Windows DLL (intel) - contributed by Michael Rich.
MacOS 7.5.3 (or later) - contributed by Petteri Kamppuri.
FreeBSD 3.3 (intel) - contributed by Michael Gottwald.
NetBSD/pmax - contributed by Joachim Buss.

The Compaq Alpha NT DLL has also been slightly updated, everything worked as it should, but the version number displayed was wrong, so Michael Rich sent me a fixed version which I've put online as BladeDLL-091-alpha-2.zip.
 

A few other minor things have also been changed, take a look at the Changelog for the details.

The trouble I've had with my computer was that I decided to buy a new harddrive (20GB) and install all of Linux, BeOS and Windows NT so that I could more easily take care of those versions of BladeEnc myself. Unfortunatelly I ran into some trouble, the NT disc was slightly corrupt so I haven't managed to install that yet, BeOS refused to boot until I found out that I for some reason needed to disable IDE DMA in my BIOS (but after that it worked perfectly) and I've been trying almost half a dozen Linux distributions before I finally settled for Caldera OpenLinux 2.3 which suited my current needs better than the rest. That means that the precompiled glibc-binary now is made under OpenLinux 2.3 and I don't know how that affects compatibility with other glibc based distributions. Please tell me if it works on your system or not. As always, the static version should work on all Linux systems.

I've put Linux i586 and BeOS x86 R4 binaries online. The Windows version will have to wait until I get my hands on a better install CD or somebody provides me the binaries.

I've also been told about problems to get BladeEnc to work on some architectures. That seems to be because we made a design misstake when including one of the optimizations for 0.88 which takes for granted that floats are organized in a certain way (position and size of sign, mantissa and integer parts). I should have known better, but didn't think about it when including that patch. We will fix that for the next release and in the meantime I've put the source for 0.82 online as well, which should work on all machines.

But now it's time to celebrate year 2000... Happy New Year everybody!
 

Windows version can be downloaded from my download page. No need to update any other version since the bug was in some Windows specific code in bladesys.c.
 

Also, my appologies to the UPX developers for jumping to conclusions. It obviously wasn't their program that caused the problems.
 

The most serious one was that BladeEnc 0.88 for Windows crashed on some machines, most likely because I had compressed the executable with UPX. UPX is a utility for compressing executables and I thought it might be a good idea since BladeEnc 0.88 had grown quite a lot in size because of the inclusion of a very big optimization table. Unfortunately there seems to be a bug in UPX (at least the Windows version), making it crash BladeEnc on certain machines. Thanks for all the bugreports I've received, I'll forward the most detailed ones to the UPX developers, hopefully it will help them fix this bug. BladeEnc 0.89 is distributed uncompressed.

The other one was that the -rawbyteorder=[] switch didn't work.

New Windows and Linux i386 binaries are online. The Mac OS port and Windows DLL were never affected by any of these bugs, so I see no use in updating them.
 

To subscribe, please go to http://lists.fix.no/mailman/listinfo/bladeenc-dev.
 

Source code and Linux i586 binaries are online and I'll put the Windows binary online later today or tomorrow (need to go to work to get access to a Windows machine). Other versions will appear when I get them from contributors (with possibly a few days delay). I have removed the Linux i386 binaries since I don't think there is any use for them. Is anyone really compressing mp3 files on anything less than a pentium? And in that case, is the i386 version really faster?

Also, as you allready must have discovered in order to read this: BladeEnc's homepage has moved to http://bladeenc.mp3.no. MP3.no were (among others) kind enough to offer me free web space. I chosed mp3.no since I got some good contact with them and feel most comfortable to be associated with another project based on non-profit voluntary work. Their goal is to help young Norwegian artists to get attention, please head over to their homepage at http://www.mp3.no if you can read Norwegian.

Moving the homepage to mp3.no is the first step in my attempt to turn BladeEnc into more of a community effort and less of the one-man-show it's been since the start. At mp3.com we can set up mailinglists, run our own CGI-scripts and give more people access to updating the homepage. People who are interested in helping out on a more regular basis are welcome to contact me. We need programmers, a webmaster (for redoing the whole infrastructure of the homepage), people who keeps track of the frontends and update the pages accordingly and people who can do quality tests on experimental unstable versions.
 

Windows, Mac OS and all UNIX variants allready supports it.  

Many thanks to Marco Amrein for coming with the idea on how to find and fix it. Looks now like it was totally harmless, but it made the output file slightly different so I couldn't be sure until I had found it. Also thanks to Vitor Sessak for providing another fix, but Marco's sollution was more right on spot.

Ok, now I just need to get the "long filenames getting truncated on certain Windows machines"-bug fixed and we'll have a new stable version.  

If you're a programmer you are very welcome to download the source and try to help me. Detailed info about the bugs can be found in the file TODO, included with the source. Both are very strange and I actually suspect that they are bugs in GCC and Windows respectively, but they need to be worked around anyway.

The sooner we can have these bugs fixed, the sooner we'll get 0.88 out the door.  

I've also put the latest unstable source (0.87) online. Unfortunatelly gave the vast speed optimisations of 0.84-unstable a decrease in sound quality. I've been trying to fix it, but after many hours of searching I could still not pinpoint the cause, so I decided to go back to 0.83 and merge in all the changes I know are harmless. 0.87 is therefore quite slower than 0.85, but it's still faster than 0.82 which is the latest stable release and generates the same quality. The best new features of 0.87 are probably the config-file and ability to encode raw samples. More details in the CHANGES-file in the source archive.

