LameXP v4.21 Final · Build #2382 (2023-12-29)

[LoRd_MuldeR]

LameXP v4.14 Alpha 6

Changes between v4.13 and v4.14 [unreleased]:
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also effects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated Opus encoder/decoder libraries to v1.1.2 and Opus-Tools to v0.1.9 (2016-01-12)
* Updated MediaInfo to v0.7.82 (2016-01-27), compiled with ICL 15.0 and MSVC 12.0
* Updated QAAC add-in to the to QAAC v2.58 (2016-01-05)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Updated ALAC decoder to refalac v1.58 (2016-01-05)

[real.finder]

hi LoRd_MuldeR

I would suggest adding an option to choose between AAC encoders, and the auto choice act like the how lamexp do now except in low bit-rate/Quality or if the HE-AAC v2 is selected lamexp will prefer nero because it better than others in that

QAAC never encode in HE-AAC automatically, and support HE-AAC v1 only (did't support HE-AAC v2)

edit1: in tvbr mode in QAAC I can't encode in HE-AAC, and after I test more I see that QAAC in very low bit-rate (8 kbps) will encode HE-AAC v2 if I choose HE-AAC v1 or HE-AAC v2 ! and in the low (64 kbps) will encode in HE-AAC v1 if I choose HE-AAC v1 or HE-AAC v2 !

edit2: with 48000 Hz even in 8 kbps QAAC will encode HE-AAC v1 if either HE-AAC v1 or HE-AAC v2 chosen

edit3: 8 kbps in QAAC made the output in 32 kbps in 44100 Hz stereo, and in another test I see that in 40 kbps QAAC will encode HE-AAC v1 and in 32 kbps will encode HE-AAC v2 if either HE-AAC v1 or HE-AAC v2 chosen, if I didn't choose HE-AAC v1 nor HE-AAC v2 QAAC will encode in LC

[von Suppé]

Hi LoRd_MuldeR,

First, awesome program, thank you.
I found out that when I convert flac files into mp3, the <track> and <Album Artist> tag info from the flac files are not properly copied to the mp3 tags.
The <track> tag automatically gets 1, 2 3, etc. The <Album Artist> tag stays empty.
I checked the Meta Data and Advanced Options tabs but can't find anything about those.

Cheers
von Suppé

[LoRd_MuldeR]

I found out that when I convert flac files into mp3, the <track> and <Album Artist> tag info from the flac files are not properly copied to the mp3 tags.
The <track> tag automatically gets 1, 2 3, etc. The <Album Artist> tag stays empty.
I checked the Meta Data and Advanced Options tabs but can't find anything about those.

As explained various times before, this is not a simple matter of "copying" the meta tags ;)

It requires to detect the meta tags from the original input file, store them in an internal data structure, and finally re-embed them into the re-encoded file.

And all this needs to work with 17 different decoders and 11 different encoders (187 combinations), which all store different tags in different ways.

Therefore, the internal "meta data" model was designed as the "the lowest common denominator" of what the various different audio formats (encoder/decoders) support.

At this time, the "Album Artist" is not maintained as a separate field in the internal "meta data" model, sorry!

About the "track number" issue: I think you probably have "Position" set to "Generate from list position" on the "Meta Data" tab. If so, then the result is as expected.

[von Suppé]

At this time, the "Album Artist" is not maintained as a separate field in the internal "meta data" model, sorry!Thanks for the explanation, it be so.
About the "track number" issue: I think you probably have "Position" set to "Generate from list position" on the "Meta Data" tab. If so, then the result is as expected.Indeed it was. Doh, I feel stupid. Simply didn't know that "Position" meant track #. Changed it to "Unspecified (copy from source file)" and it works like a charm now :-)

Thank you
von Suppé

[LoRd_MuldeR]

LameXP v4.14 Alpha 8
https://sourceforge.net/projects/lamexp/files/Snapshots%20%28BETA%29/2016-04-07/LameXP-ALPHA.2016-04-07.Release-Static.Build-1878.exe/download

