.:  Home  :.
.:  Productions  :.
.:  Misc  :.
 Not so fresh news

Not in my name 2015.11.16
----------

They all gone mad, but French politicians and executive are the worst and the most hypocrit. They can only think with their testosterone and take us for fools. They continue using the same means that previously fed terrorism here, claiming like little children that the others started it. I'm definitely not with them. I said it before and I say it again: stop bombing now. Explain yourself about your foreign policy, about your bloody business in Middle-East and Africa. Solidarity with all victims all over the world.

Related: Non Ó la guerre du mÔle poster campain.


Ohm Studio, misc. tool updates 2013.10.17
----------

In case you didn't notice, Ohm Force finally released Ohm Studio, an on-line collaborative DAW (digital audio workstation) to create music. It's something I was involved in for like. 13 years? But the story has just begun.

I also updated some tools and uploaded new ones during the last weeks. trimx264opt, rendezvous, ssa_cps_info, Screenshot Machine and modded things like RemapFrames, DeSpot. I let you find them on your own.


Screenshot machine, fmtconv, Dither 2013.03.21
----------

There were no news for a long time because I'm lazy at this task but things are still going on.

First, Screenshot Machine is a tool to capture color-accurate screenshots from a video, supporting high bitdepth and anamorphic formats (non-square pixels). Video players and workstation can take screenshots from a long time, but I found extremely difficult to produce decent screenshots with minimal display artefacts. This is particularly important for comparison purpose when doing high-quality filtering and encoding. Defects can be rather subtle, so you have to trust your whole display chain to begin with. I came with a bunch of more or less usable informal Avisynth scripts written for my own needs, then shared them on forums or IRC channels, collecting feedback from people. I finally decided to bring everything together and make an official release of it.

This is a bit older, but I also released fmtconv, a plug-in for Vapoursynth. This is a set of functions to convert between formats (resolution, colorspace and bitdepth). Most of the code is taken from the Dither package. Improvements done specifically on fmtconv will be backported to Dither, but later.

In the meantime, Dither was updated to version 1.22.1, what a surprise.

On the audio topic, I plan to update some papers and their related source code (especially the resampling library), and possibly release something new. I've got new material more or less ready for years, but organizing everything to make it usable for everyone is really a boring task...


Same boring title 2012.08.16
----------

Some updates for Dither and AddGrainC.


Dither 1.8 2011.05.15
----------

Dither update with a lot of new features.


Dither 1.7 2011.04.27
----------

Dither update again.


Dither 1.6 2011.04.10
----------

Dither update. Synchronized the modified MVTools with the main branch, improved the doc, changed prototypes and default parameter values, and added a few utility functions.


Dither 1.5 2010.11.07
----------

Another update for Dither. I added three error diffusion algorithms and fixed the noise generator, which was causing not-so-random patterns sometimes, depending on the picture width.


Dither 1.4 2010.09.24
----------

I've updated the dithering tools for AviSynth. They now have SmoothGrad, a smoothing filter working in the 16-bit land (to be used between the noise filtering and the dithering) and GradFun3, a derived debanding function for color-banding removal quite similar to GradFun2db. There were also a few minor improvements and bug fixes.


FFTReal 2.11 2010.09.12
----------

I accumulated significant changes to the FFTReal API over years and had to release an update. Specifically, I fixed a linking error with template specialisations occuring with GCC 4. Please also note that all symbols have moved to the ffft namespace. And just like with my previous other works, I decided to drop the LGPL (not suited to template code) and release everything under the WTFPL.


RectalAnarchy 1.11 2010.09.11
----------

I added the source code of my Buzz plug-in RectalAnarchy to the distribution so it's not really an update. Enjoy!


Dither 1.3 2010.06.20
----------

The dither script for AviSynth has been updated to version 1.3. The main dithering component is now a plug-in (meaning faster processing). I also synchronised my modified dfttest and mvtools2 plug-ins with the latest official releases.


AviSynth's MVTools2 update 2010.06.20
----------

To optimize an AviSynth script for compensating the rolling shutter in digital video cameras, I modified the MFlow function in MVTools2. It now accepts the motion compensation time parameter as a full clip (one value per pixel component) and not just as a scalar. I included the new version in my dithering package, even if it's not really related.


