• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Would I benefit from an audio interface?

Cifer

Member
Joined
Dec 7, 2021
Messages
45
Likes
5
(Please read the part of the text written in bold letters. I am not asking how much delay I should have or how to minimize delay on my system.)

Hello everyone,

I finally got my stage piano to work with Pianoteq 8 and the Topping DX1 and am now trying to figure out how much further I can reduce the delay between key presses and sound.

When I select a buffer size of 16 samples Pianoteq displays 0.3 ms of latency. The input latency displayed in the Topping panel is 1.33 ms and the output latency 3.50 ms in Safe Mode, 1.50 with Safe Mode disabled. I can't drop the buffer rate any lower without overloading my CPU. With a buffer size of 64 samples - which is a tiny bit more stable - I get 1.3 ms according to Pianoteq and 2.33 ms according to Topping.

Would a low latency audio interface perform any better than the Topping? I have no idea how to compare the performance in this regard.

And would an interface lower the load on the CPU so that I could select lower buffer sizes?


I'm essentially trying to understand how much latency simply comes from my system, and if I could remove more by using something else than the Topping. Sadly I have no way of measuring total delay except software like Latencymon, which to my knowledge doesn't show all of it.

Thanks a lot!
 
Last edited:

Doodski

Grand Contributor
Forum Donor
Joined
Dec 9, 2019
Messages
20,750
Likes
20,760
Location
Canada
I have no idea about keyboard latency and delay but it seems that @ the present rates of latency you are down pretty low in latency. What do other keyboard players say about respectable latency numbers and where should they be at? I am coming from a PC first person shooter environment with latency being all about the speed and it's noticeable and from the perspective of a ex-athlete whom reached performance levels where drag/delay/latency in certain operations is very bad. So I am all aboard on the latency train and seeing what this is about.
 
OP
C

Cifer

Member
Joined
Dec 7, 2021
Messages
45
Likes
5
I have no idea about keyboard latency and delay but it seems that @ the present rates of latency you are down pretty low in latency. What do other keyboard players say about respectable latency numbers and where should they be at? I am coming from a PC first person shooter environment with latency being all about the speed and it's noticeable and from the perspective of a ex-athlete whom reached performance levels where drag/delay/latency in certain operations is very bad. So I am all aboard on the latency train and seeing what this is about.
The latency for an instrument should be low enough to not bother the player. I personally want it to be as low as possible since I am used to a real instrument. I don't care about the exact numbers, but I would like to know if an external sound interface would give lower latency than the Topping. I'm sure it's easy to find out, but I am new to this. Hopefully a more experienced user will see this thread.
 

Philbo King

Addicted to Fun and Learning
Joined
May 30, 2022
Messages
669
Likes
874
Consider that 1 mSec is roughly equivalent to being 13" farther from a speaker. It might mess up your timing if you regularly play staccato arpeggios with 128th notes but is unlikely to affect performance in general.
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,227
Likes
17,806
Location
Netherlands
There may be some tweaks to the Xmos driver. See:

 
OP
C

Cifer

Member
Joined
Dec 7, 2021
Messages
45
Likes
5
Consider that 1 mSec is roughly equivalent to being 13" farther from a speaker. It might mess up your timing if you regularly play staccato arpeggios with 128th notes but is unlikely to affect performance in general.
It definitely affects my enjoyment. I notice very clear differences between every single setting. It has been shown that lowering the latency in Pianoteq by 1 ms doesn't only decrease the overall latency by 1 ms but by more than that. I've even had someone help me blind test by changing the settings randomly over and over.

Either way, my question was whether or not an audio interface will help with the latency over the Topping.
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,227
Likes
17,806
Location
Netherlands
Either way, my question was whether or not an audio interface will help with the latency over the Topping.
I don’t think a USB 2.0 interface will do much better. A Thunderbolt/USB 4 interface might do better. But you need to port and those interfaces tend to be more expensive. If you have PC with PCI-E slots, you may look or a sound card. Make sure it’s a real PCI-E card, and not one with a USB conversion on it.
 

DVDdoug

Major Contributor
Joined
May 27, 2021
Messages
2,920
Likes
3,832
There's gotta' be more latency (another buffer) somewhere... I think there's also some latency in the MIDI reading/decoding before the messages are sent to the software.

An interface with ASIO drivers MAY have lower latency (and it may be easier to adjust buffer size with ASIO). Your software also has to support ASIO, and apparently Pianoteq does.

ASIO was designed for low latency, but the buffers are needed because of multitasking... Not the audio itself, but something else interrupting the audio and hogging long enough for an input/recording buffer to overflow, or an output/playback buffer to underflow. And, your computer is always multitasking, even if you are only running one application.