I think it's about time for a new stable release now, so please download, compile and test 0.87 if you are interested and report any bug you might encounter. Unless some serious bug is encountered we should have 0.88-stable this weekend. Just need to polish up a few things and update the manual first.  

Joachim Buss have provided us with binaries for DECstations running NetBSD/pmax 1.4. Both R3000 and R4000 binaries are available and he told me to make clear that these binaries would not run on SGI-mips machines, only DECstations. Actually, I've never heard of that computer before, got to be quite uncommon. However, it's nice to see BladeEnc on yet another architecture. :)  

I've made quite a number of changes in the code during my stay in England. I haven't put it online yet, but 0.86-unstable should be up later tonight or tomorrow. I also think it's really about time to make a new stable update, but I really need to fix some things first, like a slight quality degradation compared to 0.82, so it might still take some time...

As you've probably noticed I've also moved my homepage to http://hem.bredband.net/tord which is the place provided by my new ISP. It's really about time since I now have a quite spiffy 24 hour internet connection (have an effective transfer rate of more than 400 kilobyte/second!) and the old one had to be updated through a modem connection for security reasons. So I guess I'll be updating the pages a little bit more often from now on. :)

If you have bookmarked www.bladeenc.cjb.net or bladeenc.cjb.net you won't need to update your bookmarks since those are virtual adresses which automatically will transfer you to these pages later on. However, if you have bookmarked http://home8.swipnet.se/~w-82625/ you should definitely change it to http://bladeenc.cjb.net. Don't bookmark http://hem.bredband.net/tord since that is a static adress and these pages are quite likely to move again before the end of the year.

Some minor updates have been made to these pages. David Jamroga has been so kind to read through these pages and correct any spelling errors he could find. Thanks! I've also updated the FAQ a bit (so I guess there are some new spelling errors *grin*), but the rest of the pages (quality, frontends, speed comparisons etc) are still quite out of date.

In the source archive you can now also find the source for the DLL-wrapper.

Some new ports of 0.82 have also been provided. An Irix6 port has been submitted by Thomas Svedberg, an MS-DOS port has been done by Matt Craven and a new OS/2 port has (as usually) been provided by Mikael Kjellström.

I have some more ports in my inbox which I'll put online later... 

It was originally converted from the C-Headers to Delphi by Jack Kallestrup who released it on the net. Unfortunately it contained an error which caused memory leaks, crashes etc. which Sinan now has fixed and sent to me in order to make a working version available to anyone.

I've also updated the source a bit and released 0.85-Unstable which you also can find in the source archive. As with all unstable releases, don't use this for any serious encodings since it might contain new bugs affecting the output.

It's now quite improved compared to 0.82, so I think it will be time for some serious testing and a new stable release when I get back from England.

Well, now I better get this online and start to pack my bags. It's 21:27 and I'm leaving 12:00 tomorrow...
 

I haven't managed to make any arrangements yet that will allow me to update these pages or work on BladeEnc while I'm on the islands, so it's quite possible that I'll be completely cut off from any work on BladeEnc until I get back. :(

I'll be coming home a few weekends during this time, but then I will probably have no time for BladeEnc, so don't expect any updates for some time to come...
 

Changes described in the CHANGES file in the source archive.
 

Thanks to Marc Schefer for the BeOS binary, Steven M. Schultz for the BSD/OS binary and Rich West for the one for LinuxPPC.
 

The Solaris ports (both SuperSparc and UltraSparc) have been recompiled using Workshop C 5.0 and that way gained a notable speed increase.

New Solaris ports were provided by Serg 'Ice' Tsyganenko and the Alpha stuff was done by Lennart Börjeson. 

I've also put up a new version of the source code (0.83-unstable, although i'm quite sure that it's as stable as 0.82) that isn't ready for mass consumption yet, but those who wants to play around with it are welcome. Changes described in the CHANGES file in the source archive.
 

The Windows binary has been done by me and the Solaris binaries have been provided by Marcos Theophylactou.

I've also fixed a minor compilation problem and put up an updated version of the source.
 

Source code and linux-intel binaries are immediately available and a windows version should be uploaded by tomorrow.
BladeEnc 0.82 also contains some other small fixes, provided by volunteers. See the changelog at the download page for full details.

An unknown number of volunteers have been working on optimizing BladeEnc and I have now received two very interesting patches that are claimed to speed up BladeEnc notably, so you can expect some speed improvements for the next version.

Just need to take a close look at them and make sure that they don't affect quality in a negative way before I include them...

This one gives the result that the first frame header (and possibly some others further down the MP3 file) gets lost. This might sound serious, but in most cases you won't notice anything. The sound quality is still the same except that it might skip a few milliseconds of music when a frame header gets lost and another few milliseconds might be of a slightly lower quality. This slight difference will in 99% of the cases be impossible to hear.

However, this bug is dangerous in another way since the lack of that header makes the bitstream and the headers further down the line get slightly out of sync. This is automatically corrected after just a few frames, but the frames that are affected will have some incorrect data in their headers which might confuse the player. All known players handles this nicely and goes on playing the bitstream without any trouble and some of them (for example mpg123) also prints a discrete error message.

However, there might be either present or future players (software or hardware)  or tools that won't be able to handle these slightly screwed up mp3s and that's the reason for this warning.

I don't see any need for anyone to recompress any mp3's generated with BladeEnc 0.81 unless you run into trouble, but do advise everyone to either downgrade to BladeEnc 0.76 or wait until 0.82 is released (which should be within just a few days).

