Before attempting to build a new .fxp file with version 6.5.3 that has a better possible fit of the string lengths, I tried to capture a reference sound for this instrument, adapted to the "player" position. "with a recording of each of the 88 notes with up to 11 microphones (non professional unfortunately). Here is the link to the 11 tracks (24-bit flush 96khz) of my last registration attempt
"REAL-BOSEN-200-11-MICS-TRACKS-PART1.zip" (1 Gb) at http://dl.free.fr/i1ST6UU3s
"REAL-BOSEN-200-11-MICS-TRACKS-PART2.zip" (1 Gb) at http://dl.free.fr/i9r277QWr
as well as a link on the photos of my last (amateur) real positioning of the mics:
https://photos.google.com/share/AF1QipP...F3M1lPVGtB

I'm now thinking about the mix and the selection of tracks to keep to build the reference sound of the 88 notes
(advice is welcome) to work later on the spectrum profile of Pianoteq pro, to try to set a virtual piano Pianoteq a little closer to the acoustic instrument, continuing to hope that Modartt offers "a day" a model closer...

Bruno

Edit: Something just hit me that explains the spectrum differences: I started from your already modified spectrum to generate mine, while you started from the raw U4, so what I get is a sort of combination of your changes with mine...The result is then not exact since pasting the note edit string replaces your result and does not add to it. I guess I should start again by removing your spectrum values first...sigh...

I corrected the blunder above by redoing the analysis from a raw U4 spectrum for notes 21 to 33. Not much audible difference but might as well be exact...

https://www.forum-pianoteq.com/file/azc1xt3w

Hi Gilles,
Thank you for this review.
The correction made on the lower notes in your fxp with your tool brings a deeper sound also very nice.

By analyzing the spectrum profile of your fxp on these notes and comparing it to mine, I noticed several differences:
(1) Much more contrast, explainable difference because I used here a scale factor dividing (in db) by 4 the spectrum differences initially obtained. Your fxp shows me that the scale factor I use must be less important at least for the lowest notes. (With a scale factor that divides by only 1.1 on these low notes, I can get a sound closer to the original one (scale factor, however, unsuitable for higher notes). Unfortunately, I have not yet enough perspective to choose the best scale factor note by note automatically according to the frequency of the fundamental.In my fxp, the scale factor was constant: division (in db) by 4 no manually modified, which should be reviewed)

I use the scale factor very empirically in the note edit window: I simply adjust it by ear to get rid of unwanted artefacts, usually overemphasized bell sound.
In this case A-1 was left as is but on E0 and F0 I had to rescale a lot partly because the source and destination notes were not exactly in tune (to my ears)

Edit: Something just hit me that explains the spectrum differences: I started from your already modified spectrum to generate mine, while you started from the raw U4, so what I get is a sort of combination of your changes with mine...The result is then not exact since pasting the note edit string replaces your result and does not add to it. I guess I should start again by removing your spectrum values first...sigh...

I am by no mean an expert in signal processing...I simply used an FFT algorithm written in Java I found on the web and built a specialized program around it for use with Pianoteq Pro. I experimented with it in the past and found that the default 16K sample number and rectangular window seemed to work best for the type of signal piano notes provide, mostly short, so the window limits effects are not that important in my opinion. I used for the Schimmel your longer samples (4.5s) instead of the short ones (0.4s) because they were easier to extract...(lazy...) but trying for instance A-1 with your short sample and a 65K window only produced more bell inharmonics. This is all very empirical because for instance the signal starts with a hammer sound that is separate in Pianoteq and thus wrongly included in the sound, and also there is, when recording long notes, the sound of the damper also modeled separately. I think the only justification for my approach (or yours) is that we get audibly closer to the target sound. We can never reach it though... My software only accepts 16/44.1 samples so I also am less concerned by FFT window size or type than you if you work with bigger 24/96 samples.

I don't hear anything wrong with G#2 but the "zing" resonance on C#2 can be removed by changing the string length to something like 1.03m. Using my tool only produce a very dull result. I had to cut the note short because of the important damper sound at the end.

bm wrote:

Advice and criticism would be welcome to try to improve the timbre of these virtual instruments, especially in the low medium and the medium.
From the 2 audio comparisons: (virtual versus real)
https://www.forum-pianoteq.com/download.php?id=3431
and
https://www.forum-pianoteq.com/download.php?id=3432
I note that the duration of the sound of the virtual instrument remains abnormally long.

