I know that 24 bit makes a big difference compared to 16 bit, but someone just gave me their opinion on sample rates. They said that there isn't much of a difference between 44.1 and 96, and since 96 takes up more than twice the space on my CPU that I should just record at 44.1. Also, there has been some talk about loosing quality when converting from 96 or 88.2 down to 44.1. I am trying to record album quality music, do you think the sample rate will make much difference? Should I record at 44.1, 48, 88.2, or 96?
Topic Tags
Comments
Sample rates do make a difference to those of us who can hear. A
Sample rates do make a difference to those of us who can hear. Anyone who doesn't hear the difference should be doing something else besides audio IMO.
The real questions are, at what cost and how much difference in the end product does it make?
My personal observations:
I auditioned an Apogee PSX800 converter at my studio using 2" 24 track as the source. I had the ability to toggle between analog and digital at 16 bit, 24 bit ... 44.1 48. 88.2 and 96k all at the same levels. Several other musicians sat in on the listening session and we all concurred at the conclusion that the analog playback sounded the best with natural high end extension that sounded completely open. 88.2 and 96k ran a very close second with 48 and 44.1 to follow. Of course 16 bit was the worse. Each successive lowering of the sample rates made the sound appear to close in, as if a ceiling were being lowered.
That said, all of these formats sound fine for professional audio production ... my favorite recording I ever did was on Alesis Blackface 16 bit ADATs .... Performance really does matter more than anything else.
As far as how much difference it makes at the end of the day ... my feeling is the better the source tracks are, the better the end product. Add to this that at some point, 44.1 may cease to be the delivery format of choice. At that point you might be glad you recorded the multis at a higher rate.
I record at 24/48, and subscribe to the argument that it's bette
I record at 24/48, and subscribe to the argument that it's better to capture the live sound at a higher sample rate even though you may convert down to 44.1 at some point. I've read a few well-written and researched articles suggesting that a track recorded at 48k, then converted to 44.1k has more definition and dynamics than a track recorded directly at 44.1k. As for losing quality when you down-sample, use a high quality converter and you should be ok. I use R8Brain by Voxengo. It happens to be free and it crushes the conversion built into Sonar.
That said, I don't use 88.2 or 96, or 192 for that matter, in the interest of DAW performance and disk space.
THAT said, Mr. French and others whose opinion I respect say stick to 44.1k.
Assuming you've got a dedicated audio drive, you should have plenty of room to record your project at 48k or more. Like almost everything around here, it's up to you. Why don't you try recording some tracks at various sample rates and see if you can hear a difference?
I have recently started recording at 24bit, 96khz. Now that I f
I have recently started recording at 24bit, 96khz. Now that I finally have some really good speakers (Mackie HR824's w/HRS120 sub) I can really hear the difference. The biggest benefit for me as a sound designer is that I can manipulate sounds (Pitch 'n' Time, etc.) with fewer artifacts. Vocalign performance has been especially improved. My mixes seem to have more definition and improved dynamics, the subtleties are more noticable.
Just about the only drawback is the need for increased storage. My clients must now ante up for two 200gig drives instead of two 80gig drives.
As Kurt stated, you have to balance cost vs. delivery format. If you are delivering to MP3 you may as well stick with 16bit 44.1khz.
A few random thoughts from an amateur. It partly depends on the
A few random thoughts from an amateur.
It partly depends on the sound card. It seems to me that not all sound cards that has 96k on the front panel, sounds especially good at 96k. So try your own sound card.
It partly depends on the plugins you use for processing. Some plugins are known to sound better at higher sample rates. Some plugins has been known to sound really bad at higher frequencys.
It partly depends on your computer. Maybe, 24 channels at 96k is going to be too much for your computer. So if you need 24 channels you have to opt for a lower frequency.
It partly depends on the rest of the signal chain. If you start with a typical local talent in a room that is not treated acoustically and record it with a $100 Chinese mic, well, it probaly will not make much difference.
A lot of great records has been done at 44.1 or 48k. A lot of recorders still use that. You will not be damned forever if you stay at 44.1 at the moment.
Finally, it is very simple to test for yourself. If you cannot hear much of a difference you are in the same situation as most other recorders.
Personally, I spend my time and money on getting better mics and placing them in the right position. Putting the right mics up at the right place makes and enormous difference, recording at 44.1 or 96 makes a tiny difference. Í would like to be smart and focus on the large differences first.
Gunnar.
I was slightly blown away when I saw the Director Of Sales at Ap
I was slightly blown away when I saw the Director Of Sales at Apogee said (Quoting from Pro-sound Web forums):
For my .02, I would like to say that from our standpoint, there is no sonic benefit to 192K sampling rates in and of themselves. There are converters out there that perform better at 96K and 192K, but the fact is that changes to other aspects of the converter design, namely filtering, are really what is having an impact on the sound.
The benefits higher sample rates deliver are ancillary at best and generally unnecessary, provided the converter is designed to properly perform at lower sample rates. In these instances where the converter is designed properly, there are no sonic benefits to using higher sampling rates.
http://recforums.prosoundweb.com/index.php/mv/msg/288/0/208/0 For reference (Page 14, about 1/2 way down)...
I believe that Nyquest was right (as his proven Theorem shows). This means the limitations of current 44.1 rates are in the design of the circuitry and filters, NOT a limitation of 44.1K (22.05KHz) bandwidth IMO...
If you can clearly hear a difference at high FS, it is likely the Filters acting differently in the AUDIBLE band (sub 20KHz), NOT more "sub 22.05KHz audio information" creating more resolution (unless you can hear past 22.05KHz - I surely can't).
If the guy who is in a position to MAKE MONEY selling high FS converters (Apogee) denies there is any benefit to high FS rates, then I am inclined to believe him for some reason...
Now, Up-Sampling AFTER A/D conversion for the purpose of Plug in processing CAN be beneficial in some software, but this is based on the Algorithm, and how it functions on high frequencies nearing Nyquest.
I can, however, see 96K becoming a new standard just because the Pro-Audio industry can. And so many people are set on hearing an improvement reguardless of if the problem is in the FILTER at lower FS (a fixable problem at 44.1 IMO), or if there actually is an improvement in high sample rates... I can't hear a difference, but I'm not making big bucks doing my thng. I'd like to fix the root of the problem, not apply band-aids to poor converter design by doubling or quadrupling the sample rates.
Just one guy's take...
:cool:
No
No