Mario Martinez

Hello everyone, I would like to give you an update to what we have been doing so far: Some people contacted us privately to give us very detailed feedback. The good thing is many of these people have been listening to our calibrations from the beginning (they can play all our different calibrations on their systems). That allowed them to A/B compare all those calibrations over the different instruments to let us know which instrument sounded best on which calibration. These kind of feedback is very helpful for us because it does give us a very good sense of what is going on. We have used all that information to develop a new calibration 2.0b. This new calibration is just a fine-tuning of 2.0a but we think it does make a difference. We have update our web page with this new calibration. I know all these "updating" can get tiresome, but if there is still someone "out there" willing to try I would love to share these new material with you again :)

Back then we had 6 albums that we offered with calibration 2.0 We now have 12 albums plus we have further developed our calibration to try to improve transparency. The new albums have been recorded using a new piano. All 12 albums are now processed with calibration 2.0a.

Thank you Ken for your feedback. I appreciate it The recording setup is exactly the same on all albums (old and new) so none of the differences you hear should be due to different recording techniques. The only thing that is different is the piano itself and its position on the stage. The new piano is a larger piano and it does have a much richer bass. It is also placed up front on the stage so if you were sitting in our hall it would actually be closer to you. It seems like what you are hearing is consistent with what we are doing in the actual hall (and that is a good thing because that is usually not the case)

If the calibration is right, what you see on the recordings should match what you see on the sketches. In other words, all the instruments on all these recordings should be rendered by your systems in a way that their relative position to one another matches the position of the physical instruments on the stage. Why is this important? The "black hole" we were talking about before is the one responsible for the deformation of the spectrum of all recorded material. That deformation does not just affect timbre, it also affects depth, presence, balance, position, etc... If you were to record several instruments (like we did) using a fixed two mic setup with no mixing, you would definitively be able to perceive the effects of this "black hole" on the sound of the raw takes. As of today, the industry does not have a "mechanical" way to correct this (that is what we were talking about on the "timbre solfege" post). There is no known solution to this other than the engineer fixing it "by ear". Our research project is about developing a methodology that will allow us to deal with this "black hole" in a more objective manner, but checking the quality of the results is still something that has to be done "by ear", which is what we are trying to do here with this giveaway.

Here are some sketches that may be useful as a reference on your evaluation. The blue area represents the stage. The person in red represents the mics. The following sketch represents instrument placement for the old albums : - Debussy Preludes. Book 1. - Chopin Polish Songs - Iberia - Songs of Paolo Tosti - Cabrera plays Debussy - Conversations The black piano on the sketch represents the old piano: a smaller size piano placed further back on the stage. The following sketch represents instrument placement for the new albums: - An evening with Riera - From Russia with love - Mozart - Beethoven wind quintets - Goldberg Variations - Ad Astra The brown piano represents the new piano: a bigger size piano placed up front on the stage. And these are the sketches for the classical guitar, drums, flamenco and strings: All these recordings have been done using the exact same setup: same hall configuration, same mic arrangement and position, same recording chain, same calibration. The only thing that changes from one recording to another is the instruments themselves and their position on the stage.

Here is some info that might be helpful in the evaluation of the different calibrations. The purpose of our methodology is not to make each album sound as "good" as possible. If we wanted to do that we would not use one single calibration for all our albums. Instead we would master each album separately to obtain whichever sound we liked best. For example, say we are recording a small piano that has very little bass, we could compensate the recording to make it sound bigger. That is actually the standard procedure in any regular recording project. The purpose of our methodology is to make each instrument sound as "truthful" as possible. That means portraying the timbre of each instrument "as is" with all its good and bad qualities. No instrument is perfect in every way, some have a glorious bass, others have beautiful treble, other rich mids, some sustain wonderfully; choosing an instrument is always a compromise. When we developed calibration 2.0 we only had one piano to work with. We tried to be as objective as possible on the developing of our calibration but with only one instrument to run test with it was easy to be bias by the particularities of the instrument. We were not aware of this when we presented 2.0. The recorded sound was so good we were convinced we had gotten it right. But running further tests made us realize that we could have unintentionally tweak the calibration in favor of that particular piano. That is when we decided to get the new piano. We thought having two completely different instruments placed at two different spots on the stage would help us stay objective. Running 2.0 over this new piano was an eye opener for us. It did not just confirm our suspicions but it also gave us immediate clues as to what was going on. But fixing it proved much harder than hearing it. We spend a long time trying to fix what we knew was there but we did not manage to get it right. Every time we thought we had something we would test it against the old piano just to find out that we were being fooled again. At some point (probably drawn by desperation) we wondered away from our initial philosophy and decided to use two different calibrations (2.0 for the old piano and 3.0 for the new piano) but that would have been wrong. It would have been just like mastering each album in a different way. Back then we convinced ourselves that this was the only way; we even came up with a good excuse that justified the use of these two different calibrations (that excused is explained on the 3.0 thread) But some how (and with the help of some of you) we came back to our senses, we decided we would try it again, and this time we did manage to get something that worked on both pianos.

