Virtual piano comparison

navindra Hmmm, is this an improvement for you?

Improvement compared to what?

David B Your previous criticism regarding shrillness.

navindra Hmmm, is this an improvement for you?

Well, I think it is an improvement, although not enough but it is. What did you change? However after 1:20 the short scales that I do slightly higher still have that shrillness.

CyberGene Interesting. The EQ3 effect was boosting the higher freqs, so I just set it to flat (I guess I could have just disabled the effect entirely). I also decreased the Hammer hardness at Forte from where I originally had it at 1.28 to 1.08, making the tone mellower at higher velocities.

navindra I remember when I was beta-testing one of the Pianoteq versions (Grotrian, if I remember correctly). I would do exactly the same comparisons as this one and Philippe and the team would take it into account and would attempt to fix it in the next beta version. However we didn't know what they did, it was just an iterative approach and it was always they fixed something I didn't like in a version but then something else was worse elsewhere... I've been a little disappointed by that and is when I lost my faith in Pianoteq. I guess it may be possible that what Modartt do behind the scenes, if exposed to more people in an extended Pro version would be better. However they probably do the offline recalculations and whatever their secret sauce is.

navindra Your previous criticism regarding shrillness.

I'm not sure if it's an improvement. I'm not a fan of enchiladas no matter how my wife prepares them. 😀

God Bless,
David

I think the boxy sound was a conjugation of wrong setting of render volume with Reaper and lack of normalization. I redid the renders, and since the original file uses low MIDI values, I now used the linear MIDI velocity in the Synchron player. If there is "nasalness" here, I have been calling nasal to something else...

Among all of these, I like the original PTQ from @CyberGene which I have been able to reproduce with my demo from his MIDI. The original Garritan is a close second.

All the rest are quite distant from that sense of presence, power and delicacy of these two. And many have a rather questionable timbre that I don't like, particularly the VSL! The Fazioli is fantastic for being a free library, but I like PTQ and Garritan much better.

Actually, I initially tried rendering the MIDI file with PTQ Classic (rather than "Classical Recording", see the description in CG's post) and it sound horrible, that setting is definitely the worse of the group. Since everybody can do that rendering on their own with a PTQ demo, I'm not bothering to post it, but I encourage everybody to do a comparison between the two… For this MIDI file that setting is unacceptable IMHO (so much for all PTQ instruments and settings sound all the same: to me the best and the worse are the same PTQ instrument with only slight differences in settings…. Does anybody know, perhaps @navindra, what the differences are between the two?)

RIP I don't have any insights other than what they advertise. Classical may be better for a player perspective and Classical Recording for an audience perspective.

Personally, I started from Classical Recording and created a player perspective from it for my day to day usage.

Classical: This preset offers a BA micing perspective with a classical sound.
Classical Recording: This preset is recommended for classical recording. Equipped with 4 mics: BA micing plus 2 ambience mics.

The differences are fairly extensive, howeve:

navindra I don't have any insights other than what they advertise.

Thanks, even just that is more than I knew 🙂

navindra Classical may be better for a player perspective and Classical Recording for an audience perspective.

Frankly I don't hear much "difference" from that point of view in PTQ. The mono/recording/binaural/stereo do have a difference between them, but it's not as dramatic as e.g. recordings of Matthew Lien on YT when a binaural mic. Perhaps we can start a separate thread on this topic.

navindra The differences are fairly extensive, however:

I see, that's quite a long list and of some changes that are pretty dramatic for the timbre.

But you did not comment on my main claim:

RIP Among all of these, I like the original PTQ from @CyberGene which I have been able to reproduce with my demo from his MIDI. The original Garritan is a close second.

RIP I initially tried rendering the MIDI file with PTQ Classic (rather than "Classical Recording", see the description in CG's post) and it sound horrible, that setting is definitely the worse of the group.

What do you think about this rating? I know you must prefer yours, but do you hear such a large difference?

FWIW, I re-listened to it now, and as usual my impressions of PTQ change from one day to the other… Not as dramatically as in other cases for this test, but I would not call the difference so total as I did yesterday. I relistened to some of them and still VSL is definitely not my favorite, it sounds very synthetic to me 🤯

RIP What do you think about this rating? I know you must prefer yours, but do you hear such a large difference?

Among the Pianoteq options in this thread, I still like my Bluebird preset the best -- surprise! 😃

To satisfy your curiosity, I also tried to play with the Classical preset and I hated it. It is very loud and very direct, but it feels unrefined to me in terms of tone/control/dynamic range. I'm not sure why Modartt made it the first preset, but I didn't spend any time trying to fix it either as I have good presets already.

