Ardour and MP3

Before Ardour 6.0, importing MP3 format files was not possible. I have left in place the comments on an old article that explained the licensing issues that made that tricky, and then went on to explain why using MP3 (or other lossy compression format) files as source material in a project is a bad idea. I am still very strongly convinced that it continues to be a bad idea to work with MP3 data in a DAW, but it seems counter-productive to have that debate in this context. I have left everyone's comments in place for those who want to read them.

@timrob: if you start with

@timrob: if you start with lossily-compressed material, nothing you can to the audio will ever put back the data that is missing from the beginning. It doesn't matter whether the "final" output is wav, flac or some other format - if that file is ever reconverted to some other lossy format, the result is a total mess. You're not getting the benefit of 24 bit audio - the process you describe is totally independent of the bit depth you use.

It is true that if libsndfile simply supported MP3's then we wouldn't be having most of this conversation. Instead, I get asked "why can't I import MP3's into Ardour?" by Linux users, and I proceed to explain why they should not do so even if they could. So, although licensing is the proximal cause for the situation on Linux, the core, key issue is that people should not be mixing in this way. Not with Ardour, not with Logic or any other DAW.

@timrob: if you start with

@timrob: if you start with lossily-compressed material, nothing you can to the audio will ever put back the data that is missing from the beginning. It doesn't matter whether the "final" output is wav, flac or some other format - if that file is ever reconverted to some other lossy format, the result is a total mess. You're not getting the benefit of 24 bit audio - the process you describe is totally independent of the bit depth you use.

I suspect he was referencing using the MP3 for timing purposes, and resubmitting an overdub back in full quality, in which case the quality of the MP3 is inconsequential as it isn't mixed into the final output at all. I have done similar on other projects, and while I have no issue just decompressing the MP3 to import it into the timeline, it is one of the few valid arguments I could see.


I think we are at the point

I think we are at the point of why ardor remains a niche project ...
the reason is simply that you do not add those functions that serve the people.
Take reaper .. because he had his success?
simply because it is a good cheap product and complete ..
ardor is a bomb ready to explode but fails to do so ... but for the simple fact of what it lacks.
I saw a post that a developer has proposed to add the function of markers "to cubase and ableton" and then coming to a daw full of features like the most famous ...
then why can not take off?? Only because a user is not at ease, this tread is the ultimate example. because I can not use mp3 files if I need it or want it? As we begin to actually insert a function of conversion when importing mp3 files?
Only because all the best sequencer give you a form to its sound fonts and a parametric eq on the channel and ardor not?
the goodness of a sw not only to assess the goodness of his code (although the lack of bugs is a very important thing) but also the chance to follow your workflow, if I want to use the mp3 file and mix in a project files or other wave or mp3 files sampled from the base to create music that appeals to me, to you the development team that creates discomfort there?
I do not want an open flame, or what, I want you to think only on certain points, the development of a sw, even ardor, is itself a commercial activity, and a service that you offer to people, if people can not aprezzare your efforts, and I know you've had a hard pull of ardor, perhaps there is some trick to be taken to change opinions on the trend, it is better to have an economic upside.
The result obtained is that people change daw or operating system that educates people and helps them to use, most people use Linux for daily tasks, and window or mac to produce music, this should give an alarm bell, as things that I said we reaper.
guardaqueste things for people to choose and if you have two choices choose one where there is better, even if it means that shall violate licensing or other ...
ardor is a great product in itself, but has many small changing to do to make it truly flourish, and some have mentioned here, like having your tools to work with sound, read the mp3 files, and tools to get full access to the potential of your hardware and your sw. no getting away from this, as you can not escape the fact that if there is a version of ardor that can attract customers is the version for osx and window in the future, and therefore get people to linux because if one is comfortable with here two systems that have more plugin compatibility of sound cards can rate it good instead of Linux that can run a serious card is a miracle.
I want to point out that some functions than in the magazines You must be a moment ...
as they enter steps

I'm a Music Technology

I'm a Music Technology graduate.
I have compressed files to 128, 64, 32kbps .mp3 and even as low as 8kbps using .mp2, just in the name of using the results to educate other people about lossy compression.
I understand the processes involved in mp3 compression, and have even poked around the source code of multiple mp3 codecs, which means I probably understand the topic better than 99% of people.
I am fussy about my sound quality, and I understand the implications of using lossy file formats.

