Something about the bass notes.

I checked the Kawai and yes, it's closer to what I expect the piano sound to be.
It seems to have also more presence of harmonics when the note is sustained. May be worth buying (currently have Steinway and K2)

https://www.youtube.com/watch?v=8gLgFZ2Dm5g

Having said that, not sure about an audio being mono. If you mean the sampled audio being mono then it doesn't seem so (listening to it and watching obs sound bars, where the left channel is usually slightly louder than right channel, with some variations), unless you mean stereo width. For context, the real piano samples were recorded with a Tascam DR 40.

Heya,

Individual notes of sample libraries will often sound more impressive (louder/richer/fuller) than Pianoteq. But your bass note example is a great illustration of where this difference comes from: post processing and reverb.

A lot of very high quality audio engineering went into making the raw recordings of a real piano (which to your ear, I predict, would sound really flat) into the rich and saturated sounds you hear - by carefully EQing, saturating and compressing the raw audio. The audio engineers of the sample library will make a trade-off between purity and quality of listening experience, and for the bass notes, will often err towards the latter because those have to sound "impressive".

Pianoteq does far less of that, its sound is much closer to a raw recording (I would really love it if a skilled audio engineer would compare raw recordings, Pianoteq output and processed recordings - I bet people would struggle to tell the difference between raw and Pianoteq).

The cost of that is that the post processing won't work well in every imaginable playing situation - for example, if you play very fast virtuosic passages, trills, or repeated notes, samples will start sounding muddy and/or robotic.

The other thing is that samples benefit from the resonances of their recording environment, which add a lot of flavour and warmth to the sound, separate from the instrument. Pianoteq's built in reverb is okayish but not as good as real reverb (or a high quality, dedicated reverb VST)

So, if you want to replicate the sound of a sample library, you should think of pianoteq's output as a "raw recording" to apply convincing reverb and post-processing to.

I got Ocelot Octaver, pretty wild for this purpose

Room EQ is sadly forgotten by most.

A third and final comparison:
https://www.youtube.com/watch?v=fGfWKuxf4UU

The reason I'm showing this one is because, if you go a bit up in the range, the sound feels better overall, so if I stay away from the notes below C2 things should be ok.

Having said that. I came to the conclusion that if I want more "realism" or accuracy in terms of note per note sound, maybe I should switch to sampled libraries, but since I like Pianoteq as it is I will continue using it for now.

This is very interesting from a technical standpoint, but seems a particularly inappropriate choice of MIDI to demonstrate the expressiveness of Pianoteq. I actually listened to most of the recording before stopping to read the details of its source, and immediately sensed there was something oddly mechanical about the timing and dynamics of the performances. After reading that, I downloaded the file to have a look at it in my DAW.

By the nature of how it was generated, the MIDI suffers from a lot of quantizing of notes in chords to the same tick and velocity and even the durations seem unusually/artificially consistent. The expected rhythmic patterns and cadences of these familiar pieces are not well-represented by the velocities in many sections, and the range and variation of velocities overall is much lower than I would expect to see in a direct MIDI recording. I almost expected to find that the MIDI clock was only 48 or 96 PPQ; it's actually 600 which is a somewhat unsual value but sufficient if the full precision is actually being used.

In addition to the shortcomings of timing and velocities, the sustain messages are on/off only, the soft-pedal transitions are jarring, and the sound is panned noticably to the right, almost giving the impression that the left and right hand parts were generated by two different pianos or were recordeded separately with different microphones or in different acoustic spaces. I also noticed that both the left and right-hand parts are including the same pedal messages; this doesn' seem to be causing any particular problem for Pianoteq aside from possibly making the pedal noises louder, but it seems a significant ovesight in the preparation of the files, and would make me question how much care was taken in other parts of the process.

Even if the MIDI were a direct recording from a modern MIDI controller, accurately capturing all possible nuances of the performance, it would still have the shortcoming that the performer wasn't playing and responding to the particular behavior and sound of Pianoteq. Taking a MIDI performance that was recorded on one instrument and rendering it with another will always lose something in translation.

Bottom line, I'm sorry to say that this recording does not sound at all good to me and actually emphasizes Pianoteq's deficits.

Possibly I was mistaken, but I had previously gotten the impression somehow that you were a fairly accomplished pianist in your own right and should be able to make a far better 'first person' recording to show Pianoteq in the best possible light...?

Suffice it to say, I belive it's not possible for the instrument to 'sing' if the performance doesn't. And even ignoring the glaring inadequacies of the MIDI, I found the sound pretty unappetizing overall, and few, if any, of the velocities were high enough relative to surrounding notes to really show off the impact of a solidly struck bass note. Sorry if that offends; it's just my opinion, and not meant to denigrate your production skills. You obviously experience it differently.

You began by accusing me of being argumentative when I was only stating what I personally prefer in a digital piano sound. Now you're rebutting my comments sentence by sentence and challenging me to some sort of musical dual which seems far more argumentative.

In fact I have relatively recently posted an example of the sound I get from a layered DP/Pianoteq setup alongside renderings of the same MIDI with Pianoteq and the DP alone for comparison. The composition and performance were my own. Despite my years of playing and composing, I'm not at all virtuosic in either regard. I'm not classically trained (or inclined, though my mother and sister played violin and viola professionally and the house was full of classical music as I was growing up), and don't read well at all, so producing my own performance of those pieces is not going to happen. In referring to my years of experience, I only meant to say that I believe my impression of Pianoteq's shortcomings are not likely due to my inability or lack of effort to get a more satisfactory sound from it. I'll grant it came out sounding curt and self-aggrandizing as I was trying to respond quickly on my cellphone as mentioned.

I'll let the rest of your musings go without comment, and hope that we can go both move on, sharing thoughts about all things piano- and Pianoteq- related with a mutually more amicable and respectful tone.

I'll just add one other thought which is that in the time that piano rolls were being recorded, I think most people's experience of music was through live perfromance rather than recording of any kind so I don't really see that the state of the art (or lack thereof) of recording at the time is really germaine to the topic. At the same time, I'll acjnowledge that my comments strayed from the very narrow discussion of bass-note timbres that the OP might have intended, and I apologize if I took the discussion too far off base (pun intended).

Peace,
Dave

Not sure if there's a more artful way to link/embed a post but this should work (not as recent as I thought - April 3. Time flies):

https://forum.modartt.com/viewtopic.php...3#p1001443