Bruno

Hi Bruno,
I really like your 5 mics version even though the 2 mics one is more accurate as a player perspective, the 5 mics has a very nice ambiance to it.

One thing that still sounds like the U4 is the "throaty" sound of the lowest notes that is characteristic of the Pianoteq model, while your Schimmel has a rounder low range with more bell content. I used my PtqSpecProf software and the notes you supplied to modify slightly the lowest octave (notes 21 to 33).

Here is the fxp with a short comparison of those notes:

https://www.forum-pianoteq.com/file/4wbmby63

https://www.forum-pianoteq.com/uploads....iginal.mp3
https://www.forum-pianoteq.com/uploads....=modif.mp3

I started actually with the more precise 2 mics version for the comparison and ported the results to the 5 mics one. One thing that was a bit bothersome is that some notes are more stretched downward in the Pianoteq model versus your real piano. I noticed this after beginning the tedious process of extracting and analyzing notes one by one and I was too lazy to start over again by modifying the detuning so some of the results have been rescaled to correct some discrepancies.

Still, I think this is some improvement...If you want me to try improving some other notes, let me know.

Hi Gilles,
Thank you for this review.
The correction made on the lower notes in your fxp with your tool brings a deeper sound also very nice.

By analyzing the spectrum profile of your fxp on these notes and comparing it to mine, I noticed several differences:
(1) Much more contrast, explainable difference because I used here a scale factor dividing (in db) by 4 the spectrum differences initially obtained. Your fxp shows me that the scale factor I use must be less important at least for the lowest notes. (With a scale factor that divides by only 1.1 on these low notes, I can get a sound closer to the original one (scale factor, however, unsuitable for higher notes). Unfortunately, I have not yet enough perspective to choose the best scale factor note by note automatically according to the frequency of the fundamental.In my fxp, the scale factor was constant: division (in db) by 4 no manually modified, which should be reviewed)
(2) Differences sometimes completely opposite on the level to be added or subtracted on this or that partial.
Is this due to differences in methods used for spectral analysis between your tool and my pieces of code? (at your disposal if you wish).
In practice I use (via Audacity) a spectral analysis with a Blackman-Harris type window, whose size varies according to the range of notes according to the rule (empirical): next, maybe to re-adapt: A- 1 to F1 => 65k with the 686 ms from the beginning of the sample;
F # 1 to F2 => 32k with the 346 ms of the beginning .., F # 2 to D # 4 => 16k with the 172ms of the beginning; E4 to F5 => 8k with the 88ms of the beginning; F # 5 to B5 => 4k with the beginning 46ms; C6 to B6 => 2k with the 24ms of the beginning; C7 => 1k with the 14ms of the beginning.
The results between B0 and B3 are of insufficient quality with this empirical technique (nevertheless preserved in the state in the diffused fxp).
I think you use longer samples (1s) and a 16 or 32k rectangular window with your tool, is that it?
For my part, I had avoided the use of rectangular windows because I found (with Audacity) a lower discriminating power above 4khz (not because of the frequency itself, but because the level is much lower for high harmonics), but I'm not at all certain to have chosen the right type of window, the most suitable, nor the optimal sample size, what do you think?
I had only noticed, the higher harmonics disappearing quickly, it was not necessary that the duration of the sample is not too important after the impact of the hammer ..., unfortunately in this case to the detriment of the level of the first harmonics for the bass.

Would you be able to reproduce with your tool in a fxp the following notes of my recording of T130 with 2 microphones SE8, I hope that it will be able to make disappear the horrible stamp which I obtain for example on C # 2 and G # 2 for example ...
NB: for the reconciliation of real and virtual notes, my starting point was also a fxp with only 2 microphones.

Bruno

Hi bm,

love the 5 mics FXP particularly to tweak.

Uploaded this FXP to adjust some things which I hope you enjoy. Mainly, I made changes to velocity curve (can be reset to flat of course - I just like this curve, left it in case others who haven't tried a more gentle curve find it helps their keyboards with less subtle velocity) reverb, mic balances, delay timings and added some cross-talk to rear mics, removed the delay in effects (could use that slot for another EQU3 perhaps).