You might wonder how this bug managed to sneak in since I always compare the output of a new version with the output of the previous one. The reason is that I made these changes on a laptop without my ordinary development tools over the Christmas holiday. Since I didn't have a tool for comparing two binary files accessible I took the quick route and just compared the last 10-20 bytes by hand. Since the encoding of each frame is depending on the result from the previous one we normally get a chaos effect where even extremely small changes in the process generates totally different frames at the end of the file. But since only a few bytes in the headers had changed and not the bitstream we didn't get this chaos effect. The last bytes of the files were identical anyway! :(

Like all the other serious bugs it was introduced during the vast internal changes that were made for BladeEnc 0.80, but I didn't get aware of it until after 0.81 had been released. It was reported to me within 24 hours, but I mistakenly thought that the effect was caused by another bug that was fixed in 0.81.

0.80 had some SERIOUS bugs:

• Broken MP3's were generated by all files in the batch except for the first.
• Some files couldn't be encoded due to BitHolder overflow.
• Feeding BladeEnc with 8-bit samples caused crashes.
• All AIFF-files were rejected.

Also, a big THANK YOU to all who downloaded the source code and found those bugs for me, you really prove that the open source model works!

As a result of this I've already released BladeEnc 0.81, which has fixed all these bugs and few other minor issues (see the changelog at my download page). I seriously recommend everyone to immediately upgrade to 0.81!

I haven't got a Windows version of 0.81 yet since I don't have Windows installed at home, so I've temporarily changed back the Windows binary to 0.76. I'll compile a new Windows version at work tomorrow and put it online.

Once again, sorry for all those bugs and any trouble it might have caused you...
 

This is the first release under the LGPL license and I hope it will signal the start of a new era for BladeEnc. The source code is available for download for anyone who wants to take a look at it and people who wants to contribute to BladeEnc's development are welcome to make improvements and send back the code. Take a look at the Source Code section of my homepage for more details.

The only versions currently available are the Windows version and various versions for Linux, but this will hopefully change quickly now when everyone can download the code... :)

Don't ask me when a new DLL will be available, it won't take too long...
 

I guess some of you wonders what happened with the patent issues. Well, after having looked through all the papers with my patent ombudsman and discussed the issue for a few hours we came to the conclusion that BladeEnc doesn't breach their patents. Not so much depending on the fact that it's software only, but due to some interesting details in their patent demands (which is the most central part of the patent papers). I don't want to go too much into detail yet since we still haven't resolved this issue once and for all (we still need to send some papers to the patent holders, specifying our conclusions and give them a chance to respond), but I have been recommended to just go on and release both the new version and the source code.

However, that doesn't necessarily mean that it's fully legal for you to download and use BladeEnc since the patents in your country might be different. Also, although BladeEnc in itself is a fully legal product here in Sweden, you might still make yourself guilty of patent infringements if you use it for some very specific tasks.

I'll explain everything more in detail when things have cleared up a bit and I'm not as tired as I am tonight.
 

I was made aware of this about half a year ago, but never fixed it because of five reasons:

1. Lack of time.
2. Lack of knowledge of how some parts of the encoding process worked meant that I didn't know how to fix it.
3. No player or tool seemed to bother, it all worked fine anyway, making me believe that it was commonly accepted that if the CRC-code was zero it should be ignored and not considered to be an error.
4. Nobody seemed to be using the CRC-switch anyway.
5. I recognized that this CRC-data could be filled in by a tool even after the MP3 file had been generated, so the problem could be solved later on if necessary without anyone having to throw away their MP3-files

I'm now working on a tool that will fix these MP3 files and I'll release it (with source code) as soon as it's finished, hopefully within a few days.
So don't delete or re-encode your files if they have been encoded with the CRC-switch enabled, this also goes for MP3 files encoded using all the other ISO based encoders, which also can be fixed with this tool.

BladeEnc 0.80 will of course have correct CRC calculation. In the meantime I recommend you to either not enable the -CRC switch on any new files you encode or run this tool afterwards.

First of all, I've been busy, terribly busy. For those who doesn't know it I'm working as programmer and project manager for a swedish game developer called UDS and for the last 20 months I've been responsible for what's going to be Codemaster's big release this Autumn, namely a game for PlayStation and PC called "No Fear Downhill Mountainbiking". This project currently takes an awful lot of my time and when I have some time left I'm often too exhausted to dig into the BladeEnc code.
 

Switched to Linux

I've been keeping an eye on Linux' progress during the last months and been really eager to switch over from Windows and a few months ago I finally took the big step. I've had Red Hat 5 installed on a second partition before, but since I also had Windows I never took the time to really dig into it, but now when I got myself a second computer I decided to only install Linux on it and force myself to learn the system.

It's been a rough ride with many hours of reading documentation, experimenting and reinstalling everything from scratch a few times, but now I got it all running the way I want and learnt everything I need to know, I find that I really like it! :)

For those of you who wants to try Linux I can really recommend a distribution called Mandrake. I currently use Mandrake 5.3, but 6.0 is in the works and contains a lot of improvements. I also recommend you to do some serious reading first if you (like me) don't have any earlier experience of UNIX and have a friend who knows Linux to help you install it since you can screw up your system royally if you don't know what you're doing.

What this means for BladeEnc is that I'll take care of the Linux i386 version myself and make it more of a standard Linux program with stdio-support, manpages etc. The Windows version won't suffer because of this. I know that most of my users are running Windows and I'll continue to compile the Windows version myself and keep it up-to-date.
 

