Subjective Equalization of Sample Set Recordings?

Presumably samples are subjectively monitored and equalized during the recording and/or editing process to determine an equalization curve which yields a result deemed by the sample set producer to be the “correct” sound for a particular organ.

If this be so there would be two major variables that potentially affect the final sound quality – the response of the monitoring loudspeakers, including the acoustic environment of the monitoring booth, (or the response of headphones if used) and the recording producer’s judgment.

The characteristic of the monitoring loudspeaker (or headphones) would impinge directly on the resultant sound. For instance, if it should have a rising treble response then assuredly the resultant sample set will exhibit a loss of highs and consequent brightness as the producer adjusts equalizers to compensate. This is the case in support of the Standard Monitor Loudspeaker proposed by sound engineers in professional recording studios to minimise variations in the quality of CD’s. Loudspeakers vary enormously in frequency response and distortion content so the effort to produce a “standard” speaker for monitoring purposes is a worthwhile attempt to improve the accuracy and consistency of recordings and “live” broadcasts.

I raise this matter because I notice significant differences between the sound quality of various sample sets. Some are too “bright”, others “dull”. Some are “boomy” (with peaks in the low frequencies), others lacking in bass. For example I find Salisbury Vol.1 slightly lacking in “highs” when compared to the majority of CD’s played through the same equipment. To my ears on my set-up this VPO benefits from a 3db lift at the top end (using a digital equalizer on replay) giving a result that is just magnificent. This in no way is meant to imply that my equipment is “correct”.

Maybe my assumption at the outset is incorrect and there is no subjective equalizing used in the recording of samples. Is the response “untouched” or perhaps monitored only objectively to check the frequency response of components? Is the microphone characteristic assumed to be flat or are measurements made?

I will be interested to hear comments from sample set producers on this aspect. Are subjective equalizers generally used? If so, what monitoring speakers or headphones were used?

I have specifically referred to Salisbury Vol.1 (certainly not to imply any criticism of this fantastic VPO) and would like to learn details of the recording, editing and monitoring equipment together with the techniques used to produce this wonderful sample set.


Max Albiston

Max,

You ask a very good question. I will share my thoughts from the perspective of a sample set producer of virtual theatre organs. There are many differences with sampling theatre organs than with recording classical organs. For example, it is generally not feasible to capture a particular theatre organ as a whole. Instead, for Paramount we assemble the best ranks derived from a number of instruments. Thus there is a need to select and voice ranks that will work well together for a satisfying ensemble.

To your point, I agree that it stands to reason that monitoring be done with a very neutral and accurate audio system which includes room treatment that minimizes colorations. If the end user has a similarly neutral audio chain, then the listening experience should be very similar. When voicing Paramount ranks, I monitor with a single pair of Mackie HR824 (original model) monitors which are something of a legend for their accuracy. The monitors are positioned "near field". The room has non-parallel walls and is heavily damped with Auralex foam panels and bass traps. Samples were recorded for the most part with AKG 414BULS microphones. If a Paramount user played the organ through HRs or other equally accurate speakers similarly positioned in a well controlled room, then the resultant sound will be very close to that which was intended. Of course, rarely will this be the case. Therefore, it is left to each user to adjust the audio chain to provide a balance that is satisfying. I often suggest to those asking for advice to listen with a quality set of headphones and do one's best to achieve a similar tonal balance through the speakers.

Joe Hardy

Joe -

Many thanks for your comprehensive answer which gives me the kind of insight into this process I am seeking. I wonder to what extent other sample producers are as thorough in ensuring there is minimal coloration in the monitoring set-up? I was particularly interested in your point -

If a Paramount user played the organ through HRs or other equally accurate speakers similarly positioned in a well controlled room, then the resultant sound will be very close to that which was intended. Of course, rarely will this be the case. Therefore, it is left to each user to adjust the audio chain to provide a balance that is satisfying.