To EQ, with 2 instances of EQU3 I raised lower bass shelf, removed some mid around 310Hz, added some 1400Hz, and took out some treble. Arguably, this is a first pass and sounds like a starting point - would recommend trying moving the dots to neighboring frequencies and lowering and adding a little until you have it the way you like. (a little more text in the FXP itself in the description box but that's essentially it).

In my experimenting, the EQ and change to reverb removes a lot of room cloud - but if wanting some of that roominess back, just up the reverb and reverb time etc. - although, have to say a convolution reverb (like Vienna Hall) makes an interesting result (have some pet theories about it, which I may have typed about before in other threads, mainly to do with 'player bubble', breaking down reverbs into separate overlapping mixed 'tracks', player region vs. whole room which is pure piano plus player bubble range, mix of amplification of components of this and subtractions also to suit the overall space - this work of yours bm, kind of pre-adds something positive to that dimension I feel, due to your process and note-edits/frequencies chosen - to me it results in a very delightful and inspiring experience).

Delicious to play - thanks again bm and hope these tweaks lead to some further ideas in making your piano presets sing.

Links:

FXP

Schimmel-T130-2Mics-SE8-V0.71X_5VirtualMICS - Qexl Tweak A.fxp

Audio
qexl-improv-20161406c64.mp3

Hi Qexl,
Thank you for this review. The sound of your fxp is actually more woody and very nice. Thank you also for your equalizing technique to get this woody sound. The actual instrument, however, was a little less woody when recording with open lid and front box. An important contribution in your fxp is the use of a velocity curve much more adapted for the keyboard, the virtual instrument was indeed sometimes difficult to play with the linear velocity curve initially proposed. For the amateur I am, I still have so much to learn to understand better how to build the sound of this exciting instrument, and all your remarks are a great help.

Bruno

Advice and criticism would be welcome to try to improve the timbre of these virtual instruments, especially in the low medium and the medium.
From the 2 audio comparisons: (virtual versus real)
https://www.forum-pianoteq.com/download.php?id=3431
and
https://www.forum-pianoteq.com/download.php?id=3432
I note that the duration of the sound of the virtual instrument remains abnormally long.

Bruno

Hi Bruno,
I really like your 5 mics version even though the 2 mics one is more accurate as a player perspective, the 5 mics has a very nice ambiance to it.

One thing that still sounds like the U4 is the "throaty" sound of the lowest notes that is characteristic of the Pianoteq model, while your Schimmel has a rounder low range with more bell content. I used my PtqSpecProf software and the notes you supplied to modify slightly the lowest octave (notes 21 to 33).

Here is the fxp with a short comparison of those notes:

https://www.forum-pianoteq.com/file/4wbmby63

https://www.forum-pianoteq.com/uploads....iginal.mp3
https://www.forum-pianoteq.com/uploads....=modif.mp3

I started actually with the more precise 2 mics version for the comparison and ported the results to the 5 mics one. One thing that was a bit bothersome is that some notes are more stretched downward in the Pianoteq model versus your real piano. I noticed this after beginning the tedious process of extracting and analyzing notes one by one and I was too lazy to start over again by modifying the detuning so some of the results have been rescaled to correct some discrepancies.

Still, I think this is some improvement...If you want me to try improving some other notes, let me know.

love the 5 mics FXP particularly to tweak.

Uploaded this FXP to adjust some things which I hope you enjoy. Mainly, I made changes to velocity curve (can be reset to flat of course - I just like this curve, left it in case others who haven't tried a more gentle curve find it helps their keyboards with less subtle velocity) reverb, mic balances, delay timings and added some cross-talk to rear mics, removed the delay in effects (could use that slot for another EQU3 perhaps).

To EQ, with 2 instances of EQU3 I raised lower bass shelf, removed some mid around 310Hz, added some 1400Hz, and took out some treble. Arguably, this is a first pass and sounds like a starting point - would recommend trying moving the dots to neighboring frequencies and lowering and adding a little until you have it the way you like. (a little more text in the FXP itself in the description box but that's essentially it).

In my experimenting, the EQ and change to reverb removes a lot of room cloud - but if wanting some of that roominess back, just up the reverb and reverb time etc. - although, have to say a convolution reverb (like Vienna Hall) makes an interesting result (have some pet theories about it, which I may have typed about before in other threads, mainly to do with 'player bubble', breaking down reverbs into separate overlapping mixed 'tracks', player region vs. whole room which is pure piano plus player bubble range, mix of amplification of components of this and subtractions also to suit the overall space - this work of yours bm, kind of pre-adds something positive to that dimension I feel, due to your process and note-edits/frequencies chosen - to me it results in a very delightful and inspiring experience).