Changes between v4.13 and v4.14 [unreleased]:
* Upgraded build environment to Microsoft Visual Studio 2015 with Update-2
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also effects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated Opus encoder/decoder libraries to v1.1.2 and Opus-Tools to v0.1.9 (2016-01-12)
* Updated MediaInfo to v0.7.82 (2016-01-27), compiled with ICL 15.0 and MSVC 12.0
* Updated QAAC add-in to the to QAAC v2.58 (2016-01-05)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Updated ALAC decoder to refalac v1.58 (2016-01-05)

[Yonah]

I'm having an issue with Vorbis Comment tags on Ogg files encoded with LameXP. My source files are FLAC files and have all been properly tagged. However, after encoding with LameXP, all the tags are duplicated with 2 backslashes (\\) except for the year.

For example, if the artist tag in the FLAC file is "Disturbed", the encoded Ogg file is tagged as "Disturbed\\Disturbed." The comment contains "Encoded with LameXP\\", followed by the original comment from the source FLAC file.

This bug does not occur if the Normalization Filter is enabled.

[LoRd_MuldeR]

I'm having an issue with Vorbis Comment tags on Ogg files encoded with LameXP. My source files are FLAC files and have all been properly tagged. However, after encoding with LameXP, all the tags are duplicated with 2 backslashes (\\) except for the year.

For example, if the artist tag in the FLAC file is "Disturbed", the encoded Ogg file is tagged as "Disturbed\\Disturbed." The comment contains "Encoded with LameXP\\", followed by the original comment from the source FLAC file.

This bug does not occur if the Normalization Filter is enabled.

Hello, Yonah.

If the selected encoder is capable of reading the input file format directly (and if no filters are enabled!), then LameXP takes a "shortcut" and just invokes the encoder directly on the input file – instead of going the usual "decoder → filters → encoder" route with intermediate WAV files. Now my guess is that, when OggEnc2 (i.e. the Ogg/Vorbis encoder) is reading your input FLAC file directly, then it keeps the meta tags found in the original FLAC and also appends the tags specified explicitly by LameXP – rather than "replacing" redundant tags. Not sure whether this is a bug or a feature. Enabling the "normalization" filter probably fixes this, because the "shortcut" can no longer be used when filters are enabled ;)

[EDIT]

Okay, I think it's a feature and we need to use the --discard-comments command-line switch of OggEnc2 to get the desired result. A new version is underway...

[Yonah]

Okay, I think it's a feature and we need to use the --discard-comments command-line switch of OggEnc2 to get the desired result. A new version is underway...

Thanks, Lord_Mulder! The command-line switch is just what I needed. Before I moved overseas I ripped my entire CD collection to FLAC. Since I use Vorbis exclusively on my phone this tag issue was problematic for me. No more manual tag editing after I encode an album. :)

[LoRd_MuldeR]

LameXP v4.14 Alpha 9
https://sourceforge.net/projects/lamexp/files/Snapshots%20%28BETA%29/2016-04-16/LameXP-ALPHA.2016-04-16.Release-Static.Build-1880.exe/download

Changes between v4.13 and v4.14 [unreleased]:
* Upgraded build environment to Microsoft Visual Studio 2015 with Update-2
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Fixed embedding of meta tags with OggEnc2 when reading directly from OGG/FLAC input file
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also affects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated Opus encoder/decoder libraries to v1.1.2 and Opus-Tools to v0.1.9 (2016-01-12)
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated MediaInfo to v0.7.82 (2016-01-27), compiled with ICL 15.0 and MSVC 12.0
* Updated QAAC add-in to the to QAAC v2.58 (2016-01-05)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Updated ALAC decoder to refalac v1.58 (2016-01-05)

[MCFish]

You already sent me this exact same screenshot via PM. See my reply. Please don't double-post. Thanks!

Hi i need that answer too. Ive got this error on any multichannel file

[LoRd_MuldeR]

Hi i need that answer too. Ive got this error on any multichannel file

Here we go:
I got this error when i tried to convert a. dts audio file to aac. What should i do in order to fix this?

http://i.cubeupload.com/JmcNnL.jpg

Hello.

