Hi,
I got confused about how to app. estimate the crosstalk between opamps in dual opamp package.
It kind of depends on implementation and pcb layout.
But f.ex. here, in the NE5532 specs:
- is that figure the maximum possible isolation in db? (110 db)
Regads
I got confused about how to app. estimate the crosstalk between opamps in dual opamp package.
It kind of depends on implementation and pcb layout.
But f.ex. here, in the NE5532 specs:
- is that figure the maximum possible isolation in db? (110 db)
Regads
My point is not to degrade another component's declared performance.
It is also a dual chip (vol control) with a stated crosstalk of 100 db.
Using both opamps in a dual package would save space.
It is also a dual chip (vol control) with a stated crosstalk of 100 db.
Using both opamps in a dual package would save space.
The opamp datasheet pictured in the first post shows typical crosstalk between the two opamps in a single package at 1kHz (typical, meaning the parameter is not guaranteed). Probably looks worse as frequency goes up.
practiacal is real electronics theory is not real . if you are happy with the performance enjoy the result
try available semiconductor if you are happy with the performance use it for your product .......
Trolling off topic posts have been removed, continued trolling will be rewarded with bin time. Stay on topic.
Yes, since the cause is capacitive coupling, the separation would degrade at 20dB/decade at high frequencies.
Last edited:
In practical terms crosstalk above 1kHz cannot ever be audibly detected if it's better than 50dB. Crosstalk in source material (music) doesn't even approach that. Acoustic recordings seldom hit 20dB. Electronically recorded music rarely presents hard-panned sources. In-room acoustic crosstalk negates all of this. It's only barely detectable in headphones.
100dB at 1Khz, reducing 20dB/decade would be considered audibly perfect in practical terms. If the project was measurement or instrumentation, then of course, no.
100dB at 1Khz, reducing 20dB/decade would be considered audibly perfect in practical terms. If the project was measurement or instrumentation, then of course, no.
Often crosstalk is "out of phase" and not uniform with frequency, which is not so benign.
Last edited:
Sure it is. It's the degree of crosstalk, not the phase. Actually, a little out of phase crosstalk tends to open up the soundstage, but it would have to be at a very high level to be audible, far, far more than -50dB. -100dB is completely inaudible in all conditions regardless of phase.
The audio signal in a channel is not always at full scale, but the crosstalk from the other channel could be.
This is not about a mono signal. Circuits without such quirks are quite possible, if that is the goal.
There are good reasons many like dual mono electronics.
Last edited:
First, the principles of masking. Very simplistically (and there is much more to this) if a sound is loud enough, you can't hear another that is quieter.
Second, as I detailed earlier, recorded material doesn't have channel separation anywhere near 100dB. Are you asking how I know that?
Principles of stereo mixing. Hard-panning (a signal only appears in one channel) isn't possible with acoustic stereo recording because you have a stereo pair in the acoustic space. Both mics get everything, but with spectral and timing differences. In electronic mixing you can pan fully to one channel, but when you play that recording on speakers, both ears hear both speakers all the time. You perceive stereo placement based on both inter-channel level difference and inter-channel time difference, but with speakers at 30-40 degrees apart, there is no audible improvement beyond about 35dB of separation because both ears hear both speakers.
In headphones you can discern separation a little better, but then masking takes over and you can't tell 40dB from 100dB. Try it for yourself with free software like Audacity.
All that plus 5 decades in pro audio.
Right in the absolute, wrong in the practical.
Of course they are possible. But my point is, in channel separation, it's unnecessary because of the recorded music we listen to.
I would say "many reasons", then leave "good" to the highly subjective. People think they hear a lot of things until presented with controlled testing, then suddenly, they can't.
100dB of separation is more than enough for any audio application except test and measurement.
The same is true for a phono cartridge, which struggles to beat the 20dB mark in practice and vinyl still sounds great.
An appeal to blind testing doesn't go anywhere with me.
And some of us have even more than 5 decades in audio engineering.
Just going back this...
Music doesn't have that kind of dynamic range anyway. If you had a 0dBFS signal (BTW, that would be peak, the actual "loudness" value will be 8-10dB lower) in one channel, you will not have dead silence in the other, you will have other music signals, likely down only about 35 dB or so from the other channel. They will cover the crosstalk at 50dB. With 0dBFS in one channel your ability to hear anything in the other is limited to about 40dB average, and is actually frequency dependent, and that only in non-musical test conditions.
Again, try it. Get Audacity, generate a reference level of pink noise in one channel then see what you can hear in the other with tones at reduced levels.
Go ahead, prove me wrong if you can. See if you can differentiate crosstalk better than 50dB.
Because its the difference between a jet engine taking off and a quiet wisper perhaps? A factor of 10000000000 is a large factor!
Odd that someone could have that kind of time in engineering but still reject double blind testing.