Jump to content

Suggestion: FLAC remixes?


nonoitall
 Share

Recommended Posts

Well, most of the albums have lossless files, so why not the regular remixes? I know, bandwidth would be a concern, but maybe lossless files could be a torrent-only feature? This would have benefits, not only for audiophiles, but for anyone who wanted to have the songs in a format other than MP3. Transcoding from FLAC to OGG, is very preferable over transcoding from MP3 to OGG, for example.

Link to comment
Share on other sites

Well, most of the albums have lossless files, so why not the regular remixes? I know, bandwidth would be a concern, but maybe lossless files could be a torrent-only feature? This would have benefits, not only for audiophiles, but for anyone who wanted to have the songs in a format other than MP3. Transcoding from FLAC to OGG, is very preferable over transcoding from MP3 to OGG, for example.

You'd be better off suggestions to individual artists that they make their music with your suggestion in mind; as far as OCR goes, suggesting this idea to album project coordinators would also be something you could do to promote this. But the short of it is that we presently wouldn't consider implementing anything like that as far the regular remixes.

MP3 is a standard near-universal format, so we stick with it. FLAC, correct, is for audiophiles, but it's not for the average users, and nothing is seemingly taking over the MP3 standard any time soon.

OGG/FLAC standards have been suggested and considered in the past, but unless there is a new industry standard, then it won't be considered, even on just a torrent level. We cannot reasonably expect all submitting artists to provide FLAC versions of their mixes or send us WAVs, and enacting anything like that would only limit our submissions pool.

Link to comment
Share on other sites

Yeah, my main rationale for the lossless idea isn't so much that I want perfect sound, but that lossless files can be converted to any format without cumulative compression artifacts from multiple lossy codecs. Let's say Bob wants to put all ~1,500 remixes on his 2GB player or flash drive. (He just can't choose any remixes that he can live without.) He'd probably have to squeeze them down to around 56kbps. 56kbps MP3 will sound pretty awful, and transcoding to any other codec won't produce optimal results. A 56kbps HE-AAC encode produced from a lossless source, on the other hand, wouldn't sound too bad for casual listening to most people. (Certainly a vast improvement over an MP3 of the same bitrate.) Anyway that's just one of innumerable examples I could come up with.

Do artists currently have the option of submitting lossless files, and does OCReMix retain them if they are submitted that way? I can understand that some remixes might simply be unavailable in a lossless format, but perhaps lossless could be promoted (though not required) for future submissions, and for any past remixes that artists would be kind enough to resubmit in lossless form. Then, whatever remixes are available as lossless files could be distributed via a torrent or however is convenient.

Link to comment
Share on other sites

I think the point is that OCR has a file format consistency, and any radical change at this point would be unfavourable as the previous 1600+ remixes would have to change aswell. Currently every remix has to be under 6mb, so say Disco Dan heard that we now accept lossless remixes. He'd be pretty peeved that his Triforce majeure (the most popular remix according to ormgas still i believe) had to be encoded to 113kbs when he submitted and now everyone else can submit lossless. If he doesn't have the project files anymore, then thats tough shit?

Although the lossless would be a better sound quality, the accessability to joe average aswell as the file sizes make FLAC/WAV unfavourable overall.

Pretty much the only thing I see happening is someone hosting their own lossless archive and asking for contributions from the mixers, but I doubt anything official would be done.

Link to comment
Share on other sites

The main problem is that all the current OCRemixes are MP3, and hunting down every remixer and having them give you the mix in a lossless form would be impossible. I wager a lot of the source material has even been lost, just leaving the MP3 as the only copy of the music out there.

Link to comment
Share on other sites

Torrent distribution requires seeding for literally hundreds of files, and then you'll suggest batch torrents with selective leeching but that means you're telling average joe computer user to go grab utorrent or anything with that capability. Also, consider those who don't have the computer smarts to convert their songs if we go with lossless direct download distribution.