The problem is that the WAV file (decoded from the input DTS) exceeds a size of 4 GB, which is above the WAV format file size limits. There is nothing that can be done about it :(

Using Wav64 (http://www.ambisonia.com/Members/mleese/sony_wave64.pdf/sony_wave64.pdf) would be an option, if more tools did support it. But they don't. So we are stuck with plain old WAV for now...

BTW:
* http://windows.microsoft.com/en-us/windows/use-snipping-tool-capture-screen-shots#1TC=windows-7
* https://en.wikipedia.org/wiki/Cut,_copy,_and_paste

Regards,
MuldeR

[MCFish]

Ok thanks for quick answer.
But, after i enable qaac in lamexp, qaac encodes the same file with success. Opus encoder still fails. How is that possible?
Tried it twice now, and both success with qaac...
Wav file should still be same size, right?

edit, i saw the 4gb warning in the log....yuck encoded file had a longer playtime...

[LoRd_MuldeR]

The original WAV format, the one that actually has wide support, is limited to 4 GB. Some implementations are even limited to 2 GB, because they interpret the "size" field as signed (the specs don't say it, but only unsigned makes sense).

Then there are the WAVE64 and RF64 formats: While WAVE64 is a proper extension of the WAV standard to 64-Bit and thus what everybody should be using in an ideal world (but not backwards compatible to original WAV), the RF64 format is pretty much a "dirty" hack to allow more than 4 GB in a single WAV file (note: RF64 is nothing but several standard WAV file structures dumped into one contiguous file) and provides at least some backwards compatibly to original WAV.

Now, what happens when the writing application writes a normal WAV files but the size exceeds 4 GB? Some application just fail. Some create a normal WAV file with more than 4 GB of content, which means you get a non-standard (broken) WAV file. And some switch over from WAV to RF64 "on the fly" when the 4 GB limit is exceeded. At least in the latter case there is a certain chance that the reading application will be able to read the file. But there is no guarantee it will work!

(In the end, WAV files larger than 4 GB are always problematic and things may fail on you randomly)

[Sparktank]

human wiki

Thanks for the explaination!
It makes so much more sense now.

It's quite hard to keep track of what actually supports w64 formats these days.
Thankfully, eac3to does. Custom builds of SoX. My eyes.

[mariush]

As a workaround, you may want to try saving to FLAC (lossless, fast compression) to get around the 4 GB limitations of WAV files. Then you can pipe the decoded output from FLAC straight into neroaac or some other audio encoder that supports raw / piped input

[MCFish]

Thanks you all for explaining and posting suggestions to the wav limitation. Guess i just continue to do doublework in Atak's Ripbot

[manolito]

We had this discussion about WAV files > 4GB before in this thread. LoRd_MuldeR made it very clear that he wanted LameXP to be a universal solution, and since other tools used by LXP do not (or only partly) support WAV64 of RF64 there is no way to support the conversion of large multichannel sources.

To me the only workaround was to give up on universality, so I definded some special cases and the corresponding work flows.

Generally the problem occurs for 6-ch (or more than 6) sources with a length of a little over 2 hours. For these cases I needed solutions which did not use intermediate WAV files.

AC3 to AAC:
BeSweet or EAC3To

DTS to AC3 or AAC:
Eac3To

AAC to AC3:
For this I made my own converter which uses FAAD, SoX and Aften. (Neither BeSweet nor Eac3To support AAC input.)
http://forum.doom9.org/showthread.php?p=1616904#post1616904


Of course it can be done differently using FFmpeg (and QAAC with piping).


