Anyway, I got quite a bit of work done on the song over the last few days.  It’s up to just over 2 minutes long now, and most of the themes are written and in.  For the song structure I need to finish out the rest from where it currently leaves off.  I also need to re-balance the levels, and start doing some real mixing work – the current is pretty raw.

You can listen to it here…

So, I’ll be real glad to hear feedback on the song so far!  Thanks for listening!

Today I recorded him singing “Easy Lovin” by Freddie Hart, and “After the Lovin” by Englebert Humperdink.  I then worked on the tracks in my DAW (Sonar X1) for a couple of hours… you can listen to samples of the results:

After the Lovin:

Easy Lovin:

The setup

I recorded using two tracks in Sonar – one for the music and one for the vocals.  The first thing I did was to set the volume levels on the tracks to make a decent mix.  Then I started working on the vocal track.  Using Sonar’s ProChannel strip, I added compression to even out the volume using the PC-76 W-Type compressor. This is a nice and warm compressor which handled the vocals just right.

Then I added Cakewalk’s SpecraFX processor to put a bit of a chorus and stereo expansion on his voice.  This plug-in actually does a lot more than just that, but that’s all I needed to do and it sounded nice.  After that in the FX chain is the Sonitus:fx Reverb processor.  I used it’s default reverb preset, and tweaked it just a little bit.

For the music track, I added Cakewalk Studioverb2 to , and dialed it down pretty low – just enough to give a little sheen to the music to match the vocals better.  I also turned on the PC-76 compressor in the ProChannel, and EQ’d a wide Q around 183Hz down a little.

This song is a hard one for Tommy to sing cold because of the higher range.  Thanks to Sonar’s V-Vocal plugin, I was able to take a couple of spots in the song and (ever-so) slightly adjust the pitch of a note here and there.  Made a lot of difference, though!

Finally, I went through the track a few times, tweaking the volume of the vocal track (recording the automation) to make sure to compensate for spots that were a little too loud or soft.

I copied the track settings to the recording of “After the Lovin”, tweaked a few settings, and that song was done very quickly.  I’m very pleased with the results!!

Check it out – leave comments!

The studio centers around Cakewalk Sonar X1 Producer.  This comes with a lot of built-in effects, processors, and instruments.  There are so many included sounds, loops, and sequences that it will take me months to go through and hear them all!  The workhorses include Cakewalk’s Dimension Pro synth, Rapture LE synth, and Z3ta+ synth.  There are also several other instruments, such as Pentagon I, Square I, GrooveSynth, TTS-1, DropZone, and TruePianos.  Dimension Pro also comes with a lot of built in and extras, such as the Garritan Pocket Orchestra.  There are also professional mastering plugins, which make for great sound.

I’ve started downloading some free VST instruments and effects.  I’m a big fan of Triangle I, and I was just introduced to Glitch and look forward to playing with it soon.

My computer is a 4-year old HP Pavillion home computer, with a dual core CPU and 2Gb of RAM.  Enough for my basic projects, but not enough for really aggressive compositions.  At some point I’ll have to upgrade.  Sonar can manage 32 or more cores, so I’ll probably build out a custom machine next year.

In the computer I have an E-MU 1212M PCI audio interface.  This is 2 PCI cards providing 2 audio inputs and 2 audio outputs with mastering grade 24-bit192kHz ADATs, and on-card DSPs providing 28 built in hardware based effect processing.  Their PatchMix software allows me to create many ASIO ports and route signals within the hardware very easily.

Plugged into the audio input I have my Roland Fantom Xa synthesizer/keyboard.  I can record the audio from the keyboard with nearly no latency.  For driving the keyboard from my computer, I have an M-Audio Midisport 2×2 USB-MIDI interface box.  This provides 2 independent MIDI in/out connections – I have one going to the keyboard; the other is currently unused.

In addition, I have a Korg padKONTROL (mine’s white), plugged in via USB.  I use this mainly for doing initial drum patterns, triggering loops in Sonar’s Matrix View, and mapping some of the knobs to Sonar’s mixer console.  Sonar is able to map the controller to many more functions automatically for better studio control and automation, but I haven’t delved into that yet.

