Why is it so hard to EQ Pianoteq?

Hello Mr. Construer,

I share the frustration you have described, regarding EQ.

Rather than avoiding your question, if I may speak from experience:  The "sound" of any real acoustic piano is heavily influenced by the size, shape and acoustic treatment of the space any piano happens to actually be in.

As a person who tunes pianos professionally (and sticking with grand pianos for the time being), I can attest that a piano sounds much more lively and bass-heavy when it is in a studio or smallish room.  The opposite is also true:  a studio grand piano or a concert grand piano that sounds bass heavy and lively when heard in a comparatively small sized space ... all of a sudden sounds "lean" or or "anemic" -- light in the bass region, and with far less dynamic range when that very same piano is physically re-located to a concert stage, or in a concert venue. 

What has changed?  Not necessarily the "sound" being emitted the piano; rather, what has changed is the amount of early reflections, late reflections and outright dynamic range that our ears hear, once a piano has been re-located from one place to another.

Restated, I have heard small grand pianos (on the order of 5'3" -- ~160cm -- or less) that sound absolutely "huge" in a domestic living room of one's home.  At the same time, I have heard 7' (~210cm) and 9' (~270cm) grands that sound comparatively "anemic" when on an outdoor stage, unamplified, or in a concert venue seating more than 2000 or 3000 people.

* * * * * * *
Here is a completely different analogy:  Have you ever noticed how when we hear recordings of our own voices, we complain our reproduced voices sound shrill and weak?  This is because we have grown accustomed to hearing our own voices from within the confines of our own skulls!  Restated, we hear our own speaking voice with the added benefit of extra resonance that is not heard in external recordings.  Shockingly, despite our complaints, everyone who hears our recorded voice, swears it sounds just like they hear, live.  What's the difference in perception?  I believe it is we have superimposed our own "cranial EQ" onto our own voice!

I firmly believe that the frustration we experience with products such as Pianoteq presets ... happen to make us aware that what we are hearing from Modartt (and other vendors) does not match with the expectations of what we have heard in our own lives.

* * * * *

If I may share a mistake I am guilty of having committed:
I have over-EQ'd countless recordings/performances of Pianoteq models via headphones to enhance lower and midbass, that sound horribly boomy or bass-heavy, and sluggish when played back through a stereo system of full-range studio monitors (such as B&W Matrix 801's with Krell amplification). 

Enough of my rambling.  I hope this has shed some light on the real frustrations you have experienced in your musical life.

Cheers,

Joe

Problem as I see it is we expect Pteq to sound like a good digital rendition of an acoustic.  Digitals are painstakingly designed to sell in maximum numbers.  They wouldn't do that if they actually sounded like an acoustic, with its multiplicity of keyboard and sound inequalities, unwanted resonances due to room or positioning and other undesireables.
Nothing is perfect.
Pianoteq conveys some of this and to my way of thinking, there is a realism about the product which can hit you between the eyes.  You might not like it; but you get to compare the different piano makes which tell their own story.  A Petrof might sound lovely until you compare it with something else.  So many variables also come into play!

It's just the way it is.  But I would love a volume-equaliser gradient control for the entire keyboard.  That, if it were possible, would be a first, I imagine.   And I'd gladly pay for it!

Ptq has a lot of everything, yes, and I have been tweaking (trying) Ptq since 2013. And have same experiencec as many here. When it sounds good with my headphones (Grado SR 80i right now) it doesn’t always sound good with my speakers ( Presonus eris E5 right now). So, I’m always coming back to the presets - in my opinion, after all, presets are pretty much there, needs just little tweaking. Nowadays I sometimes only change impedance or symphatetic resonance or instrument condition slider or combinations of these. And my friends have never complained about the sound, not here in the forum either. But, as always, we hear different and it comes down to personal taste, as with food 

https://youtu.be/O8aC17J6OR4

I made this video today, seemed easier to explain, at least how I see/hear things.

