Recording sound 2

Which part of the sound is essential?

It is very important to understand what part of the sound is essential and what is not. Obviously, you have to record the direct sound, but in most studios you do not record the first reflections. What is recorded is the direct sound and the "too early" first reflections (1-5 ms). Well, that is a problem and most pop-recordings sound very unnatural and have to be re-mastered. What you should record is, in fact, the direct sound and the early reflections (5-20 ms for small ensembles and for big orchestras 5-80 ms). 

If you record too much early reflections, your recording will appear to have more reverberation than it really has, but it will still be clear and open. If there is too much of the late reverberation your recording will sound distant, thick and unclear.

Note that the early reflections cannot be created in a good way in any reverberation insert or box, as these devices cannot create enough early reflections as needed to make the sound believable. If you add reverberation to your recordings – and most people do – you should add only late reflexes! Your insert or box should have the capability to add just these and no others. Note also that most inserts do not let you adjust this important thing! If you add early reflexes to your recording which already has early reflexes, the sound-image will deteriorate and the sound will be less clear.

Also to consider are the "too early" first reflections that are between 1-5 ms and which make the sound unclear. Unfortunately, this is how the mike usually is placed in most pop studios around the world. I think it has to do with the live recording situation which is transferred to the studio. In a live situation, you need to place the mike very close to any instrument for obvious reasons, but in a studio, you do not have to do this. A better way to place the mikes for a singer will be described in following pages.



The A-B and X/Y stereo recording

There are only two general ways to record. They are "A-B stereo" and "X/Y-stereo". 

The X/Y method uses the relative levels of sound to generate a sound image. You always need to use cardioid mikes for this type of recording which gives a very steady and clear sound image, but has the problem of the proximity effect described on the previous page. (M-S and Blumlein stereo recording are variants of X/Y and are also level dependent).


   X/Y-stereo with cardioids.


The A/B-stereo presents the sound image with the help of the time differences between the two mikes. It yields a sound image which is less clear than the X/Y-method, but the final presentation of the sound image over the speakers will have a wider listening span; that means that you do not have to sit exactly in the middle of the speakers to get a good presentation of the sound image. Usually two omni-directional mikes are used, although cardioids can also be used. The French radio uses a special variant of this system which is called ORTF. Recording with a dummy-head is another variant of A/B-stereo.  


A/B-stereo, to the right ORTF, 17 cm and 110 degrees


Example of artificial heads for recording



Recording with A/B-stereo



A/B-stereo with the mikes 145 cm above the floor and 70 cm from the instrument, the mikes separated 28 cm. This gives, in my mind, the best blend between the direct sound and the first reflexes. The late reflexes are not dominant and can easily be added afterwards, and are useful if the hall is small or sounds bad. The mikes are placed in the "turn" of the piano, where the grand piano's radiation of treble is at its strongest value. With a full concert grand piano (273 cm), the mikes should be placed a bit further to the right to get the best balance of the sound. 

In a humorous vein, I call these recordings: OPPR recordings or Omni-directional Pianist's Perspective Recording, which they truly are.



A/B stereo recording is shown in the pictures above and below. The setup employs omni-directional mikes, with the mikes at a slight angle from each other (the ORTF variant). This is, indeed, the easiest way to capture the direct sound and the early reflections and at the same time has the least effect on the frequency curve. With cardioid mikes you will affect the frequency curve with every adjustment of the mike's position relative to the piano. With omnis you do not affect the frequency curve at all, just the balance between the direct sound and the first reflections


This was a setup for a recording of a CD with music of Mozart. It was recorded during early 
2005 and you can buy the CD from Lovestream records at:


Stig Jacobsson in "HiFi & Musik" writes in December 2005:
"He plays very transparent and clear, with a beautiful recorded sound". 


This is how the front cover was designed. 
Remember the Omni-directional Pianist's Perspective Recording? OPPR!

At the download page, you can listen to the two last variations from Twinkle, Twinkle Little Star 
to evaluate in this context, the sound and also a tuning-test for you of the grand piano!



This is a picture from the recording of the Chopin's Waltzes in 2009. It was made in the same hall in Järfälla, Stockholm, and on the same Yamaha Concert Grand CF Piano as above. I was the recording engineer, piano technician, tuner, made the digital editing, and was co-producer with the pianist Johan Westre. The cover is below. 

It is a "OPPR 24 bit recording" (Omni-directional Pianist Perspective Recording). With this type of recording setup (omnis), you will get a very clear and open sound. The bass will be extended to beyond the lowest notes, and the treble will not sound hard or unpleasant resonant. It will be almost as "sitting at the keyboard yourself". 

I made an test of another brand of mikes, but stuck in the end with my AKG 414. The recording was set up in the same manner as the Mozart CD above, as I was happy with the sound from that CD. The mikes was moved only a few centimeters to the right, to slightly improve the balance of the bass and treble from the piano. So, if you have a good tried out set-up and measured the mikes positions carefully, you would only need to move the mikes around a few centimeters at the session, to get the "best" sound picture. No more fiddling around with the set-up.


The new four channel Edirol R-44 worked exemplary during the recording. I didn't quite trust the internal mike-amplifiers, so I used the Mackie mixer. But, the sound is of excellent quality from the recorder. It's completely noise-less, non-mechanical with no fans, as it records on SD memory cards. Much better than sensible DAT-tapes.



This is how the front cover was designed.


More recordings and common errors and mistakes with omnis  Page 3