Link to comment
Share on other sites

I agree that it would be nice to have lossless files of OCR music. I'd kill to be able to convert all 1600+ songs to OGG format... but it's just not going to happen. Even if the DJP Cru decided to allow for FLAC submissions (let alone offer old remixes in lossless format), it would be a logistical nightmare. There's almost no doubt bandwidth costs would rise. It would be extremely difficult to keep track of what remixes had the files and which ones didn't. And you can forget about contacting all the remixers.

That and it would be a real bitch to upload them all.

Link to comment
Share on other sites

Something I've wondered about is mp3 bitrate. I've spent quite some time on certain private music torrent trackers and these days I have a hard time listening to anything below 192kbps. On one hand I can understand the 6mb filesize limit, on the other... what are production values worth if we're listening to a poor quality file?

Link to comment
Share on other sites

FLAC is just not feasible on a lot of levels. To respond to a few things..

Something I've wondered about is mp3 bitrate. I've spent quite some time on certain private music torrent trackers and these days I have a hard time listening to anything below 192kbps. On one hand I can understand the 6mb filesize limit, on the other... what are production values worth if we're listening to a poor quality file?

The vast majority of people don't have a problem with 192kbps. Hell, the vast majority of people don't have a problem with 128, which is the iTunes standard. Not everyone is an audiophile. In fact, hardly anyone is, so catering to that tiny niche is not worth a great expenditure of time and effort. Additionally, saying "what are production values worth" is highly ignorant... artifacts from encoding are negligible at 192, 160, 128, and really even at 64, compared to the impact of a wide range of production issues. If you choose an unrealistic flute sample it's going to sound unrealistic at any bitrate.

Yeah, my main rationale for the lossless idea isn't so much that I want perfect sound, but that lossless files can be converted to any format without cumulative compression artifacts from multiple lossy codecs. Let's say Bob wants to put all ~1,500 remixes on his 2GB player or flash drive. (He just can't choose any remixes that he can live without.)

I don't think this is a good hypothetical. It is inexpensive to purchase a player that can store all of OCR's library. Most people have pretty nice MP3 players these days. If you don't have one, you have your computer and/or CD/MP3 players. It's not the end of the world to have to switch CDs or files around every so often, not that many people even do that these days.

A 56kbps HE-AAC encode produced from a lossless source, on the other hand, wouldn't sound too bad for casual listening to most people.

How many people actually know how to produce good encodes? And of this niche that doesn't have a decent-sized MP3 player and can encode properly, what are the chances their tiny player will support that encoding format..?

Do artists currently have the option of submitting lossless files, and does OCReMix retain them if they are submitted that way?

It's not an option, and we don't retain them if they are submitted that way. We have a 6MB filesize limit for bandwidth reasons, why would we allow people to submit files several times that size?

I can understand that some remixes might simply be unavailable in a lossless format, but perhaps lossless could be promoted (though not required) for future submissions, and for any past remixes that artists would be kind enough to resubmit in lossless form.

I would estimate that over 80% of the OCR repertoire is not available in lossless format and the source files are either gone or corrupted. That is a conservative estimate, too.

Then, whatever remixes are available as lossless files could be distributed via a torrent or however is convenient.

There is no convenient way of distributing them via HTTP; that is really out of the question. If we were to create a lossless torrent, it would at best be 20% complete, and only provoke an unending stream of inquiries.. "what's flac?" "where are the rest of the remixes?" etc...

The bottom line here is that lossless/FLAC would be appreciated by SUCH a small niche that it really is not worth all the hassle required to make it an option on a site-wide or official level. If you want lossless versions your best bet is to contact individual remixers.

Link to comment
Share on other sites

I think the point is that OCR has a file format consistency, and any radical change at this point would be unfavourable as the previous 1600+ remixes would have to change aswell.

Other netlabels, like Kahvi, did mp3 for a while, then went to .ogg for a long time, then went to both .ogg and .mp3 after that. Of course, this isn't kahvi, but the point is that mp3 and ogg coexist well and you can have either or...