Version 0.80 in the pipeline

Despite my lack of time I have done some improvements to BladeEnc and when I look at my changelog it's actually quite a few. The most noticeable features are that it now woks better under UNIX-systems (doesn't fuck up the terminal anymore and CTRL-Z, CTRL-C etc works as it should). supports streams (both stdin and stdout), some bugfixes and the fact that I've totally rewritten how the commandline is read, making it much more flexible. You can now specify individual output files and bitrates for each file in the batch and it remains backwards compatible with both the old BladeEnc and L3Enc commandlines. However, this version will not be released until my new patent problems have been resolved...
 

New patent problems. :-(

I'm having a new situation with the MP3 patent holders (mostly Thomson and Fraunhofer). 

They contacted me once again a few months ago, saying that hadn't answered their earlier letter (which I had) and repeated their demands. I answered this letter promptly, saying that I indeed had answered their previous letter and repeated my demand that they specify what patents I might be infringing so that I can investigate it further. 

This time I took the precaution to enable Netscapes "send receipt upon arrival" and "send receipt upon read" options. Just five minutes later I got a receipt from their web server saying that the letter had landed in the correct mailbox. I then waited two weeks for either a reply or the "has been read receipt" but never received any of them. I'm quite sure they read it, but didn't want to give me any indication of it. (I guess it could be seen as an indirect acceptance of me using the patents if it could be proven that they had read my letter but didn't answer, but I'm not a lawyer so I really don't know).

I then decided that I didn't want them to repeat the same procedure once again (and possibly threaten me with a trial), so I wrote them a new letter, telling them that I was still waiting for their answer and adding a line saying that I would consider all their demands to be void until they had specified what patents I might be infringing.

That got things moving. Just a few days later I got what seems to be a complete list of all their MP3 related patents. To my surprise I found that this list included a large number of european patents related to MP3 encoding that were registered as accepted by the Swedish patent authorities! :(

After some brief phone conversations with "Patent & Registreringsverket" (the Swedish patent and trademark authorities) I could just establish the fact that they had valid, registered patents on MP3 encoding in sweden, despite the fact that the law clearly says that software-only implementations and mathematical algorithms can't be covered by swedish patents! The people I spoke to were very helpful and knew the rules well, but could only agree that this whole situation seemed quite bizarre. I decided to order copies of all these patents so that I could investigate them myself at home and a few weeks later they arrived (something went wrong, so I had to order them again after about two weeks).

I had no problem to see that these patents indeed described the exact process needed to generate MP3 files, so I then decided to contact a "patentombud" (an expert on the legalities and issues surrounding patents). After having described the situation for him, he said that these patents most likely were fully valid, but that it was very doubtful if they could be applied on my situation (if some kind of hardware was involved the situation would have been different though). However, he needs to look into the details in order to help me, so I have now decided to hire him and will send him the papers on Monday. Hopefully he will come to the conclusion that I can happily continue developing BladeEnc, but if not, he will always be able to help me find the best possible solution. This initial investigation will cost me somewhere around $500-$1000 :(

Although this situation is very irritating for the moment, I'm glad to finally be able to shed some light on this situation so I finally get to know exactly what I can do and what I can't. I've found a lot of possible solutions which I'm going to discuss with my patentombud and although I don't want to give away any details yet, I must say that I'm confident we'll come out on top of this. 
 

Source release

The source for 0.80 will be released at the same time as the binary. Can't promise any date though since I want to solve this patent issue first and have so much to do at work.
 

Homepage updated

I've just updated some parts of my homepage, especially the frontend section where I've added a whole bunch of new programs and fixed a few broken links. I can really recommend the Windows users among you to try Exact Audio Copy since this ripper both is free to use and can guarantee 100% accurate ripping without glitches! Us Linux users can of course use CD-Paranoia instead... ;)

Linux i386
Linux MIPS
Linux Sparc
BeOS i386 (Release 3 only)
Solaris 2.6 i386
Solaris 2.6 SUN
NeXT m68k
HP-UX

They are all available from my download page...
 

First of all, I've moved to a new apartment. I did that just before Christmas so I just moved my things, cleaned the old one and left town for Christmas and New Year that I celebrated with my family and friends in Skåne (the south tip of Sweden, right next to Denmark). When I got back in town I didn't have a phone connection since there were some troubles and the phone company needed to get into my apartment to check a few things and since I had deadlines it remained unconnected for another 3 weeks. When I finally got my phone connected and could restore my internet connection I had 402 unanswered mails in my mailbox!!!

I still haven't read them all yet (I guess I'm a little bit more than halfway through) since I've both been busy at work and busy with getting everything set up in my new apartment. Btw, it's really a nice apartment for a bargain price. 86 square meters on the sixth floor with balcony and a splendid view over about a fourth of the town. It contains three rooms plus kitchen so now I have a nice bedroom, living room and computer room instead of everything squeezed into one small room. What a relief! I also took the opportunity to buy myself a new aquarium (375 litres, with a "back to nature" background) that I probably will keep some Guramis in and a new Computer (AMD K6-2 350MHz, 16 MB Riva TNT card, only 64 Mb RAM but with a 16 Gb Deskstar HD). I'm also going to get a direct internet connection (10 Mbit) in a few months and plan to run my old machine as a web server for these pages and some other stuff.

Anyway, back to BladeEnc. I have a few ports of 0.76 lying here (and probably more lurking in my unread mail) that I'll put online asap (including a Linux i386 one so you can stop asking me about it). I also took the opportunity to do some coding between Christmas and New Year and have now removed all copyrighted code from the source (at least as far as I know) and added a few switches. That means that the BladeEnc source now is almost ready for release, I'll just need another weekend on my own to fix the last things and then send it to the volunteers doing the ports so they can check that it can be compiled on as many systems as possible.