AviSynth tools for dithering 2010.05.06
----------

Long time no update! I just added an AviSynth script to "transdither" a video by denoising the 8 bit source, processing it with 16 bits per component and dithering it back to 8 bits. This should avoid the colour banding effect while retaining as much information as possible on the very low gradients. This is a conceptually different approach of debanding compared to the well known GradFun2db which proceeds by generating artifically the lower bits on the low gradients (with a huge blur + threshold).


New email address, article update and source release 2008.08.10
----------

 Please note that I've got a new email address (to reach me, check this page). The old one @club-internet.fr will be discontinued at the end of the month, my previous Internet provider being stopping the contract.

 The article on number rounding has been updated.

 I also released the Pampurfe's Buzz machine source code at the request of the Zzub developer community.


ay2wav 1.10 update, z80span and ay2sna release 2007.04.04
----------

ay2wav has been updated. What's new?

 Amstrad CPC speaker emulation, for more realistic sounds, with different quality settings,
 Debug mode, to ensure musics are correctly ripped and/or encoded,
 Possibility to fix tonality of early AY3 tunes ripped on CPC but modified in order to work as if they were played on a ZX Spectrum,
 A few bugfixes

I also released two other tools for AY3 chiptune hacking, z80span, a program to report Z80 code coverage, and ay2sna, a program to convert an AY3 file into a memory snapshot for Amstrad CPC emulators.


Captive God 1.01 2007.01.28
----------

I've just released Captive God. It is a help program for Captive. Captive is a game working on Atari ST computers. You can run it on a Windows PC with an STE emulator called STEem. Captive God is plugging into STEem to interract with Captive when it's running.

The main Captive God feature is the path finder, which shows on a maze map the path to a user-selected location. It takes into account clipboard collection and locks to open required to reach the desired position.

The second feature is the Monitor, a small window to be put on the top of your emulator screen. It shows you the direction to take to follow the programmed path, and the password/bar/button codes to open any door. Normally, it will show only puzzle solutions which you have found, but you can force it to cheat and reveal any code.

Captive God is quite similar to and has been inspired by the Super Captive Companion (SuperCC), by Pierre Fournier. A few icons were just lamely ripped from it. Thanks to everybody who contributed to this project (ideas, reverse-engineering, pieces of code, etc.)


ParamEQ source code released 2006.11.09
----------

Paniq is developing an open source Buzz clone called Aldrin. My ParamEQ Buzz plug-in seems to be pretty popular, so he asked me to release the source code in order to have it in Aldrin. So I released the source code under the LGPL license.


Fiche de lecture de « Que la crise s'aggrave ! » 2006.08.30
----------

J'ai publié un petit résumé de ma dernière lecture, « Que la crise s'aggrave ! » de François Partant. Livre très intéressant, ma foi ! Merci à Paul des R.G. pour le prêt du bouquin.


Added AY2Wav utility + source code 2006.07.15
----------

I just completed the AY2Wav utility, something I was developping from a long time and which I never finished until today... .AY files are music files from Amstrad CPC and Sinclair Spectrum computers. They contain music data as well as the code required to play it back on the original computer. This program emulates the Zilog Z80 microprocessor and the General Instruments AY-3-8910/8912 (or Yamaha YM2149) Programmable Sound Generator, in order to render an audio sample file in the Wav format. Currently there is no support of the ZX beeper, so some tune may be silent or not working at all. The program features a high quality resampling system, something which is missing in most AY players or CPC/ZX emulators.

The Z80 emulator code is from Marat Fayzullin, I had to modify it a bit in order to make the I/O functions work correctly. Also, I got special tricks for the AY-3-8910 from the MAME project. YMEmu by Bulba S.V. contained helpful information to distinguish between CPC and ZX because I know nothing to ZX/Spectrum! You'll find a lot of .AY files there.


Update for the article on number rounding 2006.06.17
----------

My article on number rounding has been updated to fix a few typos.