Also, I can definitely release in .flac if you want, but only for my newer stuff that I still have the source files for.

Maybe put a notice out to all the artists, that maybe they'd like to reencode their songs to .flac and you can put together your own batch torrent with select songs? Obviously not all are going to respond because they're either long gone or... dead :( but hey I'm interested.

Also: I kept suggesting .flac to zircon for the VotL distribution, but he didn't even put the .wav files in any archive compression whatsoever. Just to spite me... bastard.

Link to comment
Share on other sites

Has the file size limit been increased since OCR's inception? If not, it seems really lame to me that a limit created at a time when dial-up users probably outnumbered broadband users hasn't been changed to be consistent with the times. It really sucks to be listening to one of the longer mixes and to hear artifacts because of the limit. I hear artifacts at 128 kbps and once you get into lower than that its just....blah. Sure, I guess the people who made long mixes with the current rule would feel a bit screwed if the file size got increased, but why hold back for a few bruised egos and hurt feelings?

If it has been increased, ignore the above rant.

Link to comment
Share on other sites

The file size limit has not been increased. It has less to do with people being on dialup vs. broadband (though I guess that's part of it) and more about our own bandwidth and storage. I think we (the staff) have tossed around the idea of possibly raising it, but I don't think there is a pressing need to. It's one in a thousand that we get a mix that legitimately suffers because of the rule.

Link to comment
Share on other sites

Another advantage of the 6 mb limit is that it pushes people to do away with excessive filler and get to the mix. For example, the only time this rule was EVER broken was on April Fool's Day, 2006, when Dave posted the 11 mb, 15 minute 'Terra Fortnight Megamix' (Which I still have) . Even though it was trying to be funny, you quickly realized that if a mixer had a higher size limit, they'd also have simply too much time to play around with. The temptation to simply fill time might be too great.

Link to comment
Share on other sites

Y'know, I've yet to hear that mix.

But yea...what's been said is all true about higher file sizes, higher bitrates, and that whole shebang. I don't really think it necessary myself and it's more of a hassle, isn't it? Don't FLAC files use something other than your typical MP3 program? (I wouldn't know, at highest "quality" I've used mp3 and wav.)

Link to comment
Share on other sites

  • 2 months later...

Hey, I know it's been a couple months, (I was out of the country for a while and kind of forgot all about this topic). But I was reading an article about YouTube the other day and it brought up some fresh thoughts that seemed bump-worthy. (YouTube is experimenting with higher quality videos.) Besides, it's my topic and I'll bump if I want to. You would bump too if it happened to you! (Ahhh, is anyone old enough to get that? :-))

Anyway, I guess it just made me think about the argument that it would cause logistical problems. It's not like YouTubers who submitted clips in a low quality form are really upset over the upgrade - it's just a technology upgrade that was to be expected. The thing is, while MP3 is pretty much ubiquitous, more and more people want better quality or smaller size, even if they don't understand exactly how it works yet. (Honestly, do you think all the people who buy an HDTV realize that they're not taking advantage of it when they watch (unbeknownst to them) standard definition material like DVDs and regular satellite TV?)

So, as far as it only being appreciated by a niche of people, I kind of have to disagree. I would agree that only a small (but growing) niche of people understand what lossless files are and know to take full advantage of them at this point, but there are many people who just "want better quality" or "want smaller files".

I'm about to quote you a number of times, zircon. I consider all of your points to be valid and am not trying to argue with you at all - I'm just presenting my thoughts on the issues you remarked about.

I don't think this is a good hypothetical. It is inexpensive to purchase a player that can store all of OCR's library. Most people have pretty nice MP3 players these days. If you don't have one, you have your computer and/or CD/MP3 players. It's not the end of the world to have to switch CDs or files around every so often, not that many people even do that these days.
Aw, but Grandma just bought Bob that 2GB player, and Bob has enough trouble paying for college without having to worry about buying another (still rather expensive) MP3 player. (No, I'm not Bob - this is still just a hypothetical. :-)) And, like you said, people don't swap files/CDs around their players so often these days (it's inconvenient) so, all the more reason for Bob to want all of them on there at once.

