Request 32 or lower (16) Audio Buffer Size for Future Update

Recently I played Pianoteq with 32 Audio Buffer size through DAW. The response key is amazing and accurate. Another benefit is connection between our ears and fingers that is even closer as playing real acoustic Grand Piano. Now I finally can produce pp velocity for every key stroke effortlessly. But there's a problem, there's no 32 Audio Buffer size option on Pianoteq 8 standalone GUI. Playing Pianoteq through DAW definitely demand more power and introduce crackling sound on old MacBook Pro like mine. I am also looking forward to upgrade my machine in the future. I believe there are many Pianoteq 8 users that still have old machine. Adding 32 Audio Buffer size is definitely a dream come true for classical Pianist who often playing songs that demanding wide dynamic from pp to fff.

Another important note that I should added:

Shorter Audio Buffer size like 32 or even lower (16) will allow Piano key travel even higher than before. This way our fingers can feel and be able to executed the key velocity even better than before. The finger fatigue also reduces.

PS:
Excuse my broken English for I have no time to use Grammar check to make it readable for public reader due to work. For me, this 32 Buffer size or even lower (16) on Pianoteq 8 standalone is very important. I really really enjoy my Keyboard and Pianoteq 8 even more than before.

Thank you for reading, and happy playing.

The lower the buffer, the harder the CPU has to work. That’s a general rule.

The other thing: buffer size shown isn’t always the true size. Sometimes there’s what’s called a safety buffer.

Some audio solutions may show low butters but in testing they aren’t as low as that.

The most consistently low latency audio interfaces are from RME. They’re expensive. But they provide playable low latency without needing to go to crazy low buffers.

It’s useful to go through the first messsge on the linked thread and all the links within, especially understanding what all the numbers mean.

You’re looking for the combination that provides low latency at high workload achieved.

Not just low latency.

dikrek wrote:

The lower the buffer, the harder the CPU has to work. That’s a general rule.

The other thing: buffer size shown isn’t always the true size. Sometimes there’s what’s called a safety buffer.

Some audio solutions may show low butters but in testing they aren’t as low as that.

The most consistently low latency audio interfaces are from RME. They’re expensive. But they provide playable low latency without needing to go to crazy low buffers.

It’s useful to go through the first messsge on the linked thread and all the links within, especially understanding what all the numbers mean.

You’re looking for the combination that provides low latency at high workload achieved.

Not just low latency.

Thanks for your through response dikrek.

I am about to upgrade my machine either with M3 or M4 that already quadruple more powerful than my 2014 MacBook Pro. Yes, I am aware of the CPU that will work harder with low latency 32/16 buffer size. However, I am very positive with potential great results in 32 or 16 buffer size when I got the M3/M4 in few weeks because with my current 2014 MacBook Pro I can still play Pianoteq 8 even though its crackling noise a little bit. That's 10 years old machine.

I do have RME Babyface PRO since I got my Pianoteq 6 along with my 2014 MacBook Pro. I totally agree what you said about this RME, they really made amazing product with low latency even though it using 2.0 USB which still on par or "might be" superior than 3.0 USB. For my machine, I did my homework and testing before I decided to buy one, because am gonna use them for long term.

Really love your response. Best regards.

dikrek wrote:

So if you have the RME already, I assume you’re using that for when using the app? Not the Mac built-in audio?

Absolutely, yes.

dikrek wrote:

I find that with RME even 128 buffer is OK for good latency (my Mac isn’t 10 years old though, more like 5).

OK is OK if you know what I mean. 64 is acceptable. However, recently I enjoy playing even more when I use 32 even with crackling sound on my 10 years old Mac Machine. The RME definitely help a lot.

Your 5 years old Mac must be M1 or might be Intel. I'd like to know how it plays using 32 buffer on your Mac through your RME. I am about to get M series MacBook Pro in a few weeks.

dikrek wrote:

The newer Pro FS has a tiny bit lower latency (pretty much the minimum it can be). But even yours is OK.

USB 3 doesn’t provide lower latency. RME has a whole video explaining this. Their only interfaces on USB 3 use that due to high channel counts (over 60-something channels you start needing higher throughput).

Yes, I already watched the video on RME channel. They really make good product that reliable. For my usage, Babyface PRO is overkill. But its worth every penny in the long run.

So far I don't have any other input for Modartt to improve their amazing Pianoteq other than this new one which is Optional for 32 buffer size or 16. Once I got my new MacBook Pro, I will upload the video so I can showcase how it can improve my sound on the Piano. Not only me, many people soon will enjoy the lowest latency ever from Pianoteq.

Thank you for your positive response.

dulistan heman wrote:

Recently I played Pianoteq with 32 Audio Buffer size through DAW. The response key is amazing and accurate. Another benefit is connection between our ears and fingers that is even closer as playing real acoustic Grand Piano. Now I finally can produce pp velocity for every key stroke effortlessly. But there's a problem, there's no 32 Audio Buffer size option on Pianoteq 8 standalone GUI. Playing Pianoteq through DAW definitely demand more power and introduce crackling sound on old MacBook Pro like mine. I am also looking forward to upgrade my machine in the future. I believe there are many Pianoteq 8 users that still have old machine. Adding 32 Audio Buffer size is definitely a dream come true for classical Pianist who often playing songs that demanding wide dynamic from pp to fff.