Cheers
manolito

[Motenai Yoda]

well most tools supports pipe in/out, so maybe using ie sox to do it can be a workaround to feed those tools.

also ffmpeg has its own aac codec, and can be build with libfdk_aac too

[LoRd_MuldeR]

LameXP v4.14 Beta-2

Changes between v4.13 and v4.14 [unreleased]:
* Upgraded build environment to Microsoft Visual Studio 2015 with Update-2
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Fixed embedding of meta tags with OggEnc2 when reading directly from OGG/FLAC input file
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also affects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated Opus encoder/decoder libraries to v1.1.2 and Opus-Tools to v0.1.9 (2016-05-22)
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated MediaInfo to v0.7.85 (2016-04-29), compiled with ICL 15.0 and MSVC 12.0
* Updated QAAC add-in to the to QAAC v2.58 (2016-01-05)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Updated ALAC decoder to refalac v1.58 (2016-01-05)
* Improved auto-update function (faster Internet connectivity check)

[GMJCZP]

Hi LordMulder:

In February formatted my pc and 4.13 LameXP had not installed. I switched from avast to avg and now the program only runs disabling the antivirus. If active the antivirus the pc hangs me, not even placing .exe exclusions works.
I like very much this program but clarify that I have no intention of changing AV, sorry. I have WXP SP3.

[mike20021969]

...avg and now the program only runs disabling the antivirus.,.. I have WXP SP3.I'm running Windows XP SP3 with AVG Free (I'm using the 2012 version as 2013 onwards gave a BSOD EVERY time during install). LameXP even with AVG active works fine. Launches in <5 seconds.

[GMJCZP]

I'm running Windows XP SP3 with AVG Free (I'm using the 2012 version as 2013 onwards gave a BSOD EVERY time during install). LameXP even with AVG active works fine. Launches in <5 seconds.

2012? Why? :scared:

I have last version offline updated.

[mike20021969]

2012? Why? :scared:As I said "2013 onwards gave a BSOD EVERY time during install".

[LoRd_MuldeR]

I like very much this program but clarify that I have no intention of changing AV, sorry. I have WXP SP3.

Well, it's your decision to stick with a buggy anti-virus software, even if it means that you cannot use legitimate applications.

Personally, I can not recommend to accept this new form of censorship though.

(Technically there nothing I can do on my side - unless they showed me the exact line of code in my program that justifiably classifies as "malware" behavior - which of course doesn't exist)

[GMJCZP]

IMHO, change a AV to run a program would put the garbage under the carpet.
If you are interested, Simplex264 serves me right, but by LameXP, spend several times in the Recovery Console with chkdsk /p yesterday was unpleasant.
If the manufacturer does not listen to a client nothing can be done.

It's a shame, because your programs are good. Maybe that's the problem.

[LoRd_MuldeR]

IMHO, change a AV to run a program would put the garbage under the carpet.

In what way ???

The so-called "anti-virus" software blocks a program that, without the slightest doubt, is 100% legitimate (if you want say the opposite, then show me the exact lines of code in my program that unambiguously classify "malware" behavior"), so the bug/problem is in the anti-virus software! Consequently the one and only place where the bug could be fixed is the anti-virus software. Because anti-virus software is a "blackbox" and they keep their algorithms secret, any kind of speculation why anti-virus software XYZ is failing horribly this time is totally pointless. Especially because there is a zillion of anti-virus softwars out there and because their behvious can (and most often will) change with the next update.

NOTE: You have to understand that "FALSE POSITIVES" are an important part of the business model of anti-virus companies. They sell "premium" contracts to the big software companies, so the programs of those companies get "whitelisted" immediately. If there wasn't a sufficiently high probability that 100% legitimate programs get blocked due to "FALSE POSITIVES", then this business model wouldn't work. Of course it means that if you are not one of those big companies, you stuff just gets censored! That's how blackmail works in the 21st century. For independent software developers and smaller companies this has become a HUGE problem. I also work on a couple of commercial projects, so I know what I'm talking about...

(Buying an EV software-signing certificates helps, to some degree, but the price is WAY beyond what I can afford for my hobby projects. Let alone that you need to be a registered business in order to get this kind of certificate)

[GMJCZP]

In this moment I use AVG free. I don't have internet in my pc, so it's enough for me.