And really, when you add up the niche that specifically wants lossless, and the niche that specifically wants OGG, and the niche that specifically wants 320kbps MP3, etc, etc, etc, you end up with a fairly significant "niche". Many people want music in a variety of formats for a variety of reasons. Lossless files are a one-size-fits-all solution that could address all of them.

How many people actually know how to produce good encodes? And of this niche that doesn't have a decent-sized MP3 player and can encode properly, what are the chances their tiny player will support that encoding format..?
Many players support formats other than MP3 (AAC and WMA are very common, both being superior to MP3) and often the software that comes with these players allows users to convert to those formats without even realizing it. Firmware updates can also make literally dozens of other formats available. (Anybody else a Rockbox user?)
We have a 6MB filesize limit for bandwidth reasons, why would we allow people to submit files several times that size?
If, for nothing else, to preserve the original remix perfectly and retain it in the event that, someday, alternatives to MP3 are utilized.
I would estimate that over 80% of the OCR repertoire is not available in lossless format and the source files are either gone or corrupted. That is a conservative estimate, too.
I'd agree with your estimate. But color TV was adopted, despite the fact that most of the material of the time was black and white. The same is gradually happening with high-definition television today. The quality lost in many past remixes may indeed be irretrievable, but why doom all future submissions to the same restrictions when the technology now exists to preserve them perfectly?
There is no convenient way of distributing them via HTTP; that is really out of the question. If we were to create a lossless torrent, it would at best be 20% complete, and only provoke an unending stream of inquiries.. "what's flac?" "where are the rest of the remixes?" etc...
This I definitely understand. I've been a victim of plenty uncomplete torrents myself. But don't forget the hidden distribution that goes on on P2P networks and the like. I think just about anyone who can use torrents can use Limewire, and speaking just for myself, if lossless downloads were available, I'd share them on those networks whenever I was on the 'net. If torrents just turn out to be out of the question, maybe a limited number of lossless downloads per day could be permitted by HTTP.

Anyway, everything aside, would it hurt simply to allow (not require) lossless submissions? Even if they aren't used for now, at least more remixes would have the potential to be distributed this way if it becomes more plausible in the future.

Link to comment
Share on other sites

No, people are right when they say FLAC is preferred by a niche audience. I have more music than probably anyone else on this site (could be wrong, but I've yet to come upon a single person with more music than me on a hard drive), and yet I have no need for FLAC - it still takes up an incredibly excessive amount of space, and it generally is a pain to work with compared to mp3s with that additional step of converting the file format if you want to use the music with other devices. In addition, 192 kbps mp3s sound just fine - I even have a hard time at times differentiating with mp3s encoded at lower bit rates compared to their higher quality counterparts (e.g. the FF7 project mix 'Valse Aeris'), and my hearing has always been excellent in terms of attentiveness to details.

Probably the biggest blow, most people still don't even know what FLAC is, or if they do, they still tend to not care. In fact, of all the communities related to music I've been around, over here is the only place I've seen interest for FLAC, and even here it's in the minority.

Logistically, FLAC is unreasonable for a website like OCR since as zircon mentioned, bandwidth is still a limiting issue - quite frankly, our broadband here in the US blows. Also as zircon mentioned, it is fairly uncommon that a song that would pass the standards would break the file size limit with all the alterations. Just to use the Valse Aeris example again, the OCR version takes up only 5.4 MB and at 90-something kbps (92? I forget) - the project version takes up 10.1 MB at 192 kbps. Excepting the bit rate difference, they are identical songs.

I've yet to see any argument for FLAC with more pros than cons. Maybe once larger hard drives (including larger portable storage drives as well) & far improved internet become ubiquitous, then a format like FLAC would make complete sense. Currently however, it's just too much inconvenience for a small amount of benefit.