Finally, there are my M-Audio Studiophile BX5 studio monitors.  These things are excellent.  The BX5 has basically been replaced by the BX5a, but mine are in great condition and, except for the blown capacitor which I repaired, they haven’t failed me at all.  Their sound is powerful and clear; I couldn’t be happier with them.

I want/need to enhance my instruments – I love Native Instrument‘s Massive, and want to also get Absynth, FM8, and Kontakt… but that will have to wait till I have actual disposable income.  I will be purchasing a couple of accessories soon (headphones and a small ASIO-capable audio interface – wish list here), so I can work without making a lot of noise in the house.

So that’s about it for now.  Now, back to work on music!

Well, I’ve decided I don’t really want to do that.

There are several reasons for this.  The main one is that once I’ve set up the expectation that I’ll write about some particular topics, then I feel the pressure to deliver.  I don’t want pressure to write a blog – I want to just write what I feel like writing when I feel like writing it.  And, I want to just write about whatever I find interesting right now… not what I found interesting 3 months or 3 years ago.  Besides, the number of people that would be interested in reading those kinds of posts is pretty small, and they probably don’t read this blog anyway.  Also, detailing out an interesting software design or coding solution feels more like writing documentation than writing for fun.

Another reason is that a lot of the interesting things we did and do might be considered proprietary, trade secret, or might prematurely publicize company information.  I don’t want to have to worry about that.

However, this doesn’t mean that I can’t or won’t possibly write about those topics when I feel like it!  Also, if I have a reader that is interested in a particular topic, I’m much more likely to write about it… so suggestions/questions/interest is welcome.

Right now, what I find most “interesting” is what I’m doing outside of work with creating music.  I’ve created a category in this blog for posts about what I’m doing there, as well as a top-level menu for that category.  So for the time being, that’s probably more of what you’ll see from me here in this blog.  But there will be posts from the work-front as well… I like to crow about accomplishments!

So now… onward!

The studio monitor repair went well – I ordered the part (a 6500 microfarad, 25 volt capacitor) from Mouser Electronics.  It arrived 2 days after I ordered it (the part was $2.02, and the shipping was about $4.50).  I soldered it into the circuit board, re-assembled the speaker, and tried it out.  No more humming and crackling!  Sounds good as new.

Once I had the speaker working again, I installed my new DAW software: Cakewalk Sonar X1 Producer.  It was a 16GB download that took all of last week to fully pull down and get into my machine.  But now it is installed and working.  I set up my system and my synth (Roland Fantom Xa) in a corner of my room and set it all up… and it’s up and working now!

Since I’ve had it set up, I haven’t had much time to devote to any music work yet.  I started slicing out some clips from Never to send to some guys who will be doing a remix.  I laid down a few bars of a piece I’ve had brewing in my mind for a couple of decades, just to get started with something… but that’s about it.

One of my M-audio BX-5 studio monitors developed a hissing sound. Looks like one of the big capacitors blew. Pretty easy fix. $3 is better than $300!

In other news, I’m upgrading my DAW software to Sonar X1 Producer!  Very excited about that.  I was using Sonar 6 Studio, but a virus hit my system and I had to re-install the operating system, and lost my install CDs for Sonar.  The upgrade was extremely affordable with my “invitation only” loyal customer discounts.

I’m fixing/upgrading my music system because I’ve been in touch with a very talented music production duo out of Sweeden who really liked my song “Never” and have agreed to produce their own remix/version of the song.  Early ideas sound very promising!  I’ll provide more information later when the project progresses.  But first I have to get them a remix pack, and to do that I need my DAW… hence the X1 upgrade.

When we started this, I considered building our system on a traditional platform: J2EE.  Having worked with it before, I knew that it would provide a lot of the basic needs our systems required such as threading, lifecycle management, request dispatching, etc.  I even built an initial rudimentary server using JBoss… but was soon struggling with the vastness of the APIs and the overhead of the application server itself.  What I needed was a platform tailored for fast transactionality, easy scalability, very high availability, and a simple API.  So I created one.