FWIW, I re-listened to it now, and as usual my impressions of PTQ change from one day to the other… Not as dramatically as in other cases for this test, but I would not call the difference so total as I did yesterday. I relistened to some of them and still VSL is definitely not my favorite, it sounds very synthetic to me 🤯

I don't think you can replay a MIDI against different VSTs (or even presets in some cases) and make truly valid comparisons.

Ever since David's videos, I've admired the Bosendorfer Imperial and Modern U but not the 280VC (as it just sounds like Pianoteq). I'm happy with Pianoteq U4 as Modern U equivalent, but I'd like something like the Bosendorfer Imperial in Pianoteq as well.

The Bosendorfer Imperial is quite special in my opinion.

navindra To satisfy your curiosity, I also tried to play with the Classical preset and I hated it.

Thanks, good to know it was not just me

navindra I don't think you can replay a MIDI against different VSTs (or even presets in some cases) and make truly valid comparisons.

Yes, I agree. But since we are making such a comparison anyway, I am posting my impressions 😆

I preferred number 2 for its less bright tone.

Also, I thought number 2 had too much extra sound outside the fundamental tone, perhaps resonance or harmonics.

Both pianos seemed like they were played with a lot of pedal.

navindra Ever since David's videos, I've admired the Bosendorfer Imperial and Modern U but not the 280VC (as it just sounds like Pianoteq). I'm happy with Pianoteq U4 as Modern U equivalent, but I'd like something like the Bosendorfer Imperial in Pianoteq as well.

I've owned every VSL Synchron Piano library at one time and the only one I find myself playing everyday is the Bösendorfer Imperial. Each one has a unique attribute that I appreciate, but the BI is the only one I feel has a more complete package. It's nice to have options and it's great to find something that inspires us. Tonight I switched to the N1X internal CFX sound and it only took a few minutes to realize how disappointing it is compared to the VSL BI.

God Bless,
David

vagfilm I think the boxy sound was a conjugation of wrong setting of render volume with Reaper and lack of normalization. I redid the renders, and since the original file uses low MIDI values, I now used the linear MIDI velocity in the Synchron player. If there is "nasalness" here, I have been calling nasal to something else...

I have forgotten about that. Just listened to the two files and compared them to your first versions. I think something has been totally messed up the first time because the new versions are totally different. Your original versions are almost mono like, not sure what's causing it but the new ones are night and day, nice stereo, wide.

I still sense the Steinway's nasal character but it's much better this time. To be honest, Pianoteq has the same nasal type of mids, so maybe Pianoteq is good in emulating this type of Steinway after all? 😱

For the sake of piano science, and since they are a bit forgotten in these comparisons, I asked a friend to render the Cybergene "test-pieces" in two VSTs that I am very curious and always tempted to buy: the Bechstein Digital and the VILabs Ravenscroft.
The Bechstein was a complete disappointment, too shrill and jumpy. I am not posting it because I think the VST is better than that and it is simply an unfortunate mismatch.
On the other hand, the Ravenscroft works pretty well for this type of playing. You judge...
Note: I had to further compress the mp3s to meet the 2Mb limit.

vagfilm For what it's worth, the Bechstein Digital can have a pretty hard/pointy tone. It's a good instrument, but it's not a forgiving one IMO.

@CyberGene I am a big fan of Pianoteq, but a bigger fan of Garritan, so I was surprised that I preferred the PT rendering here!

vagfilm Not bad, although I miss some ambience and rich resonance. I guess Ravenscroft would be ideal for straight ahead jazz and modern music though because it’s very clean sounding.

CyberGene I guess Ravenscroft would be ideal for straight ahead jazz and modern music though because it’s very clean sounding.

This!

It somehow sounds like a mix between a harpsichord and a piano (in a good way).

P.S.

I don’t always drink beer (samples), but when I do, I drink a Ravens!

With only twelve calories (layers), Ravens is both satisfying and good for your health; unlike that monstrosity, VSL, packing 100 calories per key, and still, I somehow feel unsatisfied.

P.P.S.

Ravenscroft is the little engine that could, and it further proves my point: the only room for improvement is transitioning to modelling; because more layers does not for a better VST make. of course, I’m not saying that there is no improvement between 1 layer and say 12 layers, but once you go past 15 layers it’s overkill!

12 or 15 is likely to be the threshold after which it is hard to hear a velocity gap between layers (if the layers are mixed correctly of course). With Pearl Concert Grand, we have 8 layers and there is a very slight gap between layers. We have to be careful to hear them.