Unfortunately it seems like I can't use the LGPL license as I originally intended since it doesn't accept code containing patented algorithms (which in fact I consider to be a good precaution, although it suits me bad in this case). So I'll either have to write my own license agreement or use a somewhat less restrictive license like BSD. I'm fully capable of writing my own license agreement but it does take some effort and I don't like people having to learn a completely new license before having the right to play around with BladeEnc's source. Anyway, I hope we some day can work around all the MP3 related patents so we can have a totally free encoder.

Also, I don't think that my Newsletter has been working very well for the last months since I updated to Better Letter 3.0. I'll reinstall Better Letter 2.5 and see if I can get it working again...

And to all of you who sent me Christmas greetings: THANKS A LOT!!!  I really liked getting those, although I wasn't able to read them until mid-january...
 

Best Regards,

Tord Jansson
 

BladeEnc 0.76 has now been released. This is just a minor update from 0.75 where I have added one feature that has been asked for by many: Specify output directory on commandline. (use -OUTDIR=[path]) I have also tidied up the code a bit (can get quite messy with defines and ifdefs when you have the same source for all ports :), changed back the URL displayed at start-up to the old working one and removed one or two really minor bugs (can't even remember what they were, so they were really minor).

Speed and quality is the same as for 0.75 so there's no need to rush for it unless you're one of those who needs that switch.

The following versions are available:

Windows 9x/NT
FreeBSD & BSD/OS (i386)
UnixWare 7
OS/2
Irix 5.x
Irix 6.x

I had hoped to have the Linux i386 port available as well, but unfortunately I haven't received the binary yet so I'm afraid it will have to wait until early next year since
I'm moving to a new apartment tomorrow and then I'll leave town for my family's farm where I'll celebrate Christmas and New Year, without Internet access.
 
 

Sourcecode release delayed... :(

I know a lot of you have been waiting for the sourcecode to be released and I have earlier stated that it should be available before the end of the year. Unfortunately I haven't had
time to do some (more or less) necessary changes before I can release it under the LGPL. There are for example still fragments of code left that are copyrighted by others (but
free to use) that needs to be replaced before I can release the whole code under the LGPL license.

I want to make clear that this is just a delay due to me being totally buried in work (during November, which was my most busy month, I worked in average 80 hours a week!). I'm NOT having second thoughts about releasing the code or anything like that.
 
 

...and some info about a new great game in development ;)

Ok, I know that this newsletter only is intended for BladeEnc news, but I thought I'd take the opportunity to tell you about what I'm doing at work.

As some of you might know I work for a Game Developer called UDS, known for games such as Ignition and Absolute Pinball for the PC. I'm currently working as project manager and one of the main programmers for our next title which has been secret until quite recently.

The name of the game is "No Fear Downhill Mountainbiking" and it will be released by Codemasters for PC and PlayStation.

No information is available on either UDS' or Codemasters' sites yet, but some short previews have been published in gaming magazines and more are certainly to come in the near future, so keep your eyes open! ;)
 

Ok, that was all for me this time, so Merry Christmas and a Happy New Year to all of you!
 

Also, a lot of CD Rippers now support the BladeEnc DLL. These includes Easy CD-DA Extractor 3, Audiograbber, CD-Copy and CDex.
 

BeOS
NeXT
Solaris 2.6
Solaris x86 2.6
HP-UX
Linux Alpha
Linux MIPS
Linux Sparc
 

Also, Mikael Kjellström has provided us with a new port for OS/2 and even more ports are planned.
 

This new version is slightly faster ( 8% on my computer and testsample ) and has the same updated encoding engine as the DLL. The quality is identical to previous versions and a small bug that resulted in a totally harmless additional byte being appended to most MP3 files has been fixed.

The Digital UNIX port has been made by Stefan & Thomas Krantz and the Alpha FreeBSD port was delivered by Marca Registrada.

Since I haven't had so much time for BladeEnc lately Jukka Poikolainen (creator of a very good CD-Ripper called Easy CD-DA Extractor 3, available at http://www.saunalahti.fi/~poiko/cdda3/ ) volunteered to do the port. Other people have volunteered to do it before, but I have always wanted to do it myself since I had some very specific ideas on how I wanted it to work. But since I'm now more busy than ever with the game I'm working on I finally realized that I had to let somebody else write the DLL if I wanted it to be released before Christmas and Jukka seemed to be the right person for the job.

A nice side effect from this is that the latest beta of Easy CD-DA Extractor 3 already supports BladeEnc DLL! More supporting products should follow shortly, be so sure...

The DLL can be downloaded from my download page and comes with the necessary H-files and instructions on how to use it. Dead simple and straight forward if I may say so myself.

For people who don't know what a DLL is, don't bother to download it. It can't be run on its own, it needs to be integrated into another program that acts a bit like a frontend. The benefit of a DLL is that it can be more tightly integrated into another program than the normal BladeEnc.exe and can for example support things like on the fly encoding for CD-Ripping or radio recordings. It makes it also possible to use BladeEnc for internet radio broadcasting (if someone writes the necessary "frontend" for it).

I'm using monoliths free redirection service for that URL and it seems like they are having serious problems with their server. :(

Until they have solved their problems I recommend you to use my real URL instead:  http://home8.swipnet.se/~w-82625/
It's a lot harder to remember, but it works.

Also, Zac has updated his Linux i386 port of BladeEnc 0.72, this latest version is smaller and compiled with pgcc. It was done a while ago, but I haven't had time to put it online until now...