Although your question is regarding EQ, which I demonstrate in this vid, VOLUME is the main reason audio sounds wrong. If you are using a DAW for example to set volume levels, 0 dB will be very loud. It's early(ish) days in my understanding of broadcast audio levels, but peaking around -12 seems about right in my opinion although this will depend on your playing, a piece might be peaking too often in this area.

The overall volumes in my video are slightly under here to allow for the equalization changes.

Also note the temperament here....the last chord especially!

Hope the video is useful to you and other Pianoteq users.

Nick

What do you use to input midi in Pianoteq? There are 0 dynamics in the video you posted, no pedal, phrasing, articulation, etc. Perhaps the problem is in there?
Edit: I can see now it's free score, that explains everything, lol!
My main suggestion is to get a good keyboard with sustain pedal. Editing a piano performance working in pure midi is a task near to impossible. After that and some good speakers you won't make so many changes to the piano sound.

Thank you all for reading through my frustration and trying to help.

To answer questions...

My workflow is based on 'bedroom producer' setup: laptop and studio headphones (at home) and jbl305 in treated room (at school I work in), the later I use only to check my 'mixes'. Workflow is: composing - score editing/midi exporting - midi editing (with great deal of tempo and velocity manipulation) - audio processing.

In the example I posted, I tried Reaper which is awful for MIDI editing, so I will probably go back to Cakewalk (I need to use free/cheap options because of my students). Reaper is good for audio processing.

At first, I was trying to achieve classical recording of some Chopin piece as a reference (I forgot what it was, used it months ago). Then, I realized it is mission impossible so I searched for a recording that is the closest to Pianoteq presets and arrived at Ludovico Einaudi - Oltremare. This reference is somewhere between classical and pop sound, more closely to pop.

Processing I use is (after turning off all Pianoteq fxs): narrow q Eq to remove frequencies I don't like (only a few), then I used transient shaping because of the next step which is 'a touch of Eq' for spectral balance.

And here started THE problem.

I'm baffled by a need to boost lows and low mids, since in my experience I can't cut them down enough without destroying the 'natural sound' (I use Bechstein). I do this because I'm constantly feeling like 'the player' is nailing down keys with big metal hammers even with velocity of 5 (!?). Shelving down lows (bellow 180Hz) and low mids (bellow 350Hz) helps but creates harshness in upper mids: now everything is overly bright, especially on smaller speakers (phone or even TV) and makes 'ears bleed'. So I cut upper mids (around 3K)  and mid mid (around 800). This helps with brightness, but bring back 'hammering' effect of lows. And all I can do is repeat this cycle indefinitely.

I tried to soften the hammers, reduce hammer noise, change impedance, increase sympathetic resonance... And then I asked for help from a guy whose example of processing Pianoteq I heard and liked, and he revealed to me that mid's transients are way too aggressive. I don't have his FXs (I believe he used Neutron's transient shaper), but I have TDR deedger and free BitterSweet. This helps but not enough, and that's why I started wondering: is pianoteq really intended only for experienced audio engineers (with expensive setup) and home playing, or am I getting something wrong? Hence this post.

I did give a thought to my 'mastering' levels. I actually don't 'master', but more normalize sound with only 1 db of limiting reduction. This gives me end levels that are lower than Einaudi's. Mine are around 10LUFS short term and around 16 LUFS integrated. His are 7 LUFS short term and, I believe, 14 LUFS integrated. But, when I look at K-system scale, I'm in the red most of the time even on K-12 scale.

(I know I sound like I know what I'm talking about, but am aware that my knowledge ends with knowing those concepts exist. )

I could really use some directions what to do. I tried many, many, many configurations of presets, including pianoteq's preEq, and also mine tweaking.... And failed to avoid that forceful hammering or overly bright sound. This basically narrows down my issue.

Thanks again to all for trying to help,
Marko

PS: It's true that this piece lacks macrodynamics. With this more like pop style sound bigger changes in dynamics between sections sound awful. Other than that and lack of pedaling, it is far from minimalistic. Velocity ranges from 40-105, and tempo ranges from 40-80, with every note having its own settings.