[lethedoom]

In this moment I use AVG free. I don't have internet in my pc, so it's enough for me.
also,"Hi LordMulder:

In February formatted my pc and 4.13 LameXP had not installed. I switched from avast to avg and now the program only runs disabling the antivirus. If active the antivirus the pc hangs me, not even placing .exe exclusions works.
I like very much this program but clarify that I have no intention of changing AV, sorry. I have WXP SP3.
__________________
By law and justice!

RescueFrame"

http://finance.yahoo.com/news/avast-announces-agreement-acquire-avg-063000656.html

What will you do ?

[Randy31416]

4.13 introduced a buglet in file naming that I believe 4.12 did not have. I have the advanced option "Rename Output Files" unchecked (no checkmark) so file names should not be changed. Yet a buch of filenames are changed. All of them end in a period -- file names such as "Maestro E.K.E." or "Wach uff myn Hört etc." that are created by another program reading external sources. The filetype I am using is ".wav", and the files are being converted to ".mp3". When a filename ends in a period, the period is stripped during conversion. So, for example, "M:\Music\For Ellington\Maestro E.K.E..wav" does not become "M:\Music\For Ellington\Maestro E.K.E..mp3" but becomes "M:\Music\For Ellington\Maestro E.K.E.mp3" instead. I do note that double internal spaces are now preserved, but I persist in thinking that not checking "rename output files" should mean that I get out the same name I put in.

[LoRd_MuldeR]

Randy31416, this is because the clean_file_name() function is applied to clean up the file name, before the final file extension is appended. And dots (.) at the end of the file name are forbidden, so they are removed.

I have changed the logic now, so that the clean up function gets applied at a slightly later point:
https://sourceforge.net/projects/muldersoft/files/LameXP/Testing/LameXP-BETA.2016-07-17.Release-Static.Build-1896.exe/download

[Randy31416]

I have changed the logic now, so that the clean up function gets applied at a slightly later pointYou do support this program exceptionally well. Thanks. Randy

[LoRd_MuldeR]

LameXP v4.14 Beta-4

Changes between v4.13 and v4.14 [unreleased]:
* Upgraded build environment to Microsoft Visual Studio 2015 with Update-2
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Fixed embedding of meta tags with OggEnc2 when reading directly from OGG/FLAC input file
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also affects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated Opus encoder/decoder libraries to v1.1.2 and Opus-Tools to v0.1.9 (2016-05-22)
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated MediaInfo to v0.7.85 (2016-04-29), compiled with ICL 15.0 and MSVC 12.0
* Updated QAAC add-in to the to QAAC v2.58 (2016-01-05)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Updated mpg123 decoder to v1.23.4 (2016-05-11), compiled with GCC 5.3.0
* Updated ALAC decoder to refalac v1.58 (2016-01-05)
* Updated GnuPG to v1.4.21 (2016-08-17), compiled with GCC 6.1.0
* Improved auto-update function (faster Internet connectivity check)

[LoRd_MuldeR]

LameXP v4.14 Beta-5

Changes between v4.13 and v4.14 [unreleased]:
* Upgraded build environment to Microsoft Visual Studio 2015 with Update-2
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Fixed embedding of meta tags with OggEnc2 when reading directly from OGG/FLAC input file
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also affects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated Opus encoder/decoder libraries to v1.1.3 and Opus-Tools to v0.1.9 (2016-09-11)
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated MediaInfo to v0.7.85 (2016-04-29), compiled with ICL 15.0 and MSVC 12.0
* Updated QAAC add-in to the to QAAC v2.58 (2016-01-05)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Updated mpg123 decoder to v1.23.4 (2016-05-11), compiled with GCC 5.3.0
* Updated ALAC decoder to refalac v1.58 (2016-01-05)
* Updated GnuPG to v1.4.21 (2016-08-17), compiled with GCC 6.1.0
* Improved auto-update function (faster Internet connectivity check)

[LoRd_MuldeR]

The QAAC Add-in has been updated to QAAC v2.61:
LameXP.qaac-addin.2016-10-03.zip (https://sourceforge.net/projects/lamexp/files/Miscellaneous/Add-ins/qaac/LameXP.qaac-addin.2016-10-03.zip/download)

