Posted Mon May 19, 2008 5:41PM
I totally agree with procreatemultimedia.
until we mix and master the final audio, the minimum should be 24/48 at least. I have mixed audio that was at both a lower (16/44.1) and the higher 24/96 and the difference it causes is too big to ignore. main things affected were hi-hats and violin samples I recorded at both rates. the 16/44.1 just sounded muddy compared with the higher quality audio. with bass it was like having a tissue on the plectrum. with vocals it had the classic "behind the curtain" sound.
so my vote is to have 24/48 as a minimum, with 24/96 preferred. people with lesser needs can always re-sample down the tracks! :P
(Edited on 2008-05-19 17:43:49 by somhairle)
(Edited on 2008-05-19 17:44:23 by somhairle)
Posted Mon May 19, 2008 6:06PM
90% of the market out there could never tell what was 16bit and what was 24bit. Also the playback decks and or speaker setups in the average joe´s studio are of such low response that it really doesn´t matter.
The only way for them to notice would be to give a free pair of audiophile quality headphones to them in order to notice the difference and probably would have a hard time still.
Recording at 24bit and downsampling does increase the quality considerably, so want to add a bit of clarity then record in a higher format and downgrade. Just like Shoot RAW and sell as JPG.
Just for the sake of argument, punch in the signal of your cable TV or HDTV via satellite or cable and the sound is pretty awful. Muddy doesn´t even begin to sdescribe the lameness. Compression will destroy it. Now internet sound? don´t get me started.
Would it be cool if we uploaded in 24bit and iStockAudio did the downsampling in the servers to offer clients different download formats and lets say $1 for 16bit and $3 for 24bit? Sure, but lets get the engine going smoothly before this happens.
Im recording at 24bit and 96khz for all my masters; I will be ready when the site offers the possibility.
Posted Tue May 20, 2008 6:17PM
I agree with most of what jamesbenet says, though to take the RAW/JPG analogy a bit further, lets bear in mind that once a sample is included with a lot of other source material, it becomes like a single photograph as part of a montage. Depending on what the montage is like, a good quality jpg (particularly if it has been carefully handled in transfer from RAW) will be indistinguishable in the final picture from the RAW source. In other words the extra detail in the RAW is insignificant.
Likewise, once an audio sample is mixed with others, in many instances it will be irrelevant whether the original was 16/44.1 or 24/96 - especially if it's at low level in the mix. (All those less significant bits in the 24 bit sample get dropped off when the maths is done to reduce level).
The most significant factor is how well the sample was recorded to begin with. A well recorded 16/44.1 will sound better than a poorly recorded 24/96 - the higher "quality" sample will just be a better rendition of the noise and distortion! The real benefits of high sampling rates and high bit depths are realised when signals are being processed - i.e. when level is changed, e.q. is applied, or signals are mixed. Having the extra bits and samples helps maintain fidelity during these processes. The bitrate and depth of the original and final output are much less significant.
I'm really surprised procreatemultimedia,that you place any significance on the specs of $59 DVD players and computer soundcards - after all, the sound quality coming from them is complete garbage! Use a cheap computer card for recording at 24 bit sample depth, and you will be no better off than recording at 16 bit, because the bottom 8 bits (and maybe more) are lost below the noise floor! I reckon that many of these devices boast such "impressive" specs, because the average Joe thinks that more bits is always better!
Higher spec. equipment does not always equate to better results - loads of bits and high sample rate is no more a guarantee of audio quality, than letting a monkey loose with a Canon EOS 1DS is a guarantee of good pictures. Let a professional photographer loose with a cheap camera and he will produce good results - and in many situations just as good as he would have achieved with expensive kit. Likewise, a good sound engineer who knows how to get the best from his gear, will produce better quality 16/44.1 samples than some amateur will get with 24/96.
Here's a test to try. Record a sample at 16/44.1 - make sure the levels are correct and get the maximum signal on there without clipping. Now record the same thing at a higher rate - say 24/48. Upscale the 16/44.1 to 24/48 and now do some playing around - say a bit of eq and compression. Mix your sample with just one other instrument or sound. The important thing is to do the exact same thing to both. Now play both to your friends, and see how many can tell the difference. Now the real test - get a friend to play them to you and see if you can tell the difference!
When your done, you really are in a position to understand just how important the bit depth and sample rate of those original samples really is.
(Edited on 2008-05-20 18:19:16 by brianaadams)