Solaris x86
SunOS
HP-UX

Also, the Solaris port has been recompiled and is about 12% faster than the previous 0.72 port. A special UltraSparc version has also been compiled which is 21% faster than the previous 0.72 port. 

All this has been made by Joel Fredrikson.
 

New Supporting Products

A lot of new frontends and rippers supporting BladeEnc have been released and nearly everyone of the old ones has been updated. I'm currently in the middle of updating my supporting products page and hopefully it should be online in a week or so. In the meantime it might be worth checking out if your favorite frontend has been released in a new version. 

Also, a frontend called DistributeEnc can be downloaded from http://consist.horizon.nl/klaasjan.te.voortwis/

That is a frontend for Windows 95/98 (sorry, not NT) THAT DISTRIBUTES THE ENCODING PROCESS OVER A NETWORK! I know this is something a lot of you have been looking forward to... :) 

There are also some Unix and OS/2 frontends available and I'll add them to my Supporting Products section when it's updated. 
 

The BladeEnc DLL

BladeEnc will very soon be released as a Windows DLL. This means that BladeEnc can be installed as a dynamically linkable system component on your computer, which very easily can be supported by the programs that has support for it. Since BladeEnc then only takes care of the encoding itself it will be possible to write programs that uses it in ways that it can't be used today. Software for Realtime Internet Broadcasting and CD Ripping with encoding in one go (i.e. without saving an intermediate WAV-file) are two things that will be possible to do with the BladeEnc DLL. 

The interest for this DLL has been huge, so you can expect a lot of Frontends and CD Rippers to be rewritten for it. 

BladeEnc will of course continue to be available as a stand-alone program. 
 

The Ports

Various volunteers will continue to port BladeEnc to more and more operating systems. However, since they all are volunteers doing the ports in their spare time I can't promise any release date for a particular port or if it will be available at all :( 

A few ports are definitely in the pipeline though and I'm still hoping for a total of 20+ ports before the end of the year. 
 

The Homepage

The homepage needs a major update and facelift, but that's quite low priority, so I guess you will have to do with the current one for another month or two. 
 

E-Mails

Nowadays I get between 10-30 e-mails a day (depending on if I recently made a new release), so I simply don't have time to answer them all. However, all e-mails are being read, so don't let this discourage you if you feel the urge to tell me something. If you want an answer, please write your question so that I easily can answer it with a short sentence or two. 

Also, some of you have written some interesting e-mails that really deserved an answer which I never gave. The reason is simply that I sometimes didn't have the time to answer any of the more comprehensive letters and therefore kept them lying in my inbox for a few days, where they simply got lost and forgotten when new mail was added daily. My apologies for that. If you have written an important letter that you still need an answer to, please re-send it to me. 
 

The Next Version

I'm also working on a new version of BladeEnc. It will hopefully be released quite soon and will included: 

+ Slightly faster encoding (8% faster for the moment, but I might do a few more optimizations, so we'll see when it's released). 

+ Support for a specified output directory. 

+ STDIN/STDOUT support. This will give the various UNIX versions of BladeEnc roughly the same abilities as the Windows DLL and has been one of the most   asked for features. 

+ Support for RAW input (i.e. sample data without any file header). 

And possibly some other features I might throw in if I find the time. 
 

The Sourcecode

Ok, all you Free Software advocates out there, hold your breath... 

THE SOURCECODE FOR BLADEENC WILL BE AVAILABLE UNDER A LGPL LICENSE! 

I've had a large number of letters from people urging me to release the source and some of them explained in great detail and enthusiasm about the benefits of free software. These letters made me see that there actually were great benefits to be had, both for me and the users, by releasing the sourcecode under a license like GPL. I eventually decided for the somewhat less restricted LGPL license, since I still want non-free products to be allowed to be linked with BladeEnc. 

People who are interested in how software can benefit from making its sourcecode available or is interested in the philosophy behind free software (which means more than 'software you don't have to pay for') or just want to take a look at the LGPL license should take a look at http://www.gnu.org . It's quite some heavy reading, but the ideas are very interesting... 

However, I have some things that needs to be fixed first, so don't expect to see it tomorrow. I think it would be safe to say that you can download it before the end of the year, but considering my current workload at UDS I don't dare to promise anything... :( 
 

I also want to take the opportunity to point out a misunderstanding: 

A lot of you have thought that I didn't want to release the sourcecode because I was afraid that someone would make a rip-off encoder or derivative product that didn't contain my name and/or generated MP3 files with a noticeably worse quality. 

That's not the case. The reason was that if someone downloaded the source and compiled it himself (without modifying it), he ran a great risk of ending up with a binary that produced worse result than my distributed binaries! 

Very few people know this, but a surprisingly big percentage of all compilers have bugs in their floating point handling procedures that give the result that they lose precision in floating point calculations if certain code optimizations are enabled. This includes well known and popular compilers such as Visual C++ 4.0 and quite a lot of versions of the GNU C compiler. 

It might be hard to believe that those bugs exist on compilers that successfully can compile sophisticated programs like the Linux kernel, but most 
programs will work properly even if the floating point calculations are slightly incorrect. Just remember how long it took before the Pentium 
floating point bug was discovered. 

BladeEnc (and the original ISO reference encoder) happens to be one of those rare programs that suffers a lot if the precision isn't kept in all the 
calculations. The result is that MP3 files encoded with a bad binary includes some kind of low frequency phasing sound that might be hard to spot, but is very disturbing in the long run (it actually gave me a headache after about an hour, although I hardly could hear it). 

