Friday, January 30, 2009

Audio Forensics: Solving a Beatles' chord mystery

Dalhousie University (Nova Scotia) math professor Jason Brown has apparently solved a musical "who-dunnit" - what was the opening chord of the Beatles' famous Hard Days Night? There was an agreed upon suspect - namely George Harrison - but what was the chord? No one could reproduce it, even knowing what guitars they used in performances.

Enter Professor Brown. The professor used tools well known to forensic audio examiners - an FFT (Fast Fourier Transform) to analyze the spectrum and logic. And what did he deduce? It appears that George Harrision had a little help from someone else - perhaps the Beatles' manager, George Martin - who simultaneously played the chord on a piano! So, in the spirit of the board game Clue (Cluedo, for you Brits), it was George AND George, in the recording studio, with a guitar AND piano. Mystery solved!

Here is an article to read more about it. And another one. Enjoy!

(Hat tip to D.E. - thanks!)

Computer Audio: Latest Audacity beta (1.3.7) is released

Audacity 1.3.7(beta) is available for download. This release primarily seems to consist of bug fixes, but they do say that they added support for DirectSound, which is a widely-used Windows(R) DirectX API (Application Programmer Interface). Here is the text from their announcement:

Audacity ( is a fast multi-track audio editor and recorder for Windows, Mac, and Linux/Unix.

The Audacity Team is pleased to announce the release of Audacity 1.3.7 (beta) for Windows, Mac and Linux/Unix. This is primarily a bug-fix release to significantly improve stability and usability, especially on Mac OS X, but includes some new features too.

Cross-Platform Bug Fixes include:

* Muting/soloing caused incorrect channel results in exported stereo files
* Fixes for Nyquist effects, Compressor and Noise Removal
* Fixed Export as WAV could be corrupted if overwriting to same file, and Export Multiple to uncompressed formats only produced 16-bit WAV
* MP3 and WMA now export correctly with all supported metadata

Platform-specific Bug Fixes include:
* Windows Vista: crash opening Preferences
* Mac OS X and Linux: fixes for spurious clipping, label typing, no shortcuts after running effects, and project rate when importing files
* Mac OS X only: fixes for inactive or corrupted menus and hidden dialogues, portable settings, and Command and Control shortcuts. Also FFmpeg installer now available.

New features/other improvements:
* F11 Full Screen mode, high-quality "Sliding Time Scale/Pitch Shift" effect, Audio Contrast Analyzer
* Windows: sound devices can now be opened using the more efficient DirectSound API
* Improved latency correction using fixed correction value

See for details.

Wednesday, January 28, 2009

Image Recognition: Facial recognition feature in iPhoto

I installed Apple's iPhoto '09 last night to test how the face recognition feature (called "Faces", of all things) worked. There are nearly 4,000 photos in my personal library, so I left it running overnight while it went through the database and, I assume, ran its face detection algorithm. I then decided to run a test. I manually identified the same person in two different photographs and had iPhoto go out and locate matches, which is its next step in its learning process - the user next needs to confirm or not confirm the identity of each one of the identifications it produces.

So how did it do at the end of this first step? Surprisingly well is the answer, particularly for a consumer-grade product. The raw numbers were 72 correct and 16 incorrect at this first step, but that doesn't tell the whole story. In the instances where it was incorrect, all but three were confused with family members (which admittedly is the most likely confusion outcome since for all pictures with faces in the database, generally at least one is a family member). In addition, it also correctly chose pictures where only one eye was visible, where the age was significantly different, and where there were different emotions clearly visible on the faces. It also did not seem to have a problem with different compression levels, low resolution (including one that was very granulated), face paint, under-exposure, black-and-white, scans (of printed pictures), and color. Finally, it made only one error out of 86 pictures in correctly detecting and locating the face.

So, even though this was not a properly constructed test, the results were more than impressive enough to warrant looking at it further for professional use in some applications.

PS. This is not a product review, per se, but I should say that my experience with this just-released version was not all positive. Namely, it was horribly slow during the step where one has to select a picture, go into the Faces mode, and then type in the subject's name. Each entry took two to three minutes to accomplish, so I saw the "spinning beach ball" that Mac OS-X uses to let you know it is busy quite a lot. Hopefully this bug will get corrected and patched quickly.

Tuesday, January 27, 2009

Evidence: Listen to the raw audio evidence

If you would care to listen to the audio and read the transcripts of the Governor Blagojevich (IL, USA) wiretaps, here is the link.

Image Recognition: Facial and license plate recognition news items

There are two on-line news articles related to image recognition that I would like to bring to your attention.

The first news article is about law enforcement in Tacoma, Washington, USA using a facial recognition package to match 16 years worth of prisoner mug shots with pictures taken by ATM (Automated Teller Machines) in a forgery and theft investigation to generate the lead needed to solve the case.

The second news article is from New Orleans, Louisiana, USA where local law enforcement used a license-plate recognition system to make 20 arrests and recover 23 stolen vehicles and license plates in a 25 day period.

The common thread here is, of course, automatic image recognition. These software algorithms have come a long way in the last decade. However, one should understand that the conditions are partially or completely controlled in both of these applications - i.e. distance, lighting, exposure, aspect (turned toward the camera), and (possibly) compression all fall within acceptable boundaries. In addition, with the license plate recognition problem, the character set and fonts were known in advance. The controlled conditions and a priori knowledge significantly increase the accuracy of the results tremendously and the chances of a successful investigation and prosecution.

(Hat-tip to JUSTNETNews, USA - I highly recommend this free service of the National Law Enforcement and Corrections Technology Center, or NLECTC, which is part of the National Institute of Justice, USA)

Saturday, January 17, 2009

Photoshop: Most blatant Photoshopped covers

I met a graphics artist today whose job is to "Photoshop" pictures for advertisements. We got to comparing stories and she mentioned an incident involving Kate Winslet, the actress, where her legs were slimmed a bit too drastically and caused a stir. Well, before long we were pulling out my MacBook and Googling for that and other instances of "Photoshopped" covers. Here is the best link we found.


Legal: Recording conversations in the USA

Michael Flynn's podcast, Legal Lad's Quick and Dirty Tips for a More Lawful Life, gives an excellent overview of the legalities of making an audio recording of someone in the USA when you are not a government agency performing lawful surveillance. When is it legal and when is it not? The short answer is: it depends - on whether there was a reasonable expectation of privacy, what state the conversation takes place in, and whether all parties being recorded gave consent.

For instance, if one is having a loud conversation on the street (a public place), how can one expect privacy from being overheard? If one can be overheard, how can he then complain that a record (or recording) was made of what he said? An example from daily life, that we can all probably identify with, is the standard phrase we hear when we dial technical support or customer service - "All calls are recorded for quality assurance." In other words - "We're recording you - you've been warned". By being advised and not hanging up, you will likely been seen as having given implied consent.

Legal Lad, being a lawyer as well as having good communication skills, gives a more thorough and educational overview of this subject.


Sunday, January 04, 2009

Careers: Profile of a Court Reporter

The Wall Street Journal (center-right financial newspaper, USA) has an informative profile of court reporting as a career option. One point from the article that caught my eye was that while training takes only two years and salaries are attractive (at least in the USA), one must learn to type at 225 words a minute with 97% accuracy!