Now, 3+ years later, we have a mature base upon which we build our services and product offerings.  Our platform is tailored for processing financial transactions in a secure environment and designed for massive scale with “five nines” availability.  In addition, we are constantly improving the architecture – thinning the transactional layer for lower latency, decoupling systems for better reliability and more flexibility, simplifying and partitioning the database for easier management and improved usage… and more!  And as cool stuff happens, I’ll write about it here!

Of course, the platform – cool as it is – is really just the foundation of our products.  We also have developed a set of services which can manage millions of individual users in a multi-tenant environment accessing a plethora of product services.  Users can identify themselves to the system using a variety of identification and authorization methods, including biometrics, and securely access their personal information and perform secure financial transactions.  Furthermore, the system manages a network of end terminals which access these services, and is able to deliver, add, and remove new services to the terminals “over the air”.  To this end, we’ve created terminal client software which runs on a self-service financial transaction machine (a kiosk) which interacts with the user via an intuitive touch screen-based GUI and various devices such as card readers and cash dispensers.  The terminal software is extremely flexible and can take delivery of new product services (screens, logic, and data) on-the-fly without on-site interaction.  In addition, the terminal software is designed to be “hardware agnostic” – i.e. the core is not tied to any particular hardware or operating system or device.

So, that, is my system in a nutshell.  Future posts in this series will dive deeper into these aspects as well as some interesting techniques and solutions we used to build it all.  Let me know if you are curious about anything in particular, and I’ll try to write about it!

David conquers Goliath.

You see, NCR machines run Windows, and all device and machine integration is done through a layer called “XFS“.  This layer is implemented as a Windows DLL, and all interaction with machine devices is accomplished by sending Windows messages through this API.  Of course, the API has C++ header files and Windows messaging requires the use of native pointers to structures defined in these files – all things which are normally inaccessible to the Java world.

Normally.

Thanks to two tools, JNAerator and JNA, I was able to build a layer upon which I can create Java APIs for each and every XFS device, thereby opening the world of XFS-based terminals to our platform.  This is a Really Big Deal!

Again, this may sound to nebulous… so for instance, consider the “ATM Service”.  In our “ecosystem”, this service delivers access to a customer’s debit accounts via ATM debit transactions.  Our self-service kiosk-based terminal platform provides this service through screens which provide traditional ATM functionality, and drives the devices in the kiosk to do such things as accept card swipes and dispense cash.  All of these services live in various places in various systems on various platfroms, loosely coupled enough for me to not consider it a single system.  Also, since the systems live to provide transactions with economic value attached, I like the term “ecosystem” (unlike others, but that’s a debate for other venues).

We’ve designed and built our systems to be “services oriented”, but not in the way you hear people typically speak of SOA (although much of our platforms do in fact follow the SOA pattern).  When we use the term “service” in our platform, we are talking about one of two things: a running piece of code which is invokable within the platform (such as a “user login service”), or an end-to-end product service which can comprise multiple components (such as “ATM” which consists of the ATM terminal service, the ATM transaction service, the ISO 8583 service, the database service, and so forth).  One of the things that makes our architecture so powerful is that we can deliver new services to the endpoints (the kiosks) on-the-fly.  Most, if not all, current terminal platform technology in existence today can not do that.

So… now that I’ve somewhat introduced what I’ve been building for the last few years, I’d like to talk about what I would like to talk about in this “diary of a startup”

Building platforms and applications, especially in a transaction processing environment, presents many challenges and opportunities.  We have systems which run the range from interactive GUI applications to web-based portal applications to online transaction processing to integration, and everything in-between!  In building this stuff, we’ve done some pretty cool and interesting things, and that’s what I’d really like to write about.  Here’s a tentative list:

• Home-grown annotation based AOP
• Fast in-memory SOA
• Clustering for performance and high-availability
• Building a distributed test harness for load and stress testing
• Command line JMX console for non-GUI environments
• A pattern for request-reply messaging in asynchronous multi-threaded multi-connection environments
• Our Ruby-Git-Maven based development/build environment
• Write-ups on some third party tools and libraries that we use
• And more!

I might be tempted to also write about some less technical topics, such as my experiences going to India to build our offshore development team, my thoughts on development practices and methods, software architecture, and other such things.  I would be glad to entertain requests for topics as well, so feedback encouraged!