If you have a Windows computer there is a free online book called Glitch Free about opitimizing your computer for audio.
 
OP
C

Cifer

Member
Joined
Dec 7, 2021
Messages
45
Likes
5
I don’t think a USB 2.0 interface will do much better. A Thunderbolt/USB 4 interface might do better. But you need to port and those interfaces tend to be more expensive. If you have PC with PCI-E slots, you may look or a sound card. Make sure it’s a real PCI-E card, and not one with a USB conversion on it.
Thank you.

So essentially the Topping already has a latency comparably low to a decent interface, is that correct? I would mostly like a solution for laptops, so PCI-E cards are not an option.

For USB 4 I would need a USB-A-to-USB-4 adapter (if that exists), and if I understand correctly the difference wouldn't be big either? Probably not worth the investment then since I just want to hear myself play and nothing more.
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,227
Likes
17,806
Location
Netherlands
So essentially the Topping already has a latency comparably low to a decent interface, is that correct? I would mostly like a solution for laptops, so PCI-E cards are not an option.
There may be some USB pro audio interfaces that are better than Xmos. There are very many measurements of latency out there I’m afraid.
For USB 4 I would need a USB-A-to-USB-4 adapter (if that exists), and if I understand correctly the difference wouldn't be big either?
That won’t work. It needs to be an actual USB 4 interface. Effectively this is just Thunderbolt 3 ;). Adapter cables won’t do you any good.

Found this:
1670618046692.jpeg



Interestingly here, the Thunderbolt interfaces aren’t doing so good..
 
OP
C

Cifer

Member
Joined
Dec 7, 2021
Messages
45
Likes
5
There's gotta' be more latency (another buffer) somewhere... I think there's also some latency in the MIDI reading/decoding before the messages are sent to the software.

An interface with ASIO drivers MAY have lower latency (and it may be easier to adjust buffer size with ASIO). Your software also has to support ASIO, and apparently Pianoteq does.

ASIO was designed for low latency, but the buffers are needed because of multitasking... Not the audio itself, but something else interrupting the audio and hogging long enough for an input/recording buffer to overflow, or an output/playback buffer to underflow. And, your computer is always multitasking, even if you are only running one application.

If you have a Windows computer there is a free online book called Glitch Free about opitimizing your computer for audio.
Most of the delay does indeed come from the computer itself. However, that is a different issue. I want to make sure that the DAC I use has as little latency as reasonably possible. I'm not worried about nanoseconds, but I'm new to this entire matter and have no idea if the latency resulting from the Topping is almost interface levels of low, or (extreme scenario) five times as high. I just know that interfaces are optimized for this use case while that isn't necessarily the true for Hifi/MidFi gear, so that's why I made this thread.

Further optimizing Windows for live audio is another step. I've already done the most important things but can surely improve it more.
 

pierre

Addicted to Fun and Learning
Forum Donor
Joined
Jul 1, 2017
Messages
962
Likes
3,048
Location
Switzerland
(Please read the part of the text written in bold letters. I am not asking how much delay I should have or how to minimize delay on my system.)

Hello everyone,

I finally got my stage piano to work with Pianoteq 8 and the Topping DX1 and am now trying to figure out how much further I can reduce the delay between key presses and sound.

When I select a buffer size of 16 samples Pianoteq displays 0.3 ms of latency. The input latency displayed in the Topping panel is 1.33 ms and the output latency 3.50 ms in Safe Mode, 1.50 with Safe Mode disabled. I can't drop the buffer rate any lower without overloading my CPU. With a buffer size of 64 samples - which is a tiny bit more stable - I get 1.3 ms according to Pianoteq and 2.33 ms according to Topping.

Would a low latency audio interface perform any better than the Topping? I have no idea how to compare the performance in this regard.

Here are the spec from a RME UFX+ for ex:

Screenshot 2022-12-09 at 21.40.45.png

Did you try to increase the sample rate?

And would an interface lower the load on the CPU so that I could select lower buffer sizes?

below 16 samples, it is hard to get a stable stream on Windows. I don't think you can go much lower but I am clearly not an expert in Win.

I'm essentially trying to understand how much latency simply comes from my system, and if I could remove more by using something else than the Topping. Sadly I have no way of measuring total delay except software like Latencymon, which to my knowledge doesn't show all of it.

Thanks a lot!
 

studio7

Member
Joined
Nov 3, 2022
Messages
37
Likes
8
To measure the absolute latency, you can perform a simple test. Strike a key on your piano keyboard with something hard, like a pencil. Record this sound, together with the sound from Pianoteq, coming from your speakers. If you don't have a microphone, use the mic in your phone. Then take the recorded file and zoom in on it in Audacity, or any other sound editor. The difference in milliseconds between the two transients is your absolute latency. In pro hardware (digital pianos from Roland, Yamaha) it doesn't exceed 3ms, but 5-6ms should be OK for home use.
 