FFTReal 2.0 and Stopwatch 1.02 2005.10.18
----------

 FFTReal has been upgraded to v2.00; the class has been templatised to accept any kind of floating point type for data. There is also a new template class for fixed-length FFTs, giving a significative speed improvement. The license has been changed to LGPL.

 Stopwatch had its interface changed a bit to something simpler.


Game add-ons and Graoumf Tracker 2 update 2005.10.15
----------

 I filled the add-ons for video games section.

 There is also an update for Graoumf Tracker 2 fixing a few annoying bugs (clicks and crashes) and adding some functionalities (instrument selection and the numeric pad).


French translation of the article on resampling 2005.10.14
----------

I translated my article on resampling in French.


Stopwatch library and Captive Clipboard program updated 2005.09.25
----------

 The Stopwatch library has been updated. It now handles correctly CPU processor frequencies greater than 2.1 GHz on Apple Macintosh systems.

 I also added the Perl and JavaScript versions of the Captive clipboard decoder program, thanks to Chris Pile for the port.


French translation of the article on denormal numbers 2005.04.19
----------

I translated my article on denormal numbers in French. The article on resampling will follow soon.


Stopwatch on Linux and DiffEvol in Delphi 2005.04.07
----------

 A user reported me that Stopwatch was compiling and working correctly on Linux for PC. Good news!

 Also, Christian Budde send me a DiffEvol port for Delphi.


HIIR library 2005.03.29
----------

I've just published HIIR, a library for up- and down-sampling signals by a factor 2, as well as producing π/2 phase shift. There was already a piece of code for doing it, but code required several optimisations to run efficiently. HIIR currently includes FPU, 3DNow! and SSE implementations. Anyone for porting it to Altivec?


Stopwatch library 2005.03.06
----------

Another tiny and helpful C++ library, currently for Windows and MacOS. Stopwatch can measure time at the CPU clock level. Very useful to profile functions.


Plitplot library 2005.02.20
----------

During my work on the article on resampling, I wrote tools to output graphs from a C++ program which was testing and measuring the described algorithms. These graphs were in EPS format and could be included straight in the article. I think these tools can be useful for numerous applications, i.e. just to obtain graphic results from a console application without messing with a complex GUI toolkit. Therefore I made a small library from them, now published.


Resampler library 1.02 2004.08.25
----------

Resampler library has been updated, in order to fix the linking problems occuring with new GCC versions (3.3 and above).


Article on rounding 2004.07.04
----------

I've published a technical article dealing with the rounding of numbers on Windows/Intel platforms. Cet article est disponible également en Français. Je vais traduire progressivement les autres articles que j'ai déjà écrits.


Homepage 2004.06.19
----------

Back to the original homepage.


Homepage 2004.04.17
----------

The homepage is now the "official" black screen against the LCEN (also called LEN).


Homepage 2004.01.22
----------

I have changed the homepage with the protest against the LCEN (French law on confidence in digital economy)


Holidays 2004.01.07
----------

I'm back from a very nice week of holidays in Champery (Swiss Alps). Photos on Aurélien's and Boubin's [broken link] pages.


Truncated interview 2003.11.23
----------

In September I was interviewed in an article for the Computer Music magazine. My talk was rewritten to fit in the space and article spirit, I had a look at it and agreed, althought its meaning was a bit altered. So I reproduce below the original questions and anwsers out of any context in order to let people understand my exact point of view:

Q: The Oddity is widely regarded to have one of the richest digital reproductions of analog oscillators and filters; technically, how were you able to capture the Odyssey's richness in your modelling process? Did you need to invent any new technologies or techniques?

LDS: We spend time on doing an exhaustive analysis of the synth, measuring signals and comparing sounds with our models. We also took inspiration from schematics and patents.

Actually there is nothing really new in the Oddity technology, it is more a well-balanced set of synthesis and digital signal processing techniques, carefully finetuned. In the current market, I think audio companies claiming to deliver ultra-revolutionar-unseen-so-far techniques are just pretentious. Most of them just come from academic researches conducted many years ago.

Q: What is it that changed/evolved with computers and music software that allowed you to pursue a software instrument of this richness and complexity?

