Mario Martinez

Hello 88Keys, this is Mario from Play Classics. It should be an automated process. After you make the donation PayPal should have taken you back to the page of the album. At that point, the blue buttons on the page should read: "TRT sound master - 24bit 96kHz flac" "MP3 files - VBR 48kHz" You can go to the page of the album now on the same computer where you made the donation and you should be able to see these buttons. Just click on the button to download whichever file you want (you can download both) If for whatever reason you do not see them please let me know and I will send you a gift code so you can download the album. Thanks, and sorry about the inconvenience.

As of yesterday, all albums have been updated to versión 2.4c. The main reason of this last revision is the feedback provided by @rando concerning the soprano album. I only have myself to blame for this, but I think I may have been "not paying attention" to the effect that the various calibrations have had on the voice albums because I was using the instrumental only as a reference. 2.4c is the result of that same analysis including both voice albums (soprano and tenor). It should be an improvement on every album (voice and instrumental)

Yes, sure, I just sent it to you. Thanks.

Hi Henry, thanks for the feedback. I sent you a code to download the "Romantic Piano" album. I think you are listening to version 2.4b on both albums (the Mozart and the Schumann). After my last post on October 22nd I started trying something out. I got good results so I decided to update all our albums like we always do (...I know I said I would not update anymore) This new version (2.4b) has been up and running in our servers since last November 12th. Anything downloaded on or after that date would be this new version. I think Henry is the only one that has done that. (you would know for sure because the tracks are tagged with a date and a version number) I understand how "tiring" all this updating can be but, should anyone want to try it, everyone is welcome to update their albums to version 2.4b.

I am sorry you lost the 2017 files. I wish I could replace them but we no longer have them in our servers. I am thankful for all the feedback (positive and negative). As a research project I just try to keep and open mind because I know everyone has their own reasons and perception to backup their comments and it is all useful to us. I understand that you might like the 2017 files better than the current versión but that does not mean that they are more transparent. Our project is not about making a "good sound" but capturing color and balance as transparently as possible no matter what the final result end up being like. Back in 2017 we were on the 1.x versión of our "TRT sound" calibration. That calibration was developed using albums 1 through 7. All those albums were recorded using only one piano (the one we had from the start), two voices (soprano and tenor) and a classical guitar. In 2018 we purchased a second piano. We spend 3 months looking for the right one trying to find an instrument with the best posible harmonic balance across its complete range. Using our old 1.x calibration on this new piano made it clear to us that we had been unconsciously deviating our calibration to compensate for the smaller bass of our first piano. So, no matter how good those recording sounded we felt that we had to correct them because we now knew they were not completely transparent. It took us a few years of tweaking to get to a calibration where both pianos would render as what they really are in real life (showing their good and bad aspects). There is no perfect piano, some have a better treble, other better bass, it is always a compromise. Getting these two pianos to sound as what they really are helped us improve the transparency of our calibration. As a result the new versions of albums 1 though 7 sounded less heavy on the bass and the calibration also worked with the new albums which now included the new piano, cello, bassoon, french horn, clarinet, oboe, flute and violin. All these things also had an effect on the voice. On the old calibration Sobotka's voice was surrounded by an artificially lifted up bass that made it sound "smoother" and that may have created a "nice" or "beautiful" effect but just as it happened with the piano it was not as transparent as it is now. About the artists, they are aware of our constant development and give us constant feedback on our calibrations. This past summer we talked with Mr. Cabrera (he regularly accompanies both artists Sobotka and Corujo) and he assured us this new calibration was the one that best render the true qualities of both voice albums (soprano and tenor). Anyway, I am not trying to justify anything. Like I said, I understand that you may like the old version better but we just could not keep it that way knowing that it was biased.

Here are some samples I have made to try to illustrate this. They are not actual recordings; I have generated them in the computer to simulate what actually happens in the real world. The purpose of these samples is to show how mixing into one track the sound of two microphones placed at different distances from an instrument artificially alters the timbre/color and balance of that instrument. This sample is the sound that we are going to use as the source material, it would be the real sound of the instrument if you had it in front of you. It is an “A4”. It is composed of 5 harmonics. The color you hear is determined by the structure of its harmonics (power and inharmonicity): Instrument.mp3 This other sample is the sound we would get if we were to make a recording placing one mic 7.8 centimetres further away than the other. 7.8 centimetres is half the wave length of the 5th harmonic. That means that by the time the sound reaches the second microphone, the first one is going exactly the opposite direction with regards to this particular frequency. So when you put them together at the mixing table this frequency practically disappears: 7.8 cm apart.mp3 This would be the same thing at a distance of 13 centimetres therefore eliminating the 3rd harmonic: 13 cm apart.mp3 And the same thing at a distance of 39 centimetres therefore eliminating the 1st harmonic: 39 cm apart.mp3 As you can see, none of these recordings have been able to capture the true color of the sound of the original note. They have all altered the timbre into something else. The first one made the note much darker, the second one made it much more nasal, and the third one made if much thinner. But the harm does not stop there. If you were to use the third setup (39 centimetre) to record every single note of the range of this instrument, what do you think would happen? If you are thinking that everything would sound thin then you are in for a surprise. The 39 centimetre setup will cancel out the 440Hz frequency of any sound no matter where that frequency appears within the harmonic structure of the note. For some notes (like our “A4”) that means it will be cancelling the 1st harmonic (therefore sounding thin) but for others, like a “D3”, it will be the 3rd harmonic thus making it sound nasal, or the 5th one (for “F2”) making it sound dark. At the end you will end up with a collage of different timbres distributed along the range of the instrument that not only are incoherent with each other but also untrue to the real color of the source.

