. I am not sure if this alters the sound if you use a headphone without a mic and don't use a TRRS to TRS adapter for connecting it. It shouldn't, because that's a common use case, but who knows.

 

[toasty]

Maybe this is much simpler: @toasty what headphones do you use to listen to the DACs? According to the Bridge Cast specs, the output impedance of the phone/headset output is 10 Ohms. That's a lot and will change the frequency response of low impedance headphones and IEMs.

The phone/headset connector is also a combined TRS/TRRS out. I am not sure if this alters the sound if you use a headphone without a mic and don't use a TRRS to TRS adapter for connecting it. It shouldn't, because that's a common use case, but who knows.

My apologies for not mentioning what headphones I was using in the original post, should have done that. 99% of the time, I use the Beyerdynamic DT 900 Pro X.

I did notice that with the Bridge Cast that it seems to add a slight V-Curve to the FR of whatever headphones / iems I use vs when I tested with the SMSL C200. So I do think that the output impedance is affecting the overall sound. But it shouldn't affect the performance of the device, right?

I would love so too, only I don't see an issue electrically as he measured it.
Maybe real world use changes something, foreign to the loopback? I can't tell.

I was doing some critical listening tests last night with all external noise sources minimized and I could actually hear some white noise coming from the device. It was way more audible with IEMs (Tripowin Vivace, KZ EDC Pro, Tangzu Wan'er - all cheap IEMs as I barely use them).

It was originally volume dependent, but I noticed that Roland has an article saying that if nothing is plugged in to the Aux-in port, that some noise is audible & to mute that channel when it's not in use. Muting it dropped the noise by a decent amount... But when I kept listening, I could still faintly hear it. However, now it was not volume dependent, it stayed the same no matter if I maxed out the volume or muted everything. The only time it stopped was when I unplugged my headphones/iems from the device.

 

[kemmler3D]

what measurement metric would I be looking at to determine the DAC's ability to have normal good stereo imaging?

You'll want to look at distortion, particularly high frequency (THD and IMD), crosstalk (less is better), frequency response (look for any deviation in the treble) and MAYBE phase distortion. Looks like you already got some of those measurements done. Apparently this device has a significantly non-flat frequency response so I'd say that's your likely answer... except @RandomEar 's point about high output impedance might make more sense (due to causing more FR changes.)

If you are using IEMs with this device, that almost feels like a smoking gun to me.

I do think that the output impedance is affecting the overall sound. But it shouldn't affect the performance of the device, right?

In a real sense this amounts to low performance, depending on what device you're using with it, because the impedance mismatch causes the frequency response to deviate from what it's supposed to be. That can easily screw up a stereo image... 1-2dB here and there is all it takes. Audible noise would also impair stereo image by hiding low-level details.

 

[Chrispy]

tldr but how can a dac affect "directional imaging" whatever that is?

 

[RandomEar]

Yeah, the 900 Pro X is 48 Ohms. That's too low for 10 Ohms output impedance. You want 100 Ohms or more on the headphone with such an output. At least that's what the rule of thumb says.