However, sometimes someone comes to me with a compressed file and says "Here's the backing, let's record." Do I send them away and say "Find me an uncompressed file, and don't even think about paying me for my services unless you have one."? Of course I don't.
I use something other than Ardour to convert to .wav, so that Ardour will accept the file.
It's inconvenient.
It's more work for me, and I'd rather Ardour would do it for me.
I'm not going to stop using Ardour because of it, but I deplore the idea that the feature is deliberately omited because it's 'The Wrong Thing'.
If I wanted programs that limited what I can do, I'd go back to Windows.
I thoroughly agree with what timrob says. Especially his opening and closing statements.

@megamasha: I'm not going to


I'm not going to stop using Ardour because of it, but I deplore the idea that the feature is deliberately omited because it's 'The Wrong Thing'.
If I wanted programs that limited what I can do, I'd go back to Windows.
I thoroughly agree with what timrob says. Especially his opening and closing statements

Apart from the arguments for and against omitting a feature because it might encourage users to (knowingly or otherwise) do the "wrong thing", in the case of mp3 there is an important patent / license issue. This is something that can't simply be ignored just because its a little inconvenient - like them or loathe them, patents are very real and can cause significant cost / problems for a project.

Of course, one of the arguments often given in favour of open source is that if a feature is missing you can just add it yourself (which depending upon your level of skill may or may not be an option - its certainly "technically possible", and in that case you can choose how to address any license / patent issues should you decide to distribute your changes)

@megamasha: Mp3s are not

@megamasha: Mp3s are not supported by libsndfile because the code to support them would carry with it an obnoxious licensing fee: In other words, they are the "Wrong Thing" because of the way a patent-encumbered de facto standard forces the developer to take on the burden of the whole legal mess surrounding the patent, licensing fees, etc.

IMHO FOSS is about choices.

IMHO FOSS is about choices. Limiting a user's choices "for their own good" has serious implications. I'm completely on board with saying "mp3 licensing prevents us from distributing ardour with mp3 support". Any other reason given is counter to FOSS, it's philosophy and culture.

FOSS==choice. That's the whole reason it exists.

If a developer of the wireless software wpa_supplicant said "I'm not going to support open access points in wpa_supplicant because they are insecure" there'd be a fork the next day or people would use iwconfig.

Of course with the way labels have their recordings mastered now, I'd be honestly surprised if you could tell the difference in the end result. Overblown mastering kind of wrecks any sound quality and fidelity you may have had in the source material no matter how pristine it was.

I think The Black Keys have it right. Write your songs and record them any way you want. If they are good songs it won't matter how you record them. The only people "pristine sound quality" matters to are audiophiles and mix engineers. They won't help you get airplay or sell sides. The average teenager or people that buy concert tickets don't give any semblance of a crap. Your song either rocks or it doesn't.

IMHO the lack of mp3 support simply means I hook up my ipod to an input and record it again. _no one will care_. Do I work that way if I don't have to? Of course not.

I used to be a purist, been mix engineering since 1988, then I realized all that matters is getting the job done, having a good clock source, and keeping the hiss to a minimum if you have more than a few tracks you are mixing. Get your balances right, use a parametric right when you need to, and the world will turn. Where your source material came from or the format it was in are irrelevant as long as the end result works and the musicianship is good.

Yea I wince when I hear a 56k mp3, but I bet I'm the only person in the room that does.

I agree with megamasha. I run

I agree with megamasha. I run a professional recording studio, and all the time customers come (expecially singers) with an mp3 base and just want me to record their voice over it, mix and master, and I do it, cause it's my job. But I always have to convert the files before so that ardour can make use of them. It is a waste of time and drive space, and I think the professional way to do things is not the one that leads you to time and drive space waste.

VST have patent issues too, but you can build Ardour to make use of them. Why can't the same be done with mp3s?

@vervelover: there are no

@vervelover: there are no known patent issues with VST.

@vervelover: The issue

@vervelover: The issue regarding VST is not about patents (as far as I know) - instead its about compatibility with GPL which essentially resolves to this:

Steinberg have chosen to make the VST SDK available for free (e.g. no cost) provided that you sign up to their license agreement. The license agreement does not allow re-distribution of the SDK

Therefore, if you make a GPL application, you cannot include the SDK in the source code (which would be necessary to successfully build it from source), as to do so would imply it could be freely (as in freedom) distributed which is counter to the VST SDK license terms.

There is no problem with distributing binaries (if they are not GPL) as this doesn't involve distributing the actual SDK (header files)

The only problem with distributing binaries arises if they are built from GPL code, in which case the GPL insists that the source code be made available (and that makes it incompatible with the SDK license terms - as previously mentioned)

The way around this is to provide compatible header files under the GPL - which have been created without reference to the VST SDK and are therefore not covered by the genuine SDK license terms. This is what the Vestige headers provide.