[LoRd_MuldeR]

LameXP v4.14 Beta-7

Changes between v4.13 and v4.14 [unreleased]:
* Upgraded build environment to Microsoft Visual Studio 2015 with Update-2
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Fixed embedding of meta tags with OggEnc2 when reading directly from OGG/FLAC input file
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also affects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated Opus encoder/decoder libraries to v1.1.3 and Opus-Tools to v0.1.9 (2016-09-11)
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated MediaInfo to v0.7.88 (2016-08-31), compiled with ICL 15.0 and MSVC 12.0
* Updated QAAC add-in to the to QAAC v2.61 (2016-10-02)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Updated mpg123 decoder to v1.23.4 (2016-05-11), compiled with GCC 5.3.0
* Updated ALAC decoder to refalac v1.61 (2016-10-02)
* Updated WavPack decoder to v4.80.0 (2016-03-28), compiled with ICL 15.0 and MSVC 12.0
* Updated GnuPG to v1.4.21 (2016-08-17), compiled with GCC 6.1.0
* Improved auto-update function (faster Internet connectivity check)

[LoRd_MuldeR]

LameXP v4.14 RC-1

Changes between v4.13 and v4.14 [unreleased]:
* Upgraded build environment to Microsoft Visual Studio 2015 with Update-2
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Fixed embedding of meta tags with OggEnc2 when reading directly from OGG/FLAC input file
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also affects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated Opus encoder/decoder libraries to v1.1.3 and Opus-Tools to v0.1.9 (2016-10-16)
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated FLAC encoder/decoder to v1.3.1 (2016-10-04), compiled with ICL 17.0 and MSVC 12.0
* Updated MediaInfo to v0.7.88 (2016-08-31), compiled with ICL 15.0 and MSVC 12.0
* Updated mpg123 decoder to v1.23.4 (2016-05-11), compiled with GCC 5.3.0
* Updated ALAC decoder to refalac v1.61 (2016-10-02)
* Updated WavPack decoder to v4.80.0 (2016-03-28), compiled with ICL 15.0 and MSVC 12.0
* Updated GnuPG to v1.4.21 (2016-08-17), compiled with GCC 6.1.0
* Updated QAAC add-in to the to QAAC v2.61 (2016-10-02)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Improved auto-update function (faster Internet connectivity check)

[LoRd_MuldeR]

LameXP v4.14 RC-2

Changes between v4.13 and v4.14 [unreleased]:
* Upgraded build environment to Microsoft Visual Studio 2015 with Update-2
* Fixed the location of temporary intermediate files for SoX-based audio effects
* Fixed embedding of meta tags with OggEnc2 when reading directly from OGG/FLAC input file
* Enabled the "built-in" resampler for QAAC encoder
* The "Algorithm Quality" slider now also affects the QAAC encoder
* Added "AVX" (Advanced Vector Extensions) to CPU feature detection code
* Updated Opus encoder/decoder libraries to v1.1.3 and Opus-Tools to v0.1.9 (2016-10-16)
* Updated LAME encoder to v3.100 Alpha-2 (2016-01-29), compiled with ICL 15.0 and MSVC 12.0
* Updated FLAC encoder/decoder to v1.3.1 (2016-10-04), compiled with ICL 17.0 and MSVC 12.0
* Updated MediaInfo to v0.7.88 (2016-08-31), compiled with ICL 15.0 and MSVC 12.0
* Updated mpg123 decoder to v1.23.8 (2016-09-27), compiled with GCC 6.2.0
* Updated ALAC decoder to refalac v1.61 (2016-10-02)
* Updated WavPack decoder to v4.80.0 (2016-03-28), compiled with ICL 15.0 and MSVC 12.0
* Updated GnuPG to v1.4.21 (2016-08-17), compiled with GCC 6.1.0
* Updated QAAC add-in to the to QAAC v2.61 (2016-10-02)
* Updated FhgAacEnc add-in to "Case" edition (2015-10-24)
* Improved auto-update function (faster Internet connectivity check)
* Updated language files (big thank-you to all contributors !!!)

