First of all I should point out that I think I now have 4 posts called “Comment spam”. I just got 200 spammy comments in 12 hours. Luckily most of them are coming from the same IP address so they were easy to deal with.

I just bought one of the games I mentioned a couple of weeks ago (actually I bought both but I’m only talking about one right now): Fahrenheit (Indigo Prophecy in America). You should probably read the other post first.

To summarise, the game is well cool. You don’t need fast reflexes or a large amount of skill as such but you do need to be able to think fast. A lot of the time you have a time limit to do something which is three times longer than you actually need - but it includes the time needed to work out what to actually do. Failing to do something isn’t always as bad as you think since the story seems quite flexible: there are times when you realise just what an impact your past actions have on the story.

The story itself has a lot of depth. While trying to solve the murder (alternately controlling the murderer and two detectives) you also have to deal with relationships of the main characters which adds more to the game than you’d think. In fact there are a lot of mundane task to you have to complete that are surprisingly fun.

Graphically the game isn’t that impressive, at least intrinsically. But the clever use of split screen during important events adds a certain style and grandeur to it. The sound for the most part is simple but many moments have an eerie sound track reminiscent of Silent Hill just to heighten tension.

This game will probably appeal most to point-and-click adventure lovers and anyone who loves story over action. But I strongly suggest everyone have a look at it (if you’re in the UK get it from Game; you have 10 days to take it back if you don’t like it :P).

Firstly, after using it for a week I’ve concluded Google Analytics is good. I can’t reasonable access any of it on a dial-up connection but that should be an increasingly small problem. (Increasingly small? Almost as good as “it’s looking more and more less likely”.)

One thing it reveals is random search engine traffic I get for almost no reason. I posted about the April fools gag on GameFAQs and I’m now on the second page of Google if you search for GameFAQs. The funny (although understandable) bit it is that it’s my internal search page for GameFAQs that is actually in the results.

This post, like many of my others on the topic ramble a bit. You have been warned :P

Any regular readers I might have will know how much I like Pimsleur products for learning experience. I believe the basic idea can be greatly enhanced with computers. I did in fact try a short time ago but suffered from a lack of voice talent which is where speech synthesis could be useful.

Synthesised voices can be imported into the current system (which at the moment I wouldn’t be able to demonstrate since the voices need to be licensed if they are to be distributed) just by recording them to audio. This solution would allow maximum support since any browser with audio capabilities could use the system.

There is an advantage to using the IE plugin though (or any other system supporting SALT) - speech recognition. It should be entirely possible to actually have a browser based system that checks your pronunciation which would instantly make it better than any system out there.

Quite why none of the companies that create these synthetic voices have tried to develop a system like this I don’t know…

I just thought I’d provide examples of the higher quality voices available for speech synthesis.

They aren’t perfect but they’re a lot better than the voices that come with Windows XP.

Cepstral voice demos

Ooooh, the blog has been up for a whole year :D

TalkTalk (from the Carphone Warehouse) are offering “free” broadband.

The deal is you sign up to TalkTalk as your phone provider (£11 line rental a month - the same as BT), take out one of the call plans for 18 months (£9.99 a month which includes unlimited* local, national and international** calls) and pay a £29.99 setup fee. For this you get up to 8Mb broadband (very likely to be lower though) with a 40Gb download limit and an ADSL modem.

Total cost (for 18 months): £407.81, about £22.66 a month. Seems like a good deal. TalkTalk

* Calls must last less than 70 minutes. ** To a list of 28 countries including America, Australia and most of Europe.

The speech engine that I was talking about in my last article about speech synthesis is an add-in for Internet Explorer. Follow these instructions to install it yourself (you can get Microsoft Speech Application Software Development Kit (SASDK) Version 1.1 now though).

To actually create speech enabled web pages, you need to use SALT. Speech Application Language Tags is now fairly standard and being supported by Microsoft means it is almost guaranteed to survive in some form. SALT is an XML markup so you would generally embed it straight into a HTML (or more usually an XHTML) page.

The first requirement is to add the SALT namespace to your XHTML document: <html xmlns:salt="http://www.saltforum.org/2002/SALT"> This probably isn’t a requirement but you should do it anyway. The next thing you need to do is to load the add-in and tell it what to handle. This is definitely vendor specific and only applies to the add-in for IE: <?import namespace="salt" implementation="#saltobject" /> All the code does is create an object and then tell it what namespace to look for SALT tags in (in this case the “salt” namespace). There is one potential sticking point. The standalone IE add-in is not the same as the one you get if you install the whole SDK so for debugging purposes so the classid will be different.

After that it’s just a matter of adding the SALT tags for handling the speech. In this article I’ll just deal with the simplest one, prompts. A prompt is just a piece of speech. Just write your speech inside a <salt:prompt> tag: <salt:prompt id="InstructionsPrompt"> Hello. Thank you for using this salt demo. </salt:prompt> The id can be anything you want as it is only used to uniquely identify the prompt.

We now have to get around the semantic separation of form and function: this simply defines a prompt and doesn’t actually do anything with it. To actually make it do something we have to use JavaScript. Prompt objects all have a Start() function accessible from JavaScript so to make it play, just call InstructionsPrompt.Start(). Although it’s not an ideal solution for testing purposes just attach it to the onload event of the body tag. You can see (and hear) the whole SALT demo.

A final note about voices. Windows XP comes with a good voice called Microsoft Mike, but this may not be the default. Some of the others sound really bad. To set Mike as the default: Start -> Control Panel -> Speech -> Text-to-speech Windows Vista will come with new voices (the ones made from sampled sounds I talked about before), one of which is called Anna.

I have content from various ad companies on my site that I get paid for. It’s not much but it can add up and at least pays for the hosting. I realised I was owed some money from one of them that I hadn’t logged into for ages. So I logged in, checked my details were correct, set payment to “PayPal” and left. Unfortunately I changed it over too late. They sent my first payment as a cheque. As I said it isn’t that much; yesterday I received a cheque for $4.33. I’m not even sure I can deposit that in my account - whenever I’ve done it before with foreign cheques there is quite a large charge (larger than $4.33).

Silly company.

I’m currently working out how feasible computer based language-learning-type software using synthesised speech would be. Speech synthesis has improved a great deal recently and although it’s still not as good as a real person (although I believe it could be soon - at least for non-emotive situation) it could just good enough.

The traditional way to generate speech with a computer is algorithmically. Essentially someone works out how to overlay tones with different pitches and wave-shapes to form each sound. The newer way is to actually record each sound manually and essentially play them back one after the other.

There are more stages it to it than that - assuming you don’t want to write the speech phonetically (in IPA for instance) there also needs to be a way of turning text into phonetic information. This is usually half dictionary based for common words and syllables and half rule based (to avoid having a big dictionary and for coping with languages constantly expanding and evolving).

So we now have technology (almost freely available) that can produce speech that is good enough given the correct phonetic information - it’s the actualy language processing that is problematic. Most of the work is done by American companies and therefore most of the work is done processing English (American English at that).

This is not an insurmountable problem. The engine I’ve been playing with (available as an addin to Internet Explorer and as standard on Windows Vista) works fairly well with foreign words transcribed in dodgy-phonetic English. For example to get it to pronounce “Entshuldigung” (German) correctly you need to type “Enshooldicken”). This is workable for an semi-automated system - it could include a dictionary of sorts replacing words with their English-phonetics version.

I know the whole of this article is rather rambling - I’ll post something more readable later :P