Another important note that I should added:

Shorter Audio Buffer size like 32 or even lower (16) will allow Piano key travel even higher than before. This way our fingers can feel and be able to executed the key velocity even better than before. The finger fatigue also reduces.

PS:
Excuse my broken English for I have no time to use Grammar check to make it readable for public reader due to work. For me, this 32 Buffer size or even lower (16) on Pianoteq 8 standalone is very important. I really really enjoy my Keyboard and Pianoteq 8 even more than before.

Thank you for reading, and happy playing.

Hello,
Are you sure you can feel the difference between 64 and 32 samples ?

I ask that because some years ago I did some tests at home : I asked to my son to change the buffer from 64 to 128 in a random way. There were 30 attempts at 128 samples and 30 attempts at 64, I had to play without knowing which was the value, and give my opinion on the buffer value.

After a basic statistical analysis, my conclusion was that I couldn't give consistently the correct buffer value between 64 and 128.

It's very subjective. Now you could be hyper sensitive to latency.

dikrek wrote:

It also depends on the interface. Some don’t do much below a certain number. Meaning they the true latency is similar.

The only way to properly test it is using the RTL utility (you connect a cable from the output to the input and it sends pulses).

You are right for the interface. Mine is a Babyface Pro, expensive but solid and hyper stable.
Anyway, there is certainly some difference with the lowest buffer sizes. In my case it's just that I don’t feel it under 128. So, for me, it seems useless to go to play at 32 or 16.

Are you sure you can feel the difference between 64 and 32 samples ?

I ask that because some years ago I did some tests at home : I asked to my son to change the buffer from 64 to 128 in a random way. There were 30 attempts at 128 samples and 30 attempts at 64, I had to play without knowing which was the value, and give my opinion on the buffer value.

After a basic statistical analysis, my conclusion was that I couldn't give consistently the correct buffer value between 64 and 128.

It's very subjective. Now you could be hyper sensitive to latency.

dulistan heman wrote:

For Jazz piano, Gospel, Pop and Rock genre that using less velocity, using 64 or 128 buffer size is great.

For Classical and Church music that requires wide range velocity from pp to fff, using 32 buffer size can give our fingers more room to hit notes with more accurate precision. I encourage anyone to experience it on their own computer "to Feel" the difference.

Lower latency helps better control of dynamics but not because of these reasons but rather indirectly

Very fast playing such as extremely rapid arpeggios or repeated notes are in the region of 10 notes-15 notes per second max in piano playing.    So if you take 10 notes per second as a reference , it means that you have 100 ms between each note,  which is way above the  latency observed in the audio chain , so your fingers can control the keys  with no difference, whether  the audio latency is  5 ms, 10 ms or 15 ms audio latency.

BUT

when playing expressively, especially rapid dynamic shifts the pianist relies on immediate auditory feedback. High audio latency disrupts the feedback loop. If the pianist doesn't hear the sound  of  soft touch (pp) or hard strike (ff) immediately, the brain has trouble controlling with accuracy the next note's dynamic and much force is needed to press the key.

Low latency provides a near-instant feedback that  your neuromuscular system needs to allowing for fine,  instinctive dynamic control. 

Also note that Jazz or Progressive Rock are very expressive forms of music and require as much dynamic control as traditional classical music, but  pop and traditional rock usually use a reduced range.

Pianistically wrote:

Lower latency helps better control of dynamics but not because of these reasons but rather indirectly

Very fast playing such as extremely rapid arpeggios or repeated notes are in the region of 10 notes-15 notes per second max in piano playing.    So if you take 10 notes per second as a reference , it means that you have 100 ms between each note,  which is way above the  latency observed in the audio chain , so your fingers can control the keys  with no difference, whether  the audio latency is  5 ms, 10 ms or 15 ms audio latency.

BUT

when playing expressively, especially rapid dynamic shifts the pianist relies on immediate auditory feedback. High audio latency disrupts the feedback loop. If the pianist doesn't hear the sound  of  soft touch (pp) or hard strike (ff) immediately, the brain has trouble controlling with accuracy the next note's dynamic and much force is needed to press the key.

Low latency provides a near-instant feedback that  your neuromuscular system needs to allowing for fine,  instinctive dynamic control. 

Also note that Jazz or Progressive Rock are very expressive forms of music and require as much dynamic control as traditional classical music, but  pop and traditional rock usually use a reduced range.

I love reading the long analysis. Can you also comment on my last point about higher buffer size options that Pianoteq has up to 4096?  How useful 4096 vs 32 buffer size for your playing?  or what buffer size that you always use for playing, 4096?

Best regards.

dikrek wrote:

Large butters aren’t for real time playing, they’re for lower CPU overhead when playing back during, say, mixing and mastering.

  Do you think Pianoteq use that high buffer size 4096 for mixing and mastering?  Pianoteq is not DAW (still give options for 4096), but you mentioned the high functional for Mixing and Mastering even though they're not for real time playing.  Have you ever use that Pianoteq high buffer size?   What do you think the playability of 4096. Is it really beneficial for your playing using 4096 vs the lower 32 "IF" Pianoteq give the option for future update?

I am trying to find the correlation here that really beneficial for us Pianist who use Pianoteq mainly for real time playing. That's why we test, buy, and enjoy the beautiful technology to make everyone happy, particularly the audience who always love watching us playing happily to entertain them.

Best regards.

I think Modartt just used some standard dialogue box for the buffer settings, that shows what’s available with the current hardware. It changes depending on what’s connected, if the interface supports a lower number then you may see lower.