Here is a picture of a recording session in one of todays mainstream classical studios: And here is a picture of the setup we use to record every ensemble here at Musicstry Studios: The first picture is a very good example of what is regularly done today. It seems to me that engineers tend to think that the more mics the better. It does look more "busy" in the picture so I guess it also makes it easier to sell the job to the next musician/customer. The truth is, it is not better, it actually complicates everything. All those temporal incoherences you were talking about are cause by this. When you have so many mics recording the same instrument at different distances you are setting yourself up for that temporal trap. The sound that comes out of the instrument arrives to each one of those mics at a different time. If you have mics at three different distances every time the instrument produces an attack you will pick up that attack at three different points in time. What do you do with that at the mixing table? The people that use these techniques will tell you that if the distance is not big enough then the ear cannot perceive them as different attacks. That might be true, but what it certainly does not do is make the attack any cleaner. But this is not the most important side effect. To me the worst part is how this affects the timbre/color and balance of the instrument (which is the part that is almost always ignored). When you put together the sound of two mics that picked up the same instrument at different distances you are going to have phase problems that are going to affect the whole range of the instrument in different ways. You can try to "fix" one part of the range of the instrument by sliding the file of one of the mics a bit backward of forward on the software but you will be messing up even more other parts of the range of the instrument. This kind of problem cannot be fixed, so the only solution is once again the same as before. Just as with the attack, the "theory" is, bring up the sound of a set of mics that was placed at a similar distance and use the others at very low levels. If you add them to the mix at very low levels then the human ear cannot perceive these phase problems. But why would they want to do that? Their answer will probably be: ...to enrich the sound. Really! ...so you are telling me that a grand piano needs his naturally gorgeous sound to be artificially enriched? Maybe the problem is that their main set of mics was not able to capture the richness of the natural sound properly to begin with. Had they been able to do that they would not have needed any of the other mics.

There is a Spanish cellist that just recorded an album with Sony. I have heard the album to check the technical aspects of the recording. It does suffer from that common effect that makes it seem as if the cello and the piano had been recorded in different spaces and then put together in the mixing table. Plus it seem they have manipulated the dynamics of both instruments separetly placing them at different floor levels with different levels of compression. I would like to contact this artist to offer him the possibility of recording his next Sony project in our studios. This would be a great opportunity to introduce our sound into the market. But if I call him and tell him that we can do a better job he is just not going to believe me, and if I tell him to listen for himself he might not know what to listen for. I do not want to disclose the album info here into the open, but if anyone wants to listen to it please send me a PM and I will send you a link where you can stream it. I would like to know just how obvious you think the difference is, and if you are ok with it I could use your testimonies to try to convince him.

I might be too much of an idealist but, once you get to the point we are at, it wouldn't seem so farfetched to me: famous artists recording their albums in our studios to release them with their usual labels. I believe that an innovative breakthrough of this magnitude should be able to make its way into the industry on its own. After all, this methodology eliminates the middle man. It allows musicians to take control of the sound of their productions. It allows listeners to evaluate the music that the musicians do without anyone’s interference. And being transparent, it produces a natural sound that (at least to me) sounds better than anything else. These are all benefits. But the recording industry is extremely hermetic. You can not just pick up the phone and talk to someone on a big label, or a famous artist, or an important journalist. All these people are completely inaccessible for a small company like ours. We have tried it many different ways and we have never gotten any response. Everyone wants endorsements, the journalists, the artists, the labels. The only endorsement we could possible get would be from the people that have taken the time to listen to any of our albums, …and that is mostly you in this forum and a couple other audiophile forums where I have also posted during these past years. The way I see it, it is up to you to decide whether you want first line artists and big labels to record their chamber music and solo albums with our methodology. If you do so, just speak up, try to convince other forum users to try it out, if we get enough people involved we might be able to get the industry’s attention. Imagine we made a poll asking this and everyone on the forum participated. If we had a thousand people saying yes would not that be enough to make the industry listen?