Delicious to play - thanks again bm and hope these tweaks lead to some further ideas in making your piano presets sing.

Links:

FXP

Schimmel-T130-2Mics-SE8-V0.71X_5VirtualMICS - Qexl Tweak A.fxp

Audio
qexl-improv-20161406c64.mp3

bm wrote:

bm wrote:

Pianoteq 6.5.2 / 6.5.3 brings a very interesting evolution with the enrichment of the strings plans according to the length of strings set in Pianoteq.
This now allows you to build custom presets (fxp) using the length of the string that best fits the partials of the acoustic piano you want to simulate. (In connection with pianoteq pro note by note spectrum profile)
Bruno

I just finished a first build of .fxp with my 1992 Schimmel T130 upright, from the Pianoteq U4 model, playing on the length of strings.
Link for the .fxp file:
https://www.forum-pianoteq.com/file/da7c6r82
I tested for the first time a sound recording with 7 Rode Mics (including 5 preserved), link on the sound recording:
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB
Link on the spectrum profile recovery (amateur) method, after having previously adjusted the length of the strings in Pianoteq pro for match (attempt) with the partials of each note (between virtual instrument and real instrument):
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB
and (to prepare the calculation of the partial level adjustment):
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB

The result is in my opinion moderately satisfactory because the room (7.5 x 6.5m) is quite reverberant (tiled floor, despite 2 carpets) and microphones remain cheap models.
I also encountered a pedal management problem (damper) that with the U4 model of pianoteq behaves in a bizarre way.

I will do a second modeling of the T130 soon, only from 2 new microphones acquired (SE-8 match pair) whose results are very promising, (much better than my Rodes) - ditto use an audio interface Audient ID14, to Tascam Us 16x8 place used here. These tests on Schimmel will allow me to prepare a next model of our Bosendorfer 200 grand piano, which I await the return of the keyboard currently being restored.

New FXP1: Schimmel T130, adjusted profile spectrum with 2 mics SE8 (match pair) in AB position at 50 cm from the hammers, front case removed. Here too, the length of the strings was previously adjusted for each note to adjust the partials of the virtual instrument to those of the real instrument.
Schimmel-T130-2Mics-SE8-V0.71 :
https://www.forum-pianoteq.com/file/3sv01a3r

New FXP2: Schimmel T130, same profile as V0.71 but here with 5 ptq virtual mics, to better reproduce the acoustics of a living room of 6.5 x 7.5m in player position.
Schimmel-T130-2Mics-SE8-V0.71X 5 :
https://www.forum-pianoteq.com/file/7uvadcy1

cf pictures for the sound recording on the real instrument:
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB
and
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB

Advice and criticism would be welcome to try to improve the timbre of these virtual instruments, especially in the low medium and the medium.
From the 2 audio comparisons: (virtual versus real)
https://www.forum-pianoteq.com/download.php?id=3431
and
https://www.forum-pianoteq.com/download.php?id=3432
I note that the duration of the sound of the virtual instrument remains abnormally long.

Bruno

  In the second audio file, aren't the notes on the real piano stopped early by the damper? It's hard, for me, to compare the notes when I can't hear how the sound evolves. However, I hear the first notes, the Pianoteq notes, as having more lower midrange frequencies. And I must admit that without the damper cut-off, and until I read the name of the file carefully, I might have thought that they were the notes from the real piano.

bm wrote:

Pianoteq 6.5.2 / 6.5.3 brings a very interesting evolution with the enrichment of the strings plans according to the length of strings set in Pianoteq.
This now allows you to build custom presets (fxp) using the length of the string that best fits the partials of the acoustic piano you want to simulate. (In connection with pianoteq pro note by note spectrum profile)
Bruno

I just finished a first build of .fxp with my 1992 Schimmel T130 upright, from the Pianoteq U4 model, playing on the length of strings.
Link for the .fxp file:
https://www.forum-pianoteq.com/file/da7c6r82
I tested for the first time a sound recording with 7 Rode Mics (including 5 preserved), link on the sound recording:
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB
Link on the spectrum profile recovery (amateur) method, after having previously adjusted the length of the strings in Pianoteq pro for match (attempt) with the partials of each note (between virtual instrument and real instrument):
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB
and (to prepare the calculation of the partial level adjustment):
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB

The result is in my opinion moderately satisfactory because the room (7.5 x 6.5m) is quite reverberant (tiled floor, despite 2 carpets) and microphones remain cheap models.
I also encountered a pedal management problem (damper) that with the U4 model of pianoteq behaves in a bizarre way.

