About digital amp latency

[jamius19]

Hi everyone,
I'm an audio gear newbie who's looking into passive speakers to connect to my Focusrite Scarlett 2i2 (3rd Gen) to minimize latency for live monitoring.
For amplification I was looking into the Fosi Audio TB10D. It uses the TPA3255 chip and is a digital amp.

It's my understanding that once audio leaves my interface, the only way latency can be added is via this amp as passive speakers should not add any latency.
Now my question is if I use this amp with passive speakers, how much latency can I expect from this kind of digital amplifiers (and not from other factors like PC/Interface)?

Thanks for any help!

 

[RayDunzl]

Grok says:

 

[voodooless]

TPA3255 chip and is a digital amp

No, it’s not a digital amp. The D in Class D does not stand for “Digital”!

 

[jamius19]

Grok says:

View attachment 485052

Since there are Bass and Treble adjustment knobs out in front, I'm a bit worried about possible DSP implementation.

 

[jamius19]

No, it’s not a digital amp. The D in Class D does not stand for “Digital”!

Yeah, that was my understanding as well.
I contacted Fosi Audio and they said it's a digital amp.

So now I'm not sure where does the digital part come from.
Any DSP perhaps for the Bass/Treble adjustment?

 

[voodooless]

I contacted Fosi Audio and they said it's a digital amp

They are wrong! @Fosi Audio , please educate your support staff!

Treble and bass are done with an analog chip as well. No DSP onboard.

 

[kemmler3D]

Welcome to ASR! As others have said it's a common misconception that the "D" stands for digital. In reality there is no ADC / DAC implementation in this amp or most other class D amps. Tone controls are not complicated to implement with analog circuits.

if I use this amp with passive speakers, how much latency can I expect from this kind of digital amplifiers (and not from other factors like PC/Interface)?

Effectively none, 10 microseconds is shorter than we can perceive as actual latency / delay.

A very low latency audio interface introduces 100-500x (1-5ms) more delay than this amp.

 

[lucian]

Welcome to ASR! As others have said it's a common misconception that the "D" stands for digital. In reality there is no ADC / DAC implementation in this amp or most other class D amps. Tone controls are not complicated to implement with analog circuits.



Effectively none, 10 microseconds is shorter than we can perceive as actual latency / delay.

A very low latency audio interface introduces 100-500x (1-5ms) more delay than this amp.

mh - its works internally with a PWM signal, which is by definition digital. here is also a conversion from analog input to PWM and (by filtering) back to analog output.

 

[DVDdoug]

mh - its works internally with a PWM signal, which is by definition digital.

PWM is quantized in time, but not in amplitude. There is no bit depth and there are no digital samples or digital data.

For DSP you need "normal" digital data in bits and bytes.

 

[LuckyNat]

mh - its works internally with a PWM signal, which is by definition digital. here is also a conversion from analog input to PWM and (by filtering) back to analog output.

You should tell that to all the analogue synths of the 70s and 80s with PWM oscillators .. Pulse Width Modulation is a type of square wave signal. Square wave signals are entirely analogue and are generated by flipping between two voltages.

You can also represent binary by choosing any two arbitrary indicators to respectively symbolise a 1 or 0 - choose two semaphore flag positions and send a digital signal by hand, flash a bulb to a light detector with on=1 etc .. or perhaps 1 mV can represent a 1 and 0 mV can represent a 0.

All of the above can be used to transmit a digital signal but it does not therefore mean semaphore, turning a torch on and off, or square wave signals are digital things !

Anyway, here is a diagram of how the *analogue* PWM signal is generated from the interaction of an analogue oscillator and the analogue input signal.

(from https://www.analog.com/en/resources/technical-articles/fundamentals-of-class-d-amplifiers.html) .

An interesting part of such an amplifier is that, yes, you can in theory also generate the PWM signal directly from digital data if one should fancy it ? There are technical issues around that from PCM audio (can't tell you what they are, but I have read discussions on DIYaudio.com) that are tricky to solve .. so it's a rare thing to come across.

Anyway, just for the fun of it .. here is the sound of an analogue synth playing it's analogue PWM oscillators:

 

[Chrispy]

Digital amp?

 

[jamius19]

They are wrong! @Fosi Audio , please educate your support staff!

Treble and bass are done with an analog chip as well. No DSP onboard.

Thanks for the clarification!
I asked them to clarify about the "digital" bit, but haven't got any replies.

 

[jamius19]

Effectively none, 10 microseconds is shorter than we can perceive as actual latency / delay.

Yeah, that was my only concern. I also assumed it'd be imperceptible, but then I got confused when they mentioned it was digital.

I own a pair of Edifier S3000 Pro Bookshelf Speaker. While they're good enough, they have abysmal latency due to the connectivity between the speakers being wireless. I measured about ~16.26ms of latency added just by these speakers alone, which is definitely noticeable when I monitor my guitar playing for example.

Thanks for the help!

 

[jamius19]

Digital amp?

Yeah, I got confused because of the Fosi Audio's support. Please see the attached screenshot above.

 

[jamius19]

Just an update: Fosi Audio support just got back to me saying TB10D EQ doesn't involve any DSP.

 

[Bwaslo]

I certainly wouldn't worry about microseconds or even low milliseconds of latency. Every extra foot distance from the speaker you listen at will add about a millisecond of latency, just from the speed of sound in air.