Since I wanted to make sure that all versions of BladeEnc produced the same high quality MP3 files I decided to not release the source and instead encourage people to volunteer to port BladeEnc for their systems. That way I could keep a close eye on every binary that was generated. However, I didn't expect that people would ask for ports to more than 30 different systems... 
 

However, now when I've decided to release the source I've found another way to ensure quality. I'll include an internal checksum routine that can determine if BladeEnc is working properly when started. If not, the process will be halted and an explaining message will be displayed. 
 

Well, that was all the announcements I had. Hope you feel that there was something interesting for you. :) 
 

The Solaris port has been updated to 0.72. 
A bug in the OS/2 port that made it impossible for you to enter multiple files on the commandline has been removed. 
The Irix6 port now works better for people that are using bash instead of the standard shell and has been optimized. 
 

This port has been made by Giao Thanh Nguyen and has been tested on both Free BSD and BSD/OS.
 

The IRIX port comes in two versions, one for IRIX 5.2/5.3 and one for IRIX 6.x. 

This release features a speed increase of approximately 14% and some minor fixes (no new functions though) like the Linux i386 version no longer getting assertion failures on certain samples. Linux i386 users will also experience an additional speed boost since this version has been compiled with pgcc instead of gcc. 

The Linux i386 port has been made by Zac Livingston, Linux DEC Alpha port by Alexey Marinichev and OS/2 port by Mikael Kjellström. 

A total of 15 volunteers have received the sourcecode and if everything goes alright we will see BladeEnc 0.72 for the following additonal systems: 

AIX 
BeOS
Digital UNIX 
FreeBSD 
HP-UX 
IRIX 
Mac 
Mk Linux/Linux PPC 
MS-DOS 
NCR Unix SV5 (MP-RAS) 
Reliant UNIX 
SiemensNixdorf SINIX 
Solaris 
Solaris x86 
SunOs 4.1.4 
UnixWare 7 

So I guess we can say that we've covered the most of them. :) 

Just remember that all ports are done on voluntary basis, so there is no guarantee when a certain port will be available or if it will be released at all. 

This gave the result that most of my colleagues immediatelly removed their products from their homepages and ceased to develop them. A lot of MP3 enthusiasts got quite angry with Fraunhofer (to say the least) for marketing MP3 as an open standard and then when the userbase is big enough start to request $15 for every distributed encoder. 

However, this situation doesn't affect BladeEnc since swedish legislation specifically states that patents regarding mathematical algorithms or software implementations are NOT valid or enforceable in Sweden. I suspect that the same goes for quite a number of countries since these kinds of  patents are very questionable. 

I also sent a letter to Fraunhofer, politely asking them to specify what patents my product was infringing. I also ended the letter by saying that "If a license is needed I hope we can work something out". They haven't bothered to reply yet (I sent it one and a half week ago), so I suspect they recognised that they were a bit far out here and just tried to get rid of some unwanted competition. :-( 

My advice to all other freeware encoder authors is therefore to check what your countries legislation says about this. You can also check if your legislation specifically states that patents regulates the commercial use of a technique. If it does you might still be able to continue as long as you don't get any income from your product, but I'm not certain about that. 

I have also made some serious internal changes in order to make this version a doodle to port, so hopefully we will now see between 10-15 ports of this version within two weeks... 

UsE! has just been updated with tests of BladeEnc at 160 and 192 kBit, so now you can finally see the difference yourself. It also includes some animated graphical comparisons that are really usefull when you want to compare the encoders in detail. 

The URL for UsE! is: http://kempelen.inf.bme.hu/~dancsi/USE!/

Please remember that the quality tests at UsE! are technical tests and not perceptual tests. Technical tests never gives you a 100% true ranking system. The only true way to measure the quality is to listen carefully and judge for yourself. However, I find the quality tests at UsE! to be nearly identical with what I have experienced myself and what I've heard from others. 

The reasons for the long wait is that we have had quite some problems with this port. One compiler generated a mutated executable that produced MP3 files of a lower quality than the other ports and two others didn't generate working executables until some changes had been made to the source. 

But now we finally have a fully working OS/2 port, thanks to Mikael Kjellström, who has been struggling for quite a while with different compilers and had to debug the source in order to get it to run properly. 

A statically linked version of Zac's Linux i386 port is now available for all those of you who haven't got libc6.0 yet. 

For all those who wonder when we'll start porting version 0.63 I can only answer quite soon. I want to tidy up the code a bit and add defines for practically every system we're going to port for in order to make it a doodle to port in the future (that way I hope we can get all future ports out no later than a week after the windows version). And there will be a lot of ports in the future... 

A new version of FrontBlade has also been released. The new release features batch encoding and has a totally new interface. And of course, those small bugs in the interface have been fixed. Still the progress indicator doesn't work, but you can always open BladeEnc's minimized window to see the ETA. 

Both can be found on my supporting products page. 

They both support batch encoding, drag-n-drop, all bitrates and features of BladeEnc 0.63 and they don't need any runtime libraries. BladeEnc Helper has some nice features like automatically scan a directory (and optionally its subdirectories) for WAV-files and automatically add them to the batch. RazorBlade on the other hand has its own progress indicator and hides BladeEnc completely (not even a minimized window), making it look and feel like an encoder on its own. 

They are both highly recommended! 

BladeEnc Helper is written by Shawn Anderson and RazorBlade is written by Holger Dors. 