[digitaltoast]

Looking good, but if you're after feedback on 4.14RC2, I just have a couple:

During install, it wants to install to folder 4.11 and won't let me change path.
When encoding mp3, even when I override the samplerate and enforce 44.1KHz, it downgrades to 32KHz etc, at anything above VBR compression level 6.

That's it, so far! Do you need screenshots or is that OK? Thanks for a great app!

[LoRd_MuldeR]

During install, it wants to install to folder 4.11 and won't let me change path.

That's because the installer will "upgrade" the existing installation - which is located wherever you originally installed LameXP.

When encoding mp3, even when I override the samplerate and enforce 44.1KHz, it downgrades to 32KHz etc, at anything above VBR compression level 6.

And that's because of LAME's psycho-acoustic model.

The sampling rate determines the highest frequency that can be retained (cf. Nyquist–Shannon sampling theorem (https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem)). So, by reducing the sampling rate from 44.1 kHz to 32 kHz, you are effectively low-pass filtering the audio. In other words, the highest frequencies are discarded before the actual encoding. At a sampling frequency of 44.1 kHz the highest audio frequency that can be retained (without aliasing) is 22.05 kHz; at a sampling frequency of 32 kHz, the highest audio frequency that can be retained is 16 kHz. Low-pass filtering the audio makes a lot of sense at very low bitrates (or the lowest VBR quality levels), because loosing the highest frequencies is way less disturbing than the compression artifacts that arise otherwise.

Note: It's possible to "force" LAME to encode at the original sampling rate, by explicitly specifying the "--resample" option with the original sampling rate. But you almost always do not want this!

[digitaltoast]

Many thanks for the explanation:

Note: It's possible to "force" LAME to encode at the original sampling rate, by explicitly specifying the "--resample" option with the original sampling rate. But you almost always do not want this!

Except of course when it's a particularly archaic piece of hardware which can only cope with multiples of 11.020 :)

Perfect, thank you.

[LoRd_MuldeR]

Except of course when it's a particularly archaic piece of hardware which can only cope with multiples of 11.020 :)

Just for the notes: The original MP3 (actually "MPEG-1 Audio Layer III") specification from 1993 defines three possible sampling rates – 48 kHz, 44.1 kHz and 32 kHz – that any "compliant" decoder must support.

(More sampling rates – 16 kHz, 22.05 kHz and 24 kHz – have been added to MP3 with the MPEG-2 standard in 1995)

[RieGo]

in the latest version (RC2) it seems that opus isn't working properly. if i encode with the same setting as before (VBR 16kbps; complexity 10; framesize 60) i get valid opus files with correct filesize, but when played back it's just extremely noisy, which sounds like compressed at 1kbps :)
anyways, keep up the good work!

[LoRd_MuldeR]

in the latest version (RC2) it seems that opus isn't working properly. if i encode with the same setting as before (VBR 16kbps; complexity 10; framesize 60) i get valid opus files with correct filesize, but when played back it's just extremely noisy, which sounds like compressed at 1kbps :)
anyways, keep up the good work!

I see what you mean. I tried encoding a piece of music at ~16 kbps and it came out as total noise. Not that I would expect anything worthwhile at ~16 kbps.

Interestingly, at ~24 kbps it came out "okay". Of course, the ~24 kbps version didn't sound very good either. But considering that it's only ~24 kbps Opus still performed pretty well at that rate. Nothing like the ~16 kbps result.

I'm not sure whether the ~16 kbps result is a bug in latest Opus or just what you have to expect at that ultra-low rate. If you got better result before, with what Opus version was that? :confused:

[RieGo]

I see what you mean. I tried encoding a piece of music at ~16 kbps and it came out as total noise. Not that I would expect anything worthwhile at ~16 kbps.

Interestingly, at ~24 kbps it came out "okay". Of course, the ~24 kbps version didn't sound very good either. But considering that it's only ~24 kbps Opus still performed pretty well at that rate. Nothing like the ~16 kbps result.