Here goes another update of how we are doing. We have given out codes to 29 people so far. This are the albums they have chosen: Stories (6) Chopin Polish Songs (3) Conversations (1) Goldberg Variation (2) Ad Astra (1) We are still giving out albums, all we want in exchange is your feedback. Your impressions are very helpful to us.

Which "sound quality" are we talking about? Are you referring to the quality of the sound that reaches your ears? The quality of that sound would be the end product of a chain formed by "recorded material"+"playback system"+"room acoustics". In my opinion, everything that is measurable falls in the realm of "playback system sound quality" and "room acoustics sound quality" and everything that is not measurable falls in the realm of "recorded material sound quality" I would say that all those aspects of sound that have to do with psychoacoustic perception really have to do with the sound quality of the recorded material more than anything else.

Here is an update on how we are doing so far: We have given out 23 codes (Thank all of you for participating on this) About the free album, 8 people have made a choice: Stories (3) Chopin Polish Songs (3) Goldberg Variations (1) Ad astra (1) You can choose your free album anytime, so please let me know when you decide which one you want. The offer is still on, if you would like to try please say so and I will be happy to send you the codes

You are welcome :) It is 75 dB(C) Slow

Here is something you might find interesting... It is a paper on "Timbre Solfege" written by Tomasz Letowski from the U.S. Army Research Laboratory and Andrzej Miskiewicz from the Fryderyk Chopin University of Music. The paper discusses the foundations of "Timbre Solfege" (a course in perceptual analysis of sound) and presents the current version of the course program from a perspective of over three decades of its existence in the sound engineering curriculum. The interesting thing about this paper is realizing that the timbre we hear in every recording is not a direct product of the timbre of the physical instruments, but a product of the psychoacoustic abilities of the engineer that made that particular recording. No matter how far technology goes (better mics, better preamps, better converters) there will always be a "black hole" on the recording chain that will prevent us from capturing the timbre of instruments "as is" That "black hole" is located on the recording chain between the instruments and the mics and, according to this paper, the only instrument we have to restore its effects is the psychoacoustic abilities of the recording engineer himself. But how can we be sure that work has been done properly? There is no objective way to check that this work has been done properly. The only thing we can do (like the engineer did) is use our psychoacoustic ability to subjectly determine if the timbre sounds right to us, but in our case we have not even had the chance to hear the physical instrument. Our research has to do with all this. The purpose of our work is to develop a methodology that will ensure we get the right timbre on the recording without the use of subjective evaluation.

I just updated the OP with the links and info of the SPL files...

Yes, It would be the same as before (like mansr pointed out) You can use the old SPL file if you have it, if not, I will upload a new one so you can download it (because I think the old one is no longer on the server)

yes, we are giving out four albums for each Computer Audiophile member. All you have to do is ask :)

Thank you rando. Do not worry, the codes are the easy part. You are the ones taking the time to listen and I am grateful for that. I am sending you a new code just in case you do want to try. On the old albums (Iberia, Chopin polish songs, Songs of Paolo Tosti, Cabrera plays Debussy and Conversations) version 2.0 and 2.0a are extremely similar. Like firedog said, you can only tell if you listen back to back, otherwise they are almost identical. We developed version 2.0 based on the instrument we had at that time (that would be what we refer to as the old piano) and it did a perfect job on those albums but it did show some problems on the drums and flamenco that we shared with you back then. So we decided to go one step forward and try to develop something that would solve those issues too. It seemed impossible to go any further with the resources that we had so we decided to get a new piano to help us develop this new calibration. So our goal now was to develop a new calibration that would work with the old piano and everything else (the new piano, the drums, flamenco, wind quintet and string quartet) Why would we want to do this? The goal of our research project was to develop a transparent audio recording system (stage+hall+recording chain) So if we can make one calibration work for all this instruments (two different pianos in two different locations on the stage, a string quartet, a wind quintet, a flamenco guitar and a drum set) Not only would that have been a first time in the recording industry, but it would also be empirical proof of the transparency of our system.