Can someone suggest me a good software and codec to create 32bit tracks.
lossy files doesn't have a specific bit depth.
edit: the decoder decides the output bit depth.
Normally audio is encoded in 8/12/16/24/32 oftenly the user chooses the bitdepth
since an 8bit track would sound different than a 32bit track. At least this is what I heard from some samples.
Do you mean 32 bit resolution or 32 kilobits per second?
Well, was this post really meant for "AAC - general" forum ?
Well, was this post really meant for "AAC - general" forum ?
[a href="index.php?act=findpost&pid=252024"][{POST_SNAPBACK}][/a]
Since AAC is the high end of professional audio, I thought that it would fit in here, if you feel its out of place, don't bother moving, or delete it.
I meant 32bit resolution.
My apologies
I entirely fail to see how your question is related to AAC format. Please learn to read forum descriptions instead of posting a question in a random forum then wondering why noone understands what you mean.
why do you want to create 32bit files? there are no a/d converters that really reach 32bit resolution. i just heard that dsps can work better at high resolution / sample rate if you for example want to use equalizers in a recording software like cubase or samplitude...
Perhaps 32bit files are better than 24bit ?
For editing if you are going to do a lot of operations on them, possibly. For encoding, absolutely not.
You officially have no idea what you're talking about. Somebody please delete this thread or move it.
What does make you think that ?
Perhaps 32bit files are better than 24bit ?
[a href="index.php?act=findpost&pid=252117"][{POST_SNAPBACK}][/a]
i tell you what: 64bit is even better.
Do you mean 32 bit resolution or 32 kilobits per second?
Strange guy who means 32kbps...
why do you want to create 32bit files? there are no a/d converters that really reach 32bit resolution.
What does the conversion when recording directly to HDD ? Is it a soundcard's hardware A/D converter or is it the software ? If the latter, isn't it A/D converted to (say) 32bit directly ?
chaotic thread, really
I asked a simple question, just wanted an answer either a yes or a no.
Since I have been playing around with 8-24 bit audio and eventually saw a great difference whilst editing, I just wonderered if it would be a better choice encoding
directly to 32bit.
If you use lossy codecs, you don't need to bother with bitdepth at all I think.
I don't know about 32bit capabilities of lossless codecs. Monkey's Audio can't.
Anyway, the quality will never become better during encoding by just increasing bitdepth, it stays the same then.
What are you trying to do?
Lossy audiocodecs don't care about input bitdepth. The difference between using a 16bit or 32bit source stream will be inaudible/negligable, since most of the differences will be 'cut away' by the codec's psychoaccoustic model anyway.
I want more bits! More more more! Infinity+1 bits of sweet sweet music!
On the input side, the highest resolution you will find out there is 24bit, 96kHz. Maybe there are 32bit cards out there, but they will be mega expensive. And as some people have said before, those 24bit ADCs most likely won't give you a real full 24 bits of resolution.
If you want, capture at 24bit, or whatever, then resample to 32bit before messing around with the files. I mean, if it'll give you peace of mind.
On the lossy encoding side, in my experience, LAME will accept 32bit integer samples, and FAAC accepts 32bit floating point. These are both from experiments with Foobar2000 and its Diskwriter module sending data to the command-line versions of LAME and FAAC. I haven't tested any other codecs.
On the input side, the highest resolution you will find out there is 24bit, 96kHz. Maybe there are 32bit cards out there, but they will be mega expensive. And as some people have said before, those 24bit ADCs most likely won't give you a real full 24 bits of resolution.
If you want, capture at 24bit, or whatever, then resample to 32bit before messing around with the files. I mean, if it'll give you peace of mind.
On the lossy encoding side, in my experience, LAME will accept 32bit integer samples, and FAAC accepts 32bit floating point. These are both from experiments with Foobar2000 and its Diskwriter module sending data to the command-line versions of LAME and FAAC. I haven't tested any other codecs.
[a href="index.php?act=findpost&pid=252269"][{POST_SNAPBACK}][/a]
Thanks LoKi128 thats the kind of answer I've expected.