I'm not sure whether the ~16 kbps result is a bug in latest Opus or just what you have to expect at that ultra-low rate. If you got better result before, with what Opus version was that? :confused:

yes! i use opus mainly to encode tons of audiobooks (mono), which results in "really good" sound quality starting at 16kbps. it worked just fine on releases before RC2 (or RC1?) and it also works with official opus encoder 1.1.3.
if you say 24kbps came out okay for you, it may be just a problem on my side. i'll try to reinstall LameXP and check.

update:
same problem after reinstalling and removing settings.
i uploaded 2 samples, so you know what i mean:
03_opusenc.opus (https://mega.nz/#!ssAX2BzS!N9SXRhmDMtiO7cfNqV0Ujb0oaYFi9e4M0B3UHAyPBAo)
03_RC2.opus (https://mega.nz/#!I5Bgya6Y!UaA4FNFht69HAMef0U5NdcYMFvfjUIgWYjtSdBUTAkE)

i also tried encoding at 24kbps and 128kbps:
24kbps has the same weird artifacts, but not as present as with lower bitrate
128kbps sounds just fine

maybe it's really just a bug in the latest Opus, which effects only lowest bitrates

btw: i have seen that opus files generated by LameXP are always 1kbps lower than set. so i encoded my sample at 15kbps with opusenc

[LoRd_MuldeR]

yes! i use opus mainly to encode tons of audiobooks (mono), which results in "really good" sound quality starting at 16kbps. it worked just fine on releases before RC2 (or RC1?) and it also works with official opus encoder 1.1.3.

Opus binaries have last been updated in RC-1, from Git "master" 2016-09-11 to 2016-10-16. There has not been any change to Opus between RC-1 and RC-2.


maybe it's really just a bug in the latest Opus, which effects only lowest bitrates

I noticed two recent commits/fixes that may be related to this problem: https://git.xiph.org/?p=opus.git;a=commit;h=6d0628493550f4ef0eaecf3058a502404bc6dca2
https://git.xiph.org/?p=opus.git;a=commit;h=2af92cd99f1f78f596a49dc951d17faeb9363dd3

Will try building latest version soon... (not this weekend though)


btw: i have seen that opus files generated by LameXP are always 1kbps lower than set. so i encoded my sample at 15kbps with opusenc

By default, Opus uses VBR mode. So, by default, the selected bitrate only "specifies the average rate for a large and diverse collection of audio". It's more a rough estimate.

Your particular audio file may, very well, come out at a somewhat higher or a somewhat lower average bitrate!

If you really whish to hit a specific average bitrate, then you have to add the "--cvbr" option, aka "constrained variable bitrate encoding". Or even use the "--hard-cbr" option.

[LoRd_MuldeR]

RieGo, can you please try again with the latest LameXP build?
https://sourceforge.net/projects/lamexp/files/Snapshots%20%28BETA%29/2016-10-30/LameXP-RC3.2016-10-30.Release-Static.Build-1924.exe/download

(After some testing, it appears that the "noisy" output occurs when SSE2 or AVX compiler-optimizations are enabled for the Opus library, so I keep them disabled for now. The recent fixes in Opus code seem unrelated)

[amayra]

i try to convert m4a " dash" but lamexp tell me this is not supported format ?
source form youtube

[LoRd_MuldeR]

i try to convert m4a " dash" but lamexp tell me this is not supported format ?
source form youtube

What does MediaInfo say about that file?
https://github.com/lordmulder/mediainfo-gui/releases/latest

[RieGo]

RieGo, can you please try again with the latest LameXP build?
https://sourceforge.net/projects/lamexp/files/Snapshots%20%28BETA%29/2016-10-30/LameXP-RC3.2016-10-30.Release-Static.Build-1924.exe/download

(After some testing, it appears that the "noisy" output occurs when SSE2 or AVX compiler-optimizations are enabled for the Opus library, so I keep them disabled for now. The recent fixes in Opus code seem unrelated)

your new release works just fine. thanks a lot!