However - mp3 is about patents e.g.

Even if a compatible 'mp3' codec was to be engineered without reference to any existing documentation - it would still be an infringement of the patents granted on the mp3 format.

- That's my understanding of it anyway - although I'm certainly not an expert in legal matters :)

let me jump in here

let me jump in here again.

first, as i think i already mentioned, i am not missing mp3 support in ardour, although exporting to mp3 would be sometimes handy. but however, i open up soundconverter and tell him to convert everything i need, so no problem there.

but my question is - how is k3b handling this? as i remember, you can install k3b from a repository and it won't be able to handle mp3's, but as soon you install lame too, k3b will find it and list it in ->preferences->programs and will be able to handle mp3's.

this is how i remember k3b woks, but it might be that i am not right.

also mhwaveedit, if lame is installed, i can compile it with lame support, and it will play mp3's after it. if there is no lame on my system installed - no mp3 support in mhwave.

could something like this be legally done in ardour too?



I understand pros and cons

I understand pros and cons importing mp3 files, but I want to import some sounds from "freesound" and I can't do that because the format is mp3. It is annoying a little.

I agree with @nowhiskey, that maybe you can add import mp3 via external tool, eg "lame"? And add path to lame in Ardour options...?


@skygge: freesound provides

@skygge: freesound provides .wav and/or ./flac format for most/many of their files too. the mp3 versions are generally just for previews.

egarding exchanging MP3/ogg,

egarding exchanging MP3/ogg, or even FLAC, as "quick and dirty" exchange format for tracks, one problem would be the scaling: Ardour uses floating-point WAV so there's no "hard 0db" limit to how loud a track can be, or any sort of real "bit quantisaton danger" level.

So at least you'd have to normalise and bounce each segment of audio before compressing it.

It might be a nice thing to have automated with "Ardour Session Exchange", as an option (not sure if ASE has been brought up to date with ardour3 or even ardour2).
It would remain "Quick&Dirty" even with HGH FLAC since it would compromise quality because you have to convert to 16bit integer audio.

Freedom is choice. With the

Freedom is choice.
With the development of Internet connections and hard drive MP3 format will lose its importance, now withdrawing from it is like prohibition.
You can always give a warning when test run using mp3 format: D

Ardour, much as I applaud you

Ardour, much as I applaud you for trying to 'educate' users about lossy codecs, I disagree with the approach of simply turning your back on MP3 for so many years for ideological reasons, especially when the LAME codec exists. Encourage people to pick a better format by all means, but don't sabotage your own bloody product by refusing to support MP3 for the sake of an ideal. So many people have asked for this now, there's clearly market demand. You can educate people without having to use the stick.

I am heavily involved with a radio network in the UK. We are open source advocates, running Ubuntu everywhere, however for our audio work we have had to choose Audacity / Reaper on WINE over Ardour for one simple reason:

- Record companies send us 320 kbps MP3s as standard format for early release demos

So, like our many producers and presenters, when I'm trying to construct mini-mixes and trails, Ardour has successfully written itself off as a suitable tool because it requires going through the process of transcoding to wav, where Audacity / Reaper support it out of the box. (See for what the network engineer has said about this)

My message to you is simple: do continue to educate users by providing information, but please abandon your arrogance and listen to what many of your users are asking about MP3 support. LAME exists, use it.

Just a brief technical note:

Just a brief technical note: LAME is an encoder, and isn't used to read MP3 files.

But I take your point. On the other hand, I'm not sure you actually got mine (which must be hard since you write this reply months after I took down the original, lengthy text). If I'm on a crusade on this issue (and I'm not sure I am) its not about licensing and FOSS. Its about audio workflow. Look, the existence of MP3 is traceable to one thing and one thing only: limited network bandwidth (ok, maybe too: relatively limited amounts of persistence storage). I see nothing particularly wrong with the format existing specifically as a workaround to these two issues, but these are issues that are already massively less important than they were when psycho-acoustic compression was first developed. What I really object to is the notion that material which has been processed in this way has some role to play in professional workflows. Everybody wants to point the finger the other way - you want to point at record companies, record companies want to point at studios, studios want to point at clients, clients want to point at their internet service providers and so on. What should have happened at some point is that people just stood up and said "yeah, sure, MP3 as a compact distribution and listening format is fine, but we don't produce music using this format". Instead, everybody wimped out because of the convenience - and other software companies ran along like hand maidens to make it possible - and now we have ridiculous scenarios where a radio station, which is likely to compress and otherwise distort the hell out of any music they play no matter what its origin, prepares broadcasts starting with psycho-acoustically compressed stuff.