I will do a second modeling of the T130 soon, only from 2 new microphones acquired (SE-8 match pair) whose results are very promising, (much better than my Rodes) - ditto use an audio interface Audient ID14, to Tascam Us 16x8 place used here. These tests on Schimmel will allow me to prepare a next model of our Bosendorfer 200 grand piano, which I await the return of the keyboard currently being restored.

New FXP1: Schimmel T130, adjusted profile spectrum with 2 mics SE8 (match pair) in AB position at 50 cm from the hammers, front case removed. Here too, the length of the strings was previously adjusted for each note to adjust the partials of the virtual instrument to those of the real instrument.
Schimmel-T130-2Mics-SE8-V0.71 :
https://www.forum-pianoteq.com/file/3sv01a3r

New FXP2: Schimmel T130, same profile as V0.71 but here with 5 ptq virtual mics, to better reproduce the acoustics of a living room of 6.5 x 7.5m in player position.
Schimmel-T130-2Mics-SE8-V0.71X 5 :
https://www.forum-pianoteq.com/file/7uvadcy1

cf pictures for the sound recording on the real instrument:
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB
and
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB

Advice and criticism would be welcome to try to improve the timbre of these virtual instruments, especially in the low medium and the medium.
From the 2 audio comparisons: (virtual versus real)
https://www.forum-pianoteq.com/download.php?id=3431
and
https://www.forum-pianoteq.com/download.php?id=3432
I note that the duration of the sound of the virtual instrument remains abnormally long.

Bruno

The version 0.71b corrects the hollow B2 and sounds very good to my ears!

That Schimmel looks a lot like the Kawaii KS3F I used to own. I still have metal cassettes (Dolby C) recordings of me playing it. Unfortunately my old cassette player died so I can't listen to them any more...I should have digitized them when it was still working...

Nice job! Very faithful emulation of the Schimmel and the room presence with those 5 mics is really outstanding!

Small remark: B2 has a hollow resonant quality that is different from the source. Maybe some mic interaction since changing the spectrum does not correct it.

Hello Gilles,

It remains - indeed - important differences between the real instrument and the .fxp v0.71 and many adjustments to make, especially on low medium and low ... (This T130 has the particularity of a tuning the mechanism by my tuner to strengthen a little the sound level of the last octave)
To quickly try to limit this problem of bass, but with other edge effects, I added 2 variants of .fxp

(V0.71a): Variation with reversed phase on the rear microphone.
https://www.forum-pianoteq.com/file/9871vvo5
(V0.71b) Only small adjustments on spectrum profile for B1 & B2
https://www.forum-pianoteq.com/file/4qeindti

I also added below, a link to load an extract of the sound recording with the tracks of the 7 original Rode microphones (24bit 96Khz) for Audacity (about 1 GB to unzip): http://dl.free.fr/nh0Y3hbgi

I hope that the best sound taken, opened T130 piano with only a couple of SE8 mics matched & an interface Audient ID14, will allow my next .fxp to avoid this problem in part. (see 2 pictures for the 1st recording:
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB
and:
https://photos.google.com/share/AF1QipM...ZXSWxyNlBB

Overview of this first sound recording with these 2 mics SE-8 (real instrument T130) (0.42 GB flac file): http://dl.free.fr/wfMgBjKhd

Bruno

Small remark: B2 has a hollow resonant quality that is different from the source. Maybe some mic interaction since changing the spectrum does not correct it.

Wonderful work bm,

it sounds remarkable!

+1. I'm hugely impressed by this 'Schimmel' fxp - excellent 'you are there' presence, very much looking forward to the follow-up version!

it sounds remarkable! Interesting seeing your screenshots - helps to kind of understanding better what you're doing - excellent.

Just listening to Chopin's Berceuse DflatMaj with it, (just now the final chords, quite beautiful), so vital in so many ways, a real delight - I do almost feel the space and the Schimmel piano.