I am currently trying to remaster a CD of my band for 24/96. My problem is that no matter what I use to upsample the material from 44.1khz/16 bit to 96khz/24 bit I get audible distortion in the form of slight crackles and pops. All of the attempts I have made have been software based. Any help anyone can offer with regard to fixing this problem would be greatly appreciated. Thank you.
what is the file format you are using, and what programs have you used to convert the file. It is also possible that the upconvertion is fine, and you're audio card is having trouble playing the file if its not meant to play 96k files. are you planning on changing any of the audio once its upconverted? if not its kinda pointless, its not going to sound any better. for best results you should re-record your band and record at 96k, if your hardware allows it
man I always forget about that encyclopedia site until I see you post it! I'm finally bookmarking it this time for future reference and referrels. thanks Ced.
Have you tried using the superb application R8Brain Pro, from Voxengo? I use this all the time for SRC as it is simply the best out there. There is a free version too, but it's not as good. Feel free to ask for tips on usage. Another thing that comes to mind here is the following: 1 - Is there any DC offset in the 16/44.1 files? 2 - Have these files been dithered already? 3 - Have they been heavily compressed or maximized? If any of the above apply, try this: Re record the 16/44.1 files back into your system again, as if you are recording a new track. Do this from your CD player into your converters, straight into your Audio workstation at 24/96. This should then eliminate all or any of the above issues. Good Luck.
Id say its pretty obvious what the problem here is. If the original track is encoded at 44KHz you cant make the quality any better than 44KHz re-encoding the track with a higher bit rate will add distortion. Youll either have to re-record the tracks to WAV format (wav is true sound and perfect quality) or just make do with the 44KHz files you have
It may be "pretty obvious" for you, but upsampling is actually pretty common. Especially if you want/need to work with the audio, add filters and equalize. If you do this on upsampled audio you're limiting roundoff errors. What is needed, is as Wilkes wrote, a program that is good at resampling.
up-sampling doesnt do anything but make files bigger, you cant add sound quality onto something if it has none in the first place.
Twaddle, sir. If you are going to carry out any further processing at the higher sample rate, then of course it is going to be all improved in quality. Especially if the tools you are using are giving superior performance & quality at the higher sample rates too. What is important, as has been posted already, is the quality of the SRC used. If you are using a crap SRC, then you will be right - but if you are using a high quality one - with no audible artifacts in the transition band and so on, then it will be a superior sounding file at the end of the process. Best to avoid any Linear Phase stuff here, as you run a risk of adding pre ring, but if you use a linear phase converter, or a seriously good quality outboard converter, you are effectively re-digitizing from analogue at the higher rate. This is where the quality jump comes from.
how can you expect to take a track thats original sample rate is 24Kbps and therefore has no information within it for sound outside of that bandwidth and increase the sample rate, sure the sample rate will increase but the whole reason that there are crackles and pops are because of how much more the rate has been increased. to show you what i mean take a look at this, this is an mp3 encoded @ 64KBps - notice the size of the peaks and here is a picture of the same mp3 re-encoded from 64Kbps - 256Kbps the original track is at the top and the re-encoded one is beneath it - if the quality had been increased then the peaks should be larger but as you can clearly see theres no difference at all - the only thing thats changed is the file size - now 4 times larger than it was originally - no increase in quality at all
How on earth can you even consider using MP3 as an example here? This is a perceptual process, and has already thrown away all information the codec thinks you cannot hear. You cannot use MP3 to "prove" anything at all about high resolution audio! I will post a spectrum plot if you like of a file that is 24/44.1 in one incarnation, and 24/96 after proper resampling & processing. You will be able to see - clearly - the increased resolution in the top end, as after processing with a special EQ that is sensitive up to 192KHz (it uses this resolution internally), the slope extends all the way to 34KHz - if there was no difference, it would still stop dead at 22.
I wouldn't try to prove that the tools and methods Wilkes has been using professionally for 20(?) years now actually doesn't work... Probably it's easier to understand a somewhat similar example with a TV picture. Take a SD NTSC/PAL picture and double both the horisontal and vertical resolution. It doesn't look a bit different, just as in the "MP3 Example". Still the same jaggies. But take a good picture/video processing program and do some anti-aliasing on the higher resolution picture...