But hey, I can't fight this battle alone. The producers and engineers wimped out very early, and I'm sure that MP3 import and export will come to Ardour in the not too distant future.

I have a a pound of salt In

I have a a pound of salt In my cupboard but that doesn't mean I'm going to sit down and make a meal of it. The option to use it is what's important. By not allowing me to even use an mp3 file will only keep me using other DAW's. This is why I shall continue to use Abelton.

@Brotherbbad: Ardour does not

@Brotherbbad: Ardour does not disallow the use of mp3 files. It just doesn't support it at present as part of an integrated work flow. You can decode them to wav (or whatever) and use the result. Abelton is a fine program - please continue to enjoy it.

Paul, it is most certainly

Paul, it is most certainly NOT true that Freesound sounds are typically offered in multiple formats (maybe you're thinking of Bandcamp?). In doing some small "audio theater" bits for my podcast, I have had to download some sounds from Freesound that were ONLY offered as MP3s. I'm not in love with the idea of using MP3s in my podcast, but ya gotta do what ya gotta do. Every time I found a sound that is MP3 only, I always left a slightly annoyed comment asking why they're not using a lossless format.

Additionally, we play music on my podcast, and artists will often send me 224k-320k MP3s to use on the show. So at the end of the day, I end up having to use MP3s way more often than I want to. Converting them with a different app is no big deal, but being able to import them directly would save one step.

This is a bit after the fact

This is a bit after the fact here (about a year!) but I wanted to put in my 4 cents, adjusted for inflation. I've worked as a studio engineer, a professional musician, a computer geek and have recorded, transcribed, converted, listened to, and appreciated more audio music files than I can ever hope to count in my many years in the field. Yes, mp3 is lossy. On the other hand, given the current limitations on email and cloud file storage and, especially in the US, consumer bandwidth, there are a number of compelling reasons to use a popular, compressed, lossy file format - many of these similar to the reasons people should use plain text for emails rather than the generally bloated and non-standard HTML or rich-text formats.

If I play you a song, I am conveying to you a set of musical ideas. If I need to learn a song, I want access to a copy of the song which contains all the musical ideas which the performer or composer put into it. It doesn't matter if it's audiophyle quality or not. That's not important. The mp3 format has become widely popular for good reasons, and _most_ people don't notice or mind the lessened quality of the format. I, personally, prefer to use ogg instead of mp3 since it's an open source format, but the fact that compressed and lossy file formats, whatever they are, are the most popular formats for exchange of music is a fact of life, and a very logical one. On the web, the most popular image format is jpeg, which is likewise a lossy format, and few if any browsers can display a tiff (lossless) image, nor would many people have the patience to wait for a web page containing multiple tiff images to load.

What _is_ an error is to pass off an mp3 file as a full-spectrum, high quality audio product, and I, also, deplore the popularlity of mp3 downloads as a popular replacement for CDs. An artist may spend thousands of dollars in a studio recording a product with a 16 - 25K audio spectrum range, and this quality is lost in an mp3. Nor can one expect to produce a quality CD product using mp3s as source material. Bicycles and SUVs both have their uses, but it's a big mistake to try to put bicycle wheels on your SUV.

The incredible explosion of musical creativity in the US and other countries in the 20th century was largely a result of the development of radio as a communications media for music. For a substantial portion of this time, the radio medium was AM, which necessarily imposes a 5K limit on the audio spectrum. This didn't stop the musical communication and inspiration spread about by late night clear channel broadcasts of shows such as the Grand Old Opry, nor did it stop the many great writers and pickers who grew up listening to this music on the radio.

So high audio quality has its time and place, but this is not all the time, nor everywhere. If I want to record a CD, I'll keep my recordings as wav files. If I want to share songs with my friends, many of whom are songwriters, I'll use mp3 or possibly ogg files. Likewise if I want to make a quality audio recording in a studio I'll probably be using ProTools rather than Ardour.

It would be very nice if Ardour for Linux could handle mp3 files. I have built it from source, and as I recall it built with mp3 support, but Ubuntu desktop doesn't provide it with this support as a pre-compiled binary. Ardour for my Apple iMac, on the other hand, does, so if I need to work with mp3 files in Ardour I just switch to the Apple side of my virtual machine setup and use it there.

Ardour 3 can import and

Ardour 3 can import and export .ogg files, so whatever may have been said in the past, its lack of MP3 support clearly isn't about audio quality.
The reason why Ardour doesn't "do" MP3 must therefore be the legal issues around free software, patents and licensing.

Locking the thread as more of

Locking the thread as more of the posts are spam than content.

-- Seablade