- which, I think, highlights the problem. This is the need for a user, who regularly plays more than one sample set on his Hauptwerk VPO/VTO, to have different replay equalizations for each if he is to achieve the best results. Of course if the optimum is desired from several organs it is necessary to voice them on a note by note basis, using the amplitude and "brightness" adjustments provided within the Hauptwerk voicing facility. However, it becomes messy when individual equalization curves are needed to counter unusual characteristics in sample sets, probably brought about by colorations in monitoring during recording and editing. Countering the latter can be readily achieved by the use of a suitable graphic equalizer in the replay chain but how can this be simply readjusted for different organs? It can be done, of course, but is likely to complicate replay equipment beyond that desired by many users.

"Standard" monitoring and procedures would certainly minimize the need for individual replay equalization but I think this is a bit of a "pie in the sky", bearing in mind the wide range of sources of sample sets and the corresponding array of recording equipment and procedures, not to mention the difficulty of introducing such a scheme. As an alternative I wonder if it would be practical to include a software graphic equalizer in the "voicing" parameters of Hauptwerk - as perhaps an extension of the "brightness" feature? Perhaps Martin and Brett might like to comment on the feasibility of this.

Max.

[B. Milan]

Hello Max,

We plan to revisit the Hauptwerk voicing facilities for Hauptwerk version 4.1. Whether or not we decide to implement new EQ features remains to be seen. Any real-time effects will ultimately cause a loss in performance, even if minuscule, so there has to be enough reason to implement even more EQ options than are already present in the voicing screens. The current EQ settings should allow a fairly decent amount of adjustment if you haven't tried them already.

Regarding your questions about Salisbury, We used a matched pair of Earthworks QTC-30 microphones with a freq. range of 6Hz to 30kHz and also a matched pair of Neumann KM183s. We used preamps from PreSonus (ADL600 and MP20) as well as an RME Fireface 800, recorded direct to disk on a laptop at 32-bit 96kHz.

We do not perform any post equalization processing since that is not the goal of the projects, rather we capture the organ exactly as it stands and bring that to the end users with as little modification/editing as possible.The samples are given a treatment of noise reduction and that's it, other than a final scaling and balance check once loaded in Hauptwerk.

Having heard the real organ I would tend to think that perhaps any of the CDs you own may have used their own post processing EQ to brighten it up a bit, the real organ is not that bright in my opinion, rather it is somewhat smooth and dark sounding.

If you want the best experience, in my opinion, using a high quality set of headphones will bring you virtually into the church. Speakers of course are fine, however you are then bringing the church virtually into your room. If you want a simple plug and play sound as intended to be heard by the producer then headphones are really the only way to do it accurately without having to adjust and balance settings to suit your listening room.

I hope that helps to some degree.

Thank you.

Brett - thanks very much for all the info.

I can already see vast differences in approach! The Paramount sample set(s) does utilize subjective equalization while the Salisbury does not - and I appreciate the reasons for both techniques.

I now believe there is only one way to overcome differences in response between sample sets (that cannot be satisfied using the existing HW voicing adjustments) and that is by equalizer switching on replay. This is a pity because it is messy but in my experience it is the only way if one is to achieve optimum results on a range of organs.

I agree with your point re headphones. So much has been said on the merits of these versus speakers. However, there are two main reasons why I prefer speakers. Firstly, I rapidly tire of my ears being "clamped". Secondly, l find the lack of directionality - i.e. the ability to sense the source of the organ by the slight turn of the head - disconcerting.

I am looking forward to the release of Salisbury Vol.2. Has there perhaps been a recent estimate for the date of this that I may have missed?

Regards,

Max.

I can already see vast differences in approach! The Paramount sample set(s) does utilize subjective equalization while the Salisbury does not - and I appreciate the reasons for both techniques.

Note the Salisbury is one particular organ recorded at the same time with the same hardware, while the Paramount is a compilation of ranks from different instruments, possibly recorded with different mics and amps, whence it was necessary to use some equalization.