LDS: Processing power. It was impossible to do it five years ago, but in ten years the Oddity sound could be refered as a rough emulation regarding the synthesis techniques one could commonly use.

Q: Technically, how does the Oddity's Morph time feature work?

LDS: When musician clicks on a new preset, current parameters are memorized and compared to the selected preset to morph to. Then regularly, the Oddity engine computes new positions of potentiometers to make them slide smoothly within the given duration, reaching finally the new settings.

Q: How do you think piracy is currently affecting the software instrument market? What do you think can/should be done?

LDS: Piracy is affecting a lot the software instrument market, although it's quite difficult to evaluate the sold/pirated units ratio. The reason is that software belongs to the "immaterial goods" category, it costs almost nothing to duplicate, unlike hardware. The only costs are packaging and distribution, when they exist. So people believe they don't harm the manufacturer by copying software, and in the strict point of view of the duplication, they are perfectly right. This is a very narrowed view. Development has a cost, as well as support and marketing; legit customers pay for it. If they are not enough, company just goes bankrupt or needs to fire employees. In audio software market, the line is thin.

Piracy enforce companies to be competitive, reduce costs and find alternate way to make customer honnest. Good support is probably the most valuable of them, because it is easier to provide for a small company, giving them a chance against big ones. There are also bad consequencies, as the reinforcement of copy protections, which main effect is customer annoyance and increase of support load.

What should be done ? First, as I said previously, companies can improve their support. They can also reduce their cost, especially marketing, in order to lower the prices. It will give a better brand image because big companies are often criticized for spending all their money in marketing just to sell more and more and make substantial profit. Internet-only distribution is also a way to reduce costs; however people are still reluctant to buy on-line. Second important point, people should be educated about piracy and its consequences, like one does about robbery. This can be included in school programs for example, along with initiation to computers. This is more a political issue, but probably the most efficient in the long term.

In my opinion, the classic business model is unsuited to software and more generally to immaterial goods (such as music, information, etc). I don't have a ready-for-use solution, but there is definitely a need to find something benefiting equally to customers and companies.


Photos 2003.11.08
----------

Stopub action again, great success, photos here.


Photos 2003.10.18
----------

The Stopub group did an anti-advertisement campain yesterday in the Parisian subway, I took some photos, an opportunity to test my new digicam ;).


W32.Sobig-F worm, new email address 2003.08.21
----------

I used to get spam and worms in my mailbox. A lot of junk mail. But I have never seen that before: 2000 junk mails PER HOUR!!! It seems I am one of the favourite targets of the fresh new W32.Sobig-F. Consequently, I decided to change my email address. Check the Misc section.


Nothing new so far... 2003.07.14
----------

Today is the French National Day. The president will speak soon, probably the same boring and useless speech as previous years... One thing for sure, we will be far away from the Citizen's claims and needs (translated in English).


Fed up with worms and junk mail 2003.06.28
----------

Junk mail is becoming more and more annoying. Waking up this morning, I checked my mail and found more than 300 virus and worms collected in one night, in addition to a notification from Club-Internet (my ISP) that my mailbox is full! This is the first time I get such a notification, but I fear it could be more and more common in the future if things don't change. I am theoretically protected by an anti-virus on the ISP server, but it seems it doesn't work very well... W32.Bugbear.B, probably 95 % of my current worm collection, is royally ignored. I tried to mail my ISP several times about the uselessness and malfunctions of their anti-virus, but technical support service seems to be animated by robots replying with standard mails. Shame!!!


Resampler update 2003.06.23
----------

Resampler library was updated. Macintosh port and miscellaneous compilation glitches fixed.


Resampler library 2003.06.21
----------

I have just put online Resampler v1.0, a C++ implementation of the technique described in my recent article. It is a free library, ready to use in any software synthesis project. Check the production page. Most code is portable, but some snippets are architecture- and compiler-dependent. If you happen to port Resampler (should be quick and straigthforward) to a system different from Windows/MSVC++, please send me the code in order to make it available in the next releases.


