Workflow for solo piano recordings

Hello Ian,

thanks a lot for your prompt response!

Yes, I generally agree on this. Kawai's ATX2 has sensors on the hammers (instead of beneath the keys) which - in my impression - allows for a very smooth mapping of the pianists playing to midi values.

Addressing your comment: I can consistently produce midi values between about 3 and 110 on the keyboard. Comparing this to the internal sound of the ATX2 (sample-based) 3 maps to ppp and 110 to fff. Inbetween these values the keyboard behaves more or less linear in my impression. I did modify the velocity curve appropriately - this is what I meant by "slightly modified". I'm sorry I might have not been clear enough about that.

Maybe I'm completely wrong here, still I'll try to explain why I reduced the volume:

If the volume is near to say 0 dB one might get clipping - or the limiter interferes if it is activated (which I have deactivated). So reducing the volume gives more "headroom" as I understand it. Taking into consideration that I use an audio bit depth of 32 bit for export it shouldn't matter whether the volume is reduced to for example -10 dB or -20 dB as long as it is enough to avoid clipping - after all the dynamic range is somewhere around 192 dB when using 32 bit. Afterwards one can normalise to e.g. -0.3 db.

I mentioned several different things I tried and experimented with. I don't generally use compression. I tried this because I hoped to get a more present and loud sound, just as in the first two videos I linked in my first post. I don't think it's generally a bad idea - quite a few of the presets in pianoteq actually apply some compression.

I got the idea from several articles on piano recording. You might be right that it has no effect, I will check this. But I think it should have no negative effect.

Thank you for your input. I'm looking forward to further recommendations.

Cheers,
Daniel

You're welcome! :-)

Thanks for the suggestion, I'll give that a try.

Figuring out the microphone positions seems a good starting point. It would be great to be able to talk to the people involved with the recordings. ;-)

Is there anybody else reading this post who has ideas on how to achieve these unique piano sounds in the three recordings with Pianoteq?

Best regards,
Daniel

Nice ! You confirm my suspicion: That popular reduced, strong and dark sound with distinct treble many people like sample libraries and recordings for is almost certainly the result of processing. The interesting thing is that many engineers are able to achieve a similar result. I've played more grand pianos which sounded rather like Pianoteq by default, though: more complex. But I'd like it as an option. For now we have to do without that kicking pedal noise. To me Old Man River is most similar to Pianoteq by default.

EDIT: I have to differentiate a bit. Although the sound in the video at McPartland's also has that dark character and kicking pedal it is, in contrast to the others, obviously very processed maybe to sound vintage. That effect is a combination of compression and the proximity effect due to close micing and maybe the use of vintage mics and tape saturation, I guess. Fits to the piece but is not natural.

Thank you Modellingoptimist! I like that sound, very dry and close, I wouldn't say overly processed... very saturated, and a bit muffled, but that could be the sound of the studio. Maybe it was recorded to analogue...

You're welcome . I think you're one of the best tweakers here. BTW, I changed "mellow" to "dark" some minutes ago. That's what I mean.

Hello,

the initial posting of this thread could have done by me - that's what I thought when I read it first.
I would even had choosen the Einaudi video as reference for the sound I'm looking for. As I just started to play with the various available options in PT, it'll be more a long term challenge to get close to that .. at least on a basic level.

Roberto, I'm impressed how accuratly one can match a given sound with all those settings! Really unfortunate you picked the third and not the first video as reference

Beside the 'bunch of flowers' I wanted to leave an quite interesting link here. It's about the mic settings and used equipment for one of Einaudi's concerts (In a Time Laps - Burmingham). The sound is very similar to the concert in the Royal Albert Hall. There are very bad quality clips of it on youtube (not worth to be shared) but one gets an impression of the general sound.

Here the sound engineer reveals at least a part of the secret two Schoeps MK41 are used to mic the piano (in the 'completely inside the music' section).
http://www.tpimagazine.com/production-p...lapse.html

So, the article refers to two Schoeps MK41 super-cardioid mics. I wonder if those are equal to the CMC6MK4 modeled in Pianoteq?
Of course this is just one single element.

Bernd

Haha I wanted to listen to the first video but it was taken down due to copyright issues... Do you have a reference audio file?

Thanks for the link Bernd, I'm downoading right now... I'm in school until 10 pm so I'll only have a chance to listen and try to aproximate that sound tomorrow... as for the dynamic range, I think it's a natural property of how the sound propagates. When you're close to the instrument the dB range from ppp to fff is compressed compared to listening from a distance, where the ppp gets almost inaudible. The timbral palette of the instrument though is even wider when you're close, ppp is misterious and very dark (but not lacking in certain areas of high frequencies, something I'm missing in ptq's ppp) through the various stages of mp, mf, f up to the roaring fff...

Roberto,

many thanks for sharing. Really appreciated!
So far I had the chance to listen to your example only - which is very promising. Will spend some time over the weekend so see how it feels playing your preset.

Bernd

hey all,

I'm in a similar boat, getting ready to record some solo piano works of original compositions and arrangements.

I'd like to keep things as simple as possible.  I have  a controller feeding into Pianoteq 5 Std on my Linux laptop via Jack, routed through to an external firewire soundcard (Echo AudioFire2).  I have a small 8 channel mixer I can hook up somewhere in that process, mostly after the soundcard to feed the studio monitors and headphones (so I keep everything neutral EQ'd) so I have some physical volume knobs.

In terms of workflow, I'm thinking of just
1. recording some solid midi performance files directly in Pianoteq using whatever sounds best during performance / recording
2. play back the midi performance in Pianoteq and change the models, mic placement, etc so I get something that sounds good for listening
3. using whatever settings I have settled on for #2, generate the audio in Pianoteq and route it through Jack into Audacity or even something simple like Jack Time Machine for the final capture, with minimal postprocessing (mostly, downmixing into two stereo channels if I end up with more than two signals coming out of Pianoteq).

Thoughts?

Good question! 

1.  I didn't know that I could just export an uncompressed WAV directly from a midi (i.e., process the midi data into audio using the Pianoteq engine without actually having to go through realtime playback and capture first?). 

2. I wonder whether routing the data through the firewire interface vs the internal soundcard has any impact on sound quality in this case (I know it does for mic/instrument signal capture, but if Pianoteq is creating the audio internally based on a series of mathematical calculations of midi data, then it seems we would end up with the same or similar recording quality regardless of the other hardware in use?)

3. Maybe internally generating a WAV file can allow sound production at higher quality rates than real-time given hardware limitations? IE, make time rather than CPU the dependent variable (e.g., to allow higher polyphony, more mics, etc --it would just take longer to process vs creating a process error from CPU overload)

4. I'm interested in exploring multi-mic setups, and wonder whether there's a significant difference in capturing, say, 4 mic sources as independent channels and then downmixing to stereo vs capturing 4 mic signals into 2 stereo channels.

Thanks!

So, to clarify, "let Pianoteq mix for you" = assign any/all mics to either/both of the stereo output channels, and just don't bother increasing beyond the two channels unless I plan on doing some processing to a specific microphone/channel independently before the final downmix.

Also, to clarify -- are you saying you tend to use one of the mic presets without fussing much with them before rendering the audio file?

So here's my workflow modified with the information you've provided:
1. Find/create a preset ideal for performance
2. Record the midi performance
3. Find/create a preset ideal for playback/listening.  Keep the output to stereo (2 channels), no matter how many mics are used.
4. Export the midi performance directly to WAV using the highest quality settings

Thanks, this is really helpful!