PPS: Other than all mentioned processing, there is also a very low level of parallel compression (gentle, not a NY style) with saturation as an attempt to round transients further; short plate reverb with eq to bring back some brightness without harshness, medium PnoVerb to widen stereo image, and long room reverb to give depth. Both EQed. And before limiter, a touch of Gullfoss.

PPPS: After exploring this sound design topic for months, I am quite certain that there is a very big market for a pianoteq preset or pianoteq 7 feature that will allow composers to have 'out-of-the-box' nice result. There are a lot of composers (me included) that doesn't look authenticity in the sound but in the composition. Sound only should be acceptable. Nice generic sound is welcomed.

Hi Ian,

thanks for your prolonged interest in my problem.

It is actually funny that you mentiond demos. There was a period in which I tried to match demo sounds. And concluded that either demos are scam or my copy of Pianoteq (updated to the last version) is defective sound wise. Second option seems to me quite ridiculous.

I would love to have the sound of Schubert's sonata demo, but if you compare it with my example: it is waaaay louder (which means it is heavily processed); it has so clear 3rd and 2nd octave - when I compare my out-of-the-box Sweet and any other preset to it, this range sounds like a rumble with barely audible changes in intonation (hence I need to heavily clear it up); and demo overall sounds so wide, clear and present yet massive. My Sweet preset doesn't sound anything like that, not even nearly.

For the example I posted, I didn't use keyboard. I have Casio PX-350M. For 'Serenity' video I did record MIDI live and then edited just pedal (Reaper recorded three lanes of CC64 for some reason). It didn't improve sound. So I went back to editing MIDI files exported from Sibelius. When I finish editing, I go back and adjust velocity curve if needed. Usually because all this processing influence volume of some velocity ranges.

It would be great if Modartt shares MIDIs and FX chains for their demos so we can learn from them and recreate sound we like.

Marko

The Schubert Sonata demo https://www.modartt.com/bechstein was recorded with C. Beschstein DG Sweet preset with no other postprocessing, 100% direct from audio export from Pianoteq. For checking your audio configuration, here is the MIDI file that was used: https://forum.modartt.com/uploads.php?f...Kabuli.mid. Load it in Pianoteq with C. Beschstein DG Sweet and you should have exactly the same sound.

I will find that Schubert midi file really helpful as a 'benchmark' for analysis and testing. Thanks, Philippe!

Thanks, tmyoung, for the excellent technical info. Over time I've learned to lower my output volume and disable the limiter. Glad to know it's not just me being fussy, I always suspected the limiter was introducing distortion but I couldn't understand why.
A brilliant tool for checking loudness levels is Youlean Loudness Meter - the basic version is free! https://youlean.co/youlean-loudness-meter/

Thank you, Philippe, million times. This was a game changer for me. I struggled so much to figure out what am I doing wrong - MIDI programming or audio processing. Spend so much time working on both in parallel. Once the MIDI was out of picture it took me a weekend to figure out what I should do. Now, my Pianoteq production sounds 'like a record', and I am so grateful for that.

Thank you

The file "Schubert - Sonata in A Major D. 664 - Leyla Kabuli.mid" is marvelous.
It helps a lot. Thank you, Phillippe.
However, I can't figure out what's the usage of cc81 and cc16.
There is a good use of cc67 to dim the tone color.
And, I guess that cc81 and cc16 must have some amazing effect to the sound.
Can anyone tell me what it is?

Yes, cc67 controls the soft pedal, it is a very important controller (soundwise). cc81 and cc16 are general purpose MIDI controllers, but are not used in the Schubert demo.

I meant a list of cc controllers that Pianoteq respond to by default. Like 64, 67, 88... (I believe it is a rather short list.)
This cc16 and cc81 is Yamaha thing (Schubert's midi is recorded on their piano so there are events on those lanes), and I was wondering if Pianoteq responds to them. I manually tested and doubted myself whether I'm doing something wrong. Now, you confirmed it doesn't. It could've saved you some time if we had this list.