May I ask you all a question about something that I do not understand from my perspective? I understand that everyone on this forum (there must be thousands of people here) is interested in high fidelity audio reproduction. I also understand the relationship between better gear - more accurate reproduction. But it seems to me that most people have forgotten that reproduction is just one part of the chain. Accurately reproducing a bad recording is never going to make it into a good one. My question is: have people forgotten that? Are they not aware of this? I sometimes go into other threads to see what people are recommending just to listen to other recordings and compare their production work with ours. Some of the things I encounter are quite astonishing and I am only talking about classical music. People praise recordings that are completely artificial. The type that seem to be most popular is the violin piano type where you could listen to the whole album to find yourself asking at the end: was there a piano playing in it? I can imagine what the pianist himself must have thought when he was fist presented with the results of that recording. I do not know if the cause is a lack of ever having heard the real thing or may be people just find that artificial product more enjoyable. But there is obviously something going on and I feel that I am not getting the grip of it. I strongly believe that our approach is the right one, but seeing how little repercussion it has I feel that our project may be completely offline with the general public. Music lovers from classical music forums will not even try to listen to our albums because our artists are not “famous” plus they would not want to get their ears contaminated with anything other than recordings from big labels, and audiophiles (in their majority, not the 15 of you) are to busy talking about cables and gear to pay attention to the quality of the material they are going to feed their systems with. What is going on? …am I missing something?

I am having a bit of trouble understanding this statement. Could you explain how it works a bit more? Yes, of course, lets see if this makes sense: In the real world volume goes from the ground up. Lets say we have a room with a piano in complete silence (that would be 0dB). Then we play a soft passage (at around 40dB) and then a loud one (at about 70dB) In the recording world the volume goes from the ceiling down. Had we recorded the previous example we would have ended up with a file like so: a headroom (from 0dB to for example -1dB), the loud part we played (from -1dB to -31dB) and the soft part (from -31dB to -71dB). At this point if we wanted to play back the recorded sound at natural levels we would have to use a system powerful enough to produce 71dB at full volume (or use a more powerful system and bring the volume down to adjust) Now, lets bring “real” conditions into the picture. The room is not perfect (no room is). It has standing waves that resonate in certain frequencies amplifying the power of the signal significantly on some parts of the spectrum. So when we played the loud part in the piano that was supposed to produce 70dB the mics actually picked up peaks of 85dB. Now our recording would have: a headroom (from 0dB to -1dB), the loud part (from -1dB to -46dB) and the soft part (from -46dB to -86dB) To play back this file at natural levels we would need a system powerful enough to produce 86dB. Or the other way around, if we used the same system as before (the 71dB system) we would ear the file 15dB softer. What happens then if that level turns out to be too soft to be acceptable? We would need to bring the volume up by 15dB. But we can only bring it up by 1dB (the headroom that we had) because bringing it up any higher would result on peaking at certain points on the recording. What should we do then? The proper thing to do would be to work on the acoustics of the room to try to bring the effect of the standing waves down as much as possible and record again. But you are never going to get rid of them completely. Say, of that initial 15dB boost, you are able to get rid of 10dB but you are still getting peaks amplified by 5dB into the recording. You would still need to figure out a way to reverse the effect of those 5dB in the recording (that would be the work that our calibration is supposed to do) If this is done successfully then you would end up with a file like the first one we described: a headroom (from 0dB to for example -1dB), the loud part we played (from -1dB to -31dB) and the soft part (from -31dB to -71dB), where the effects of the standing waves would have been eliminated recovering the proper volume and the original color/timbre and balance of the instruments. But this is never done this way in comercial recordings. The usual solution to this problems is the use of dynamic range compression. They simply compress the file from a dynamic range of 85dB to a dynamic range of 70dB. That way they can bring it up by 15dB without peaking to match the floor level of the first file (-71dB).

Firedog also inquired about this same issue: And this was the answer: In oder words, this is how we usually proceed: first we apply the same equalisation to all the albums with the sole purpose of restoring the timbre quality, then we evaluate the resulting peak level of the whole set of albums (all the albums regardless of the instruments except for the guitar album) to figure out how much can we raise the volume of the whole set without affecting the dynamic range of the interpretations. This way we manage to have the same "floor level" for all the albums (except for the guitar which is 6dB higher) while retaining the full dynamic range of all the performances. This final calibration has enabled us to bring the volume level higher than previous ones. But that higher volume level is not a result of dynamic range manipulation. It is a result of a different timbre calibration (or equalisation). In order to demonstrate this, I have uploaded a version of the "El Corpus en Sevilla" and the "Poulenc Sextet Allegro" with no volume adjustment just the timber calibration (or equalisation). The resulting file is 4dB lower than the 2.4a file. You can compare these two files (I have label them -4db) with their corresponding 2.4a version just by lowering 4dB on the 2.4a The -4dB and 2.4a versions should produce the same dynamic range in your room while yielding different results in the DR meter software. I am sending these files to firedog and Klassikmann, if anybody else wants to download these files please send me a PM.

For some reason the link to the samples is not working. I cannot fix it right now, I will let you know when its ready...

I already uploaded the samples in Master format. I am sending them now so anyone that is curious can try this approach... Edit: its done. Everyone should have the album they chose plus the samples. If I left anybody out please let me know. Enjoy 👍

Here is an update on the situation. We have given out 15 albums so far. We have also sold another 9. Thank you for contributing to the project! :)