The Mana World Forums

Curious anyone have an opinion on switching all the music to 48000 from 44100 ?

not me, would be a good change.

regards,

Currently, we have one track at 48000 and the rest at 44100. You have a pull request that changes one of the 44100s to a 48000.

a quick check has showed that 48000 is probably the native frequency of your sound card.

Well most modern cards default to 48000.

Most digital audio is sampled at 44.1 kHz, a standard no doubt related to CD-DA, while sound cards were often designed to process audio at 48 kHz. So, the 44.1 kHz audio must be resampled to 48 kHz.

But doing this produces digital artifacts if done wrong (clipping and popping).

I resampled it to 48000 to satisfy sound cards and improve sound quality. have straelyn working on encoding the other tracks to 48000. I put this one up as a sample.

Thanks for explaining the details (and historical reasons) between the different bitrates.
It certainly sounds like 48000 is a better choice for computer audio.

lol! You think upsampling a sound file that is 44.1 to 48 khz will improve quality? Your funny.

BTW, 44.1 == music, 48 == video and broadcasting. Every single sound card out there supports 44.1, so if you hear pops and crackles, then the problem is elsewhere.

Um, he said that if the conversion to 48 is done improperly, you get problems. And no one thinks that just because it's a bigger number means it's better. (There are, however, *other* technical improvements being done at the same time).

And while the kind of sound card that professionals use can undoubtedly play 44.1 just as well as 48, in the *real* world we have sound cards that will only play at one native frequency and the resampling is done dynamically in software. If you're lucky, this will be entirely transparent to the user, but some of us have had to mess with our ALSA settings (for some reason, plugging in a webcam makes it want to output the sound to it).

Socapex wrote:lol! You think upsampling a sound file that is 44.1 to 48 khz will improve quality? Your funny.

BTW, 44.1 == music, 48 == video and broadcasting. Every single sound card out there supports 44.1, so if you hear pops and crackles, then the problem is elsewhere.

Most sound cards already scale it before you hear it, this removes it from having to do it and 44.1 as I pretty sure people aren't playing this off their cd but through a PCM device i.e. Sound Card.
http://en.wikipedia.org/wiki/Sampling_rate

Most audio processors/sound cards contain DAC for both 44.1 kHz and 48 kHz, being able to natively output either, though some older processors include only 44.1 kHz output, and some cheaper newer processors only include 48 kHz output, requiring digital sample rate conversion to output other sample rates. Similarly, processors may be able to record natively at only certain sample rates.

There are clips and pops in the tracks. Alot has to do with things outside of sampling rate.

Most likely with a newer card the music doesn't work.

And to anyone other then an audio file the improvement isn't heard.

More on this. Redoing the sample rate is moot as SDL in the client resamples it anyways.

Just to make things clear,

Up-sampling a 44.1 KHz file to 48 KHz will not improve the sound quality. It can not, because the information isn't there in the first place. Even worst, some audio engineers, such as myself, will argue that it will degrade audio quality. The reason is simple, you need interpolation to up-sample a 44.1 file to 48. So unfortunately, you are relying on the algorithm used to upscale the sample rate, which means the values aren't exact, which means there is a loss of information.

Now, if somebody with ear-buds or computer speakers hears the difference is another question. To which I will reply no, of course not! Nobody hears the difference. People like mp3 lol.

If I up-sample in Pro Tools, I know its gonna be fine. I do not necessarily trust open source audio software, that isn't to say I am right and that their interpolation isn't good.

For sound cards, the problems you are pointing out is Linux and probably drivers. Sorry to hear that, have fun resampling

Then go correct wikipedia.
http://en.wikipedia.org/wiki/Sound_quality

There is nothing to correct.

48 KHz is better, when things have been recorded, processed and down-mixed in that format. The question was about up-sampling from 44.1 to 48, which you wont hear the difference (and neither will I) and relies on an algorithm to "fill in the blanks". Or to be more precise, interpolate the blanks between the original values (if that makes any sense).

Here's something you could relate too, because people think visually a lot more. Open a picture in Gimp. Save it as a JPEG with the worst compression (lowest quality). Now open this jpeg in gimp, and save it as a PNG or a TIFF or any RAW/uncompressed format (high-quality format). The artefacts from the jpeg lossy compression will still remain, even though it is now in a "theoretically" better format.

Nobody would even think to say "save all your jpegs to uncompressed tiff to make them look better". I hope you understand more now

I think you should definitely go 48KHz if other formats are causing problems. But just don't proclaim it will sound better in your case. ^_^

Socapex wrote:Just to make things clear,

Up-sampling a 44.1 KHz file to 48 KHz will not improve the sound quality. It can not, because the information isn't there in the first place. Even worst, some audio engineers, such as myself, will argue that it will degrade audio quality. The reason is simple, you need interpolation to up-sample a 44.1 file to 48. So unfortunately, you are relying on the algorithm used to upscale the sample rate, which means the values aren't exact, which means there is a loss of information.

Now, if somebody with ear-buds or computer speakers hears the difference is another question. To which I will reply no, of course not! Nobody hears the difference. People like mp3 lol.

If I up-sample in Pro Tools, I know its gonna be fine. I do not necessarily trust open source audio software, that isn't to say I am right and that their interpolation isn't good.

For sound cards, the problems you are pointing out is Linux and probably drivers. Sorry to hear that, have fun resampling

I totally agree with this and explanation cannot be clearer. I may just add that the ogg format wich is used here is not lossless, so upsampling will result in a second compression. thus an extra quality loss. As you have not the original files or hardware onwich they were recorded, there is nothing you can do to avoid this loss.
It is false that computer hardware do not recognize 44.1 and this still for a long time, just because audio industry would have to change it's formats. Personally I use Digital Performer & Amadeus pro, but at the moment, I didn't notice any major issue with Audacity's interpolation. Playing 44.1 at 48 rate is likely to result in a small transposition but this may depend on the libraries. I fear that the amount of work is not worth the result.

Also I already agreed it's not needed. Manaplus does it on the fly already.

If you need it, I have a script to automatically conform audio to different formats. I wouldn't mind sharing it