Link to comment
Share on other sites

But like I said, it's not just a matter of quality and people who specifically want FLAC. Sure, 192kbps MP3 sounds great, but even so, a lot of information has been removed. This doesn't make much difference to someone who is content to listen to the 192kbps MP3. It does make a difference to someone who wants to convert the remix to a different format. (Compression artifacts are cumulative.) And not all the remixes on the site can even afford to be at 192kbps.

So, if you want anything that is not exactly the format that's on the site, you're stuck with sub-par quality. Lossless files completely eliminate the problem. Audiophiles are a niche, true. But this is a remix site. It's bound to get a higher percentage of audiophiles than your average site. Not to mention, audiophiles aren't the only ones who would have use for such files. Anyone who wants something that's not exactly what the site offers could be accommodated.

But let's say that's simply not possible.

Even if there simply isn't a shred of bandwidth available for even one lossless download per month, I don't see what could possibly be hurt by allowing and retaining lossless submissions. Even if every single remix was resubmitted losslessly (which is undoubtedly impossible) they would only occupy around 30GB of space. If no one seriously has the hard drive space to store them, I'll donate a hard drive (or two for redundancy). I'll even write a script to automate the conversion from the retained FLACs to distributable MP3s, if you don't have one already.

I can't see a single con to this approach - it gives remixers an additional option when submitting their songs, and someday, when bandwidth is cheaper (or sooner if someone else can host the files), those perfect copies can be distributed. Forcing the retained copies to be lossy is like forcing your 8 megapixel camera to take 0.3 megapixel pictures because they look okay on your current equipment. Sure, they look fine now, but then when you have the means to appreciate an 8 megapixel photo, you'll realize you weren't acting with much foresight when you took the pictures.

Now, I'm not saying the site's policies are stupid or anything. But why settle for adequate when you can have perfect saved away, just in case it could be useful someday?

Link to comment
Share on other sites

Just to use the Valse Aeris example again, the OCR version takes up only 5.4 MB and at 90-something kbps (92? I forget) - the project version takes up 10.1 MB at 192 kbps. Excepting the bit rate difference, they are identical songs.

I had no idea that the OCR version of Valse Aeris was encoded so low, so I was rather curious about how well it sounded. So I downloaded and listened to it myself...

And, dude... I can hear audio compression artifacts ALL OVER the song. Even without comparing it to the original, the string section often sounds warbly, the cymbals deteriorate into a noisy mess, and whenever several instrument sections are playing simultaneously there is a terrible amount of shimmering noise in the background.

I honestly didn't expect the compression to be quite so apparent, yikes. O.O

Link to comment
Share on other sites

Valse wasn't encoded at low CBR, it's VBR.. the best I was able to do under the circumstances. Still sounds pretty good to me.

I know, site policy dictates remixes must be under 6MB. I'm not accusing anyone of encoding it poorly, as I understand that such a low bitrate was necessary for it to be posted. However, given what I hear, I think Bahamut should be careful using that song as justification for his argument. ;p

Though to be honest, I'm kinda unsure how it sounds "pretty good". Perhaps the fact that I'm listening to it at high volume in my headcans is helping the artifacts stand out.

If ya don't believe me, here are a couple short snippets from Valse Aeris, comparing a ~200Kbps VBR encode (of the project WAV) with the encode on OCR. Playing the two back to back, the cymbal distortion and string warble on OCR's encode seem pretty noticeable:

Valse Aeris clip 1 - OC ReMix @ 103Kbps

Valse Aeris clip 1 - High Quality @ 198Kbps

Valse Aeris clip 2 - OC ReMix @ 103Kbps

Valse Aeris clip 2 - High Quality @ 198Kbps

(NOTE: Because these MP3s were frame-cut to avoid any re-encoding, some audio players may display erroneous time and bitrate information. The audio itself, however, is NOT affected by this.)

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...