A thread about higher sample rates.

  • Thread starter Thread starter MessianicDreams
  • Start date Start date
MessianicDreams

MessianicDreams

New member
So here's the deal:

I track and mix at 48khz. I do this because it offers a slightly better detail to the sound, as well as allowing for slightly lesss brutal digital filtering, as you can now have frequencies up to 24khz instead of 22khz with 44.1khz SR.

But are there added benefits to recording at higher sample rates? I've heard that certain people can hear a difference up to about 96khz, but above that not really. I can hear people asking "but why, MD? why would you ever want to record higher than 48khz?". Well, recording up to 96khz allows for more higher frequencies (by that i mean >24khz, as 96khz allows recording freqencies of up to 48khz), and so less brutual digital filtering needs to be applied.


Why record frequencies that you cannot hear? Well, what about phase and frequency relationships? Surely all the HF information is going to have some sort of effect on frequencies in our hearing range, right? so by recording at 192khz, you'd be included a lot more HF information, which although they can't be heard or reproduced will have some sort of effect on what we are hearing.


I heard someone once say that a higher sample rate offers a better representation of the analogue signal, and whilst this is true, it does not offer a better representation of analogue recording mediums, which are actually pretty poor at recording HFs (the domains, due to their small size, can sometimes magnetically "cancel" each other out, which results in a lower HF content and adds to why tape feels "warm").


That's my thesis so far. Can anyone comment on this? is there some major factor which i have not thought about?
 
bennychico11 said:


thanks for your "help" (if you can call it that). If you'd taken the time to read the contents of my post, perhaps you would've seen that this post wasn't a "what sample rate should i record at", such as most of the other threads (indeed, all the threads you quoted) on this board. I want to know whether my theory on frequency and phase relationships stands true or whether it's a load of BS.


i know how to use the search function. i used it before posting this thread. perhaps i missed the thread i should've found. i don't think i did. sorry if this thread offends you :rolleyes:
 
No matter how you parse the question, sample rate is such a minor consideration in the whole process as to hardly rate more than a few lines of text...

But FWIW, high sample rates are not meant to provide a closer representation of analog recording mediums, they are meant to provide a more accurate representation of the analog source. There's a huge difference in definition there.

Analog "warmth" has several causes including, but not limited to, magnetic "smear", tape saturation characteristics, lack of digitization, etc.

The fact is good machines with good quality tape can reproduce high frequencies that extend beyond the human hearing range just as well as digital can.

All in all, I'd personally recommend not even worrying about the whole sample rate thing very much. It's about as important in the big picture as the difference between recording a guitar amp indoors with a wind screen or without a wind screen.

G.
 
SouthSIDE Glen said:
But FWIW, high sample rates are not meant to provide a closer representation of analog recording mediums, they are meant to provide a more accurate representation of the analog source. There's a huge difference in definition there.


that's why i felt it important to point this out - i was quite astounded when a friend of mine came out with this, as it seemed complete BS. and of course there are many other reasons for the "warmth" of tape, smearing being just one.



Also this thread was meant to be more of a theoretical discussion rather than a "i have a question, please answer it" type deal.

thanks for yer input southside :)
 
I think any problem with recording at higher sample rates lies with the audio interface - the converters is possesses. Sure most interfaces out there advertise 24/96 or 24/192, but how many of them actually accomplish this in a clean way in the prosumer market? Personally I still record at 24/44.1 and am happy with the quality of my recordings. If I were to jump up, I'd probably jump up to 88.2, but I tried once and wasn't "blown" away over 44.1.
 
MessianicDreams said:
i know how to use the search function. i used it before posting this thread. perhaps i missed the thread i should've found. i don't think i did. sorry if this thread offends you :rolleyes:

don't be quick to misconstrue my post as my being "offended"
and I did read what you said. but most of what you asked could be found in pieces around this board....I just gave you a place to start.
I didn't say nobody else could post and give you more direction on it, as they have already.

The Lavry article has been linked to before in the links I gave you and this post is pretty interesting as well.
 
There has been some test where brain activity was noted in the prescence of sound above 22khz, even though the subjects did'nt indicate they could hear it! That said, years of science and research basically have proven that a higher sample rate is'nt going to bring much more fidelity. ( My friend "SPOT" claims to be able to hear a difference!) :D

The transition from the analoug medium, with built in restrictions on high frequency content to a digital medium where there is less of a built in limit( thus: "warmth" ) combined with the volume wars " mastering Mafia" setting there limiters on stun and providing "digititas" square wave distortion is the reason that the boys in marketing are going to set you up with the Higher sample rates.

Up sampling can accomplish allot in signal processing, and compression and limiting probably benefit the most ( because of digitals "look-ahead" capabilities) .

http://mixonline.com/mag/audio_world_above/

How much compression and processing do you need in your choosen genre?? if you recording most ordinary stuff , I'd spend my money on converters and monitoring, not the mucho extra hardware and processing power needed to handle 64 bit 96 khz or :eek: higher sample rates. Work on getting the most out of the more that reasonable technology we have now and quit worrying about scoring the "wonder tool" "magic bullet" sample rate thingy. :(

Just my(correct) opinion ;)
 
Only time you should record in 48KHz, in my opinion, is if you're recording for video, or you're using a crappy soundblaster card that is locked in at 48KHz sample rate.

If you're going to record any higher than 44.1, I would think 88.2 would be the next logical sample rate to use because you'd think it would convert to 44.1 more cleanly. Although I could be wrong.
 
danny.guitar said:
If you're going to record any higher than 44