A lot of people wrote me and told me that the archive containing the Linux archive was corrupt, but I knew that I had extracted it to make sure that the manual was there before I uploaded it. So I thought it might have been corrupted when uploading it, so I downloaded it again (using FTP Explorer). The downloaded archive was identical with the uploaded one. Then I asked Zac (who made the port) to download it and see if he could find something that was wrong. He also said that the archive was corrupt and sent me a new one that was identical with the one I had gotten earlier. Then I finally  tried to download the archive with Netscape and  now I got a corrupted archive! 

After some testing I found that Netscape couldn't download any file with the extension ".tgz" from my homepage without destroying it! If I changed the extension to ".gz" it worked just fine. I don't know the mechanisms behind this, so I have no idea what went wrong or if it only happens with Netscape or with all browsers... 

Anyway, I have now changed the extension of both the i386 Linux and Solaris archive, so now can everyone download them correctly... 
 

Instead I have added a new site called UsE! that describes itself as  "The user oriented MP3 encoding guide". It includes detailed quality tests of nearly all encoders on the market and I find its ranking system to be very truthfull to reality. Highly recommended for anyone that wants to see some detailed testresults! 

However, keep in mind that the tests are performed at 128 kBits and that BladeEnc is geared for higher bitrates... 

Take a look at my Supporting Products page for some more info and a link to SoftechSoftware. 

The Digital Alpha Linux port has been made by Alexey Marinichev and the UnixWare 7 version has been made by Jon Coyle. 

BEShell is not just another L3Enc frontend, it's a frontend written specifically for BladeEnc and therefore supports all currently available features of BladeEnc, including all the bitrates and priority settings! 

Batch encoding is supported and since all files are sent to BladeEnc at the same time you get an estimated time for encoding the whole batch. 

BEShell is written by XeT Software and can be found on my supporting products page. 

This port will be used as the base for other Linux/Unix ports that therefore should follow shortly. 

Since I don't use Linux myself (although I'm planning to install it soon), I can't test this port myself. So I would appreciate to get some comments and suggestions from all you Linux users out there. 

I have also made some minor changes to the homepage, including some new speed comparisons. 

Since I'm working as a professional game developer I usually work in projects with a development cycle of 12-18 months. Along the project you have a number of dates when you need to reach certain milestones. When you get close to a milestone you drop everything else and concentrates 100% on the work you have to do, because if you have fallen behind schedule you really need to give 100% to not get yourself or the people depending on you into trouble and even if you're ahead of schedule you still give 100% since this is your last chance to get things as good as possible for the deadline. 

Currently we are in the final stages of our project which means that we have a deadline at least every second week (beta-versions for gametesters, advertising materials for the publisher, special demo versions for shows etc). I have just completed a very crucial deadline which included 30 hours of non-stop working between Saturday and Sunday. 

What these deadlines mean to BladeEnc is that I basically have no time at all left to work on the next version, update the homepage or answer e-mails. Sorry for those who haven't received any answers yet, I'll try to answer them as soon as possible. I won't completely stop working on BladeEnc though, but the development will slow down to a crawl until I have passed all the deadlines. 

I have tomorrow off and will then upload the long awaited PC Linux port and do some modifications to the homepage. 

I have a lot of ideas for the next version which mainly will include some quite useful features (probably no speedincrease this time), but please have patience, I can only spare a few hours a week for now. 

Now I need to get some sleep... 

But the most significant feature of BladeEnc 0.60 is a speed increase of more than 60 %. This makes BladeEnc the fastest freeware encoder currently available!

For those of you who are interested in details: I found a way to replace some very complex and slow floating-point operations with some large lookup-tables and iteration-loops. This gave me a speedincrease of 63% on my P133. It should give even a bit more on a Pentium II since my large lookup-tables should fit nicely in the P2's big cache, but I haven't been able to test that yet. Nedless to say, the quality of the output is identical to that of earlier versions. 

Also, take a look at my new "Supporting Products" page where I present different Frontends and CD Rippers that supports BladeEnc! 

I have also made some minor changes to the homepage, and updated the FAQ with some new questions that I have been receiving lately. 

Some people have voluntered to convert BladeEnc to other systems and I have therefore also made some internal changes in BladeEnc 0.54 to allow for easier conversions. Hopefully you will see a PC Linux, SUN Linux and Solaris version of BladeEnc in the near future. 

Quality and encoding speed is the same as in previous versions. 

This is the result of the main new feature of BladeEnc 0.50: an L3Enc emulation mode! 
The emulation mode is automatically entered when BladeEnc is called with a commandline that can be recognized as an L3Enc commandline. BladeEnc then mimics the behaviour of L3Enc by changing the default settings, reading the commandline and processing the specified WAV-file. 

The difference is that you are doing the compression with a freeware encoder that you don't need to register in order to use the higher bitrates! Also, the output is better when using bitrates above 128 kBit/s (see my quality section for more details) and the process is slightly faster! 

As soon as time permits I will add a Frontend section to my homepage in order to give you a detailed overview of the best frontends with direct download capabilities. 

For those who don't know :   A frontend is a graphical userinterface that lets you select what files to compress and set various options by just pointing and clicking. The frontend then calls the underlying program (in this case BladeEnc) with the correct parameters for performing the specified task.

This is now the best available combination if you want to make high-quality, high-bitrate MP3s of your CDs with least possible hassle. 

CD Copy can be found at http://www.cdcopy.sk/

I'm really sorry for that bug and promise it won't happen again. From now on I will compare the output from every new version of BladeEnc with a Reference MP3 that I have made. As long as the two files are identical no quality degradation can possibly have occured.