Article on resampling 2003.06.20
----------

Finally! I have just published my article on resampling. Azeteg from #musicdsp suggested me to add something about the cochlea response and other psychoacoustic considerations related to filter pre-ringing. Unfortunately I don't have time to conduct extensive search in this area now, maybe in a future paper update. Thanks to Lee_Dc for the rereading.

I'm going to release in a few days some source code as basic implementation of the article's method.


Article 2003.06.15
----------

My article on resampling is almost done. I'll make it reread by a few people before publishing. I also need to tweak a bit more the source code illustrating the techniques described. It will be published along with the article.


Book 2003.06.09
----------

I've read the book Jee lent me, Against Method by Paul Feyerabend. Very interesting work about epistemology and history of sciences, but sooooo hard to read. My brain hurts!


Music 2003.05.19
----------

Added some very old tunes - soundtracker modules - to the production section. Please do not laugh !


New article in the pipe, book, photos 2003.05.06
----------

I'm currently in the process of writing another technical article on signal processing applied to audio and music software. It will be focused on a cheap and very efficient technique (good quality vs CPU load ratio) for resampling a signal at a variable rate. Unlike my previous article, there will be some nice graphics for the response curves, misc comparisons, etc. Being unsatisfied with my current tools, I've developped a small program to draw these graphics in EPS files, with the help of the fine Olli Niemitalo's EPSPlot C++ class.

I've finished to read the book Ummo, de vrais extraterrestres (Ummo, real extraterrestrials). Very worth the read. This deep analysis of the Ummite language may change your mind about the Ummo case, if you happen to know it. Unfortunately, most available documents related to Ummo are in French or Spanish, only very few were translated in English.

I've also uploaded the photos taken during my week-end in Les Contamines. Outdated film + mediocre photographer = terrible pictures.


Nice week-end 2003.04.27
----------

I've spend my Saturday and part of the Sunday on graph theory. I need an algorithm for data-flow directed graphs and I'm pretty sure it already exists... OK I need first to find a precise model for my problem, this is eventually the trickiest part.

We had a good barbecue at D&D's, with the Evil Cats. Some photos here [broken link].


Reconstruction filter 2003.04.23
----------

After the controversed Iraq war, the controversed Iraq reconstruction. Yesterday I read in the newspaper that Mr Jay Gardner will be the temporary civil administrator in Iraq. The paper discussed the relevance of its position at the head of Iraq, mentioned some facts about the past of JG - retired general and weapon industry leader - and a link on a website (http://www.stopjaygardner.com) managed by his disparagers. I tried to connect, but the site was down. The domain name was even not registered nor available. This site never existed and everything is disinformation ? Wrong. There are still some pages in Google cache at the moment I'm writing these lines. What happened to this site ? Closed for abusive slandering ? There was nothing offending about Mr Gardner, just facts on his public life. This is probably what we call freedom of thought and expression.


Snow time ! 2003.04.16
----------

Long week-end at the montain. After the recent snowfalls, Aurélien, Bibi, Francky and I decided to rent a chalet in Les Contamines (French Alps). It was really nice, we had good snow, sunny weather, and our chalet was just in front of the Mont Blanc ! I've taken some photos, I'll upload them soon.


Denormal article on CiteSeer 2003.03.25
----------

CiteSeer, the scientific literature library, has now accepted to feature my paper on IEEE denormal numbers.


More source code 2003.03.09
----------

A lot of new-old stuffs from my dusty folders: Synthesis of alias-free analogue waveforms, 3DNow! biquad filter, WAV file splitter, AVR to WAV file converter, DiffEvol, SharedPtr, MS Visual C++ macros, Degas Elite picture file converter and a school schedule generator. Have fun !


Articles and source code 2003.03.02
----------

I added my article on denormal numbers and some interesting piece of code in the Production page. More to come soon.


Buzz machines 2003.02.16
----------

I uploaded my Buzz machines on the Production page.


Website started 2003.02.09
----------

My free.fr account is now set up and I have just uploaded the first elements of this website. Pages are still missing, they will come in the next few days.



[Home] [Productions] [Misc] Website by Laurent de Soras, 2003-2008