You know whats the problem with some of you guys that actually you seem to be happy with the results achieved by 24bit audio samples, yes fine but a couple of years ago people where also happy with 16bit audio samples etc. If you think this is a stupid and chaoctic thread, you should be ashamed. 32bit is something uncommon thats why some of you pretend to know alot but in reality nothing.
In not a question of having more bits +1 (lamer) its a question of finding a way to improve and not just stick to what we find optimal.
Imagine this ---> Listen to an 8bit sample
NOW Listen to a 24bit sample
Different hey (assuming you have good ears) thats what i'm trying to explain.
Then comes 32bit which certainly has much more capablities to improve sound.
Imagine this ---> Listen to an 8bit sample
NOW Listen to a 24bit sample
Different hey (assuming you have good ears) thats what i'm trying to explain.
Then comes 32bit which certainly has much more capablities to improve sound.
I think maybe the point many a few of these guys were trying to make is that going from 24bit to 32bit is even
more imperceptable than going from 16bit to 24bit. I know I can't tell the difference, either way. I imagine most people cannot.
{sigh} TOS #8 anybody? I'll do the honors...
Around here, you are not allowed to make statements about audio quality without producing some kind of objective proof. You can't say 32 bits is better than 24 bits without showing us through double blind tests that you can tell the difference. Come back when you have.
I highly doubt you will be able to, however, due to the following simple math:
0 dB SPL is the threshold of human hearing, the quietest sound that anybody anywhere in the world has ever been recorded as perceiving. Your personal threshold is probably a bit higher. If you play back a 24 bit file so the peaks are at 144 dB SPL, the smallest details that were recorded will be at that threshold. However, you will not be able to perceive those details, since the blast of sound at 144 dB SPL will render you temporarily deaf within seconds. (For reference, a rock concert is about 100-120 dB SPL)
16 bits, with a dynamic range of 96 dB, is just fine for the final master product. Higher precision is only needed to avoid rounding errors when further processing will be done to the audio, as well as keeping the noise floor managable when mixing several tracks together.
Around here, you are not allowed to make statements about audio quality without producing some kind of objective proof. You can't say 32 bits is better than 24 bits without showing us through double blind tests that you can tell the difference. Come back when you have.
As I already explained, few applications can really support 32bit files and decoding them on common soundcards (16/24bit) would be a waste of time. Therfore my intention was to try to explain that this idea of using 32bit is still open in such a manner that only the future can really prove you the right answer.
Perhaps a better place to start would be comparing 16-bit with 24-bit at the same sampling rate. You'd need a soundcard capable of outputting at 24 bits per sample directly and you'd probably want to have a 24-bit recording to start with. For getting the 16-bit version, you should create two versions, one made by simply truncating the last eight bits off the end of each sample and one dithered to see if that makes a difference.
If you're wondering, I'm not orchestrating a similar test for myself mainly because I don't have any 24-bit equipment to use with the test.
I only have some samples 16/24, right now I'm trying to play some files, which seem to play but nothing is heard. Probably something to do with the soundcard.
I said this thread is chaotic because I can't do anything with posts like this:
I want more bits! More more more! Infinity+1 bits of sweet sweet music!
On Bob Katz' site there's this:
This 16-bit CD illustrates that there is no noise floor problem at high monitor gains and that it is a myth that 16-bit CDs have to be compressed or limited to fit in the medium! After all, CDs have a measurable 115 dB dynamic range (properly dithered)--noise floor is NOT a problem.
I totally agree with analogy's post.
In fact, we could all be perfectly happy with CD for the rest of our mortal lives. These 115dB are such a great dynamic range that you get deaf from a starting jet engine at 115dB, while you don't notice anything than your heartbeat and rushing blood at 0dB.
The greatest (RMS) dynamic range that ever occurs may be about 50dB for classical. That is very very much. So there's still hughe headroom for the peaks (should be at least 20dB) and footroom left. Today's pop CDs have only a RMS dynamics of ~3dB and the same for peak -- insane, and it's ridiculous to talk about >16bit resolution.
In addition to my previous post:
For editing, a higher resolution is, of course, still recommended.
Also, 32bit is available as a float format which allows sample values above 0dBFS. That is the real advantage: no need to fear clipping. It's especially useful for clip restoration, because it then provides the neccessary headroom for the added peaks.