OP
C

Cifer

Member
Joined
Dec 7, 2021
Messages
45
Likes
5
There may be some USB pro audio interfaces that are better than Xmos. There are very many measurements of latency out there I’m afraid.

That won’t work. It needs to be an actual USB 4 interface. Effectively this is just Thunderbolt 3 ;). Adapter cables won’t do you any good.

Found this:
View attachment 248864


Interestingly here, the Thunderbolt interfaces aren’t doing so good..
I've seen this before, but my problem is that I don't know how the Topping compares. This data is about roundtrip latency between a DAW and an audio interface, and I don't have either.

I'm really just clueless as to whether the Topping delays things much more, or just by 3 milliseconds or so.
 

Doodski

Grand Contributor
Forum Donor
Joined
Dec 9, 2019
Messages
20,750
Likes
20,760
Location
Canada
I don’t think a USB 2.0 interface will do much better. A Thunderbolt/USB 4 interface might do better. But you need to port and those interfaces tend to be more expensive. If you have PC with PCI-E slots, you may look or a sound card. Make sure it’s a real PCI-E card, and not one with a USB conversion on it.
Would this work?
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,227
Likes
17,806
Location
Netherlands
I've seen this before, but my problem is that I don't know how the Topping compares. This data is about roundtrip latency between a DAW and an audio interface, and I don't have either.
No, but you only need one way, so it should be about half of what you see here.
I'm really just clueless as to whether the Topping delays things much more, or just by 3 milliseconds or so.
I’d like to real latency measurements of the Topping. I suspect it’s more than the panel shows.
That won’t fit in a laptop, will it ;)
 
OP
C

Cifer

Member
Joined
Dec 7, 2021
Messages
45
Likes
5
To measure the absolute latency, you can perform a simple test. Strike a key on your piano keyboard with something hard, like a pencil. Record this sound, together with the sound from Pianoteq, coming from your speakers. If you don't have a microphone, use the mic in your phone. Then take the recorded file and zoom in on it in Audacity, or any other sound editor. The difference in milliseconds between the two transients is your absolute latency. In pro hardware (digital pianos from Roland, Yamaha) it doesn't exceed 3ms, but 5-6ms should be OK for home use.

I knew of this method but thought it would be inaccurate. How exactly do I measure the difference? I don't know where to start and end because I have no idea where exactly the transient of the piano sound is starts. Does this just mean the part around 2.590 where the line starts getting less flat again?

1670621023053.png


You said for a good piano it should be 3 ms, but the latency you measure with this method includes everything which happens in the computer and DAC, and depends on the sample and buffer rate. I'm confused by this as it doesn't seem like measuring total latency tells you anything about the piano's performance.

The Piano I'm using most of the time is an old Kawai ES-100. No idea how much latency it has, and I don't know where/how to check.
 
Last edited:

studio7

Member
Joined
Nov 3, 2022
Messages
37
Likes
8
I knew of this method but thought it would be inaccurate. How exactly do I measure the difference? I don't know where to start and end because I have no idea where exactly the transient of the piano sound is starts. Does this just mean the part around 2.590 where the line starts getting less flat again?

View attachment 248871

You said for a good piano it should be 3 ms, but the latency you measure with this method includes everything which happens in the computer and DAC, and depends on the sample and buffer rate. I'm confused by this as it doesn't seem like measuring total latency tells you anything about the piano's performance.

The Piano I'm using most of the time is an old Kawai ES-100. No idea how much latency it has, and I don't know where/how to check.

I think that your absolute latency is the red rectangle.


latency.jpg



Btw, in my own measurements most of the latency comes from the Piano keyboard > USB > Pianoteq, and only a small fraction (around 2ms) comes from my Audient iD4 audio interface.
 
OP
C

Cifer

Member
Joined
Dec 7, 2021
Messages
45
Likes
5
I think that your absolute latency is the red rectangle.


View attachment 248948


Btw, in my own measurements most of the latency comes from the Piano keyboard > USB > Pianoteq, and only a small fraction (around 2ms) comes from my Audient iD4 audio interface.
Thanks.

I'll measure the latency again with proper settings then. I think it will be around 30 ms.

I know the vast majority comes from the MIDI travel and the software, but I suspect that the Topping is adding an excess amount despite displaying relatively low numbers, because everything else is set up well and I'm not seeing anything running on my PC show huge latency numbers.

I've read reports of people having the same issue with other Topping models and HiFi DACs and have ordered a Focusrite Solo to check. I hope the drivers aren't as bad as many people say, but it appears to me that this is just another case of hating on something because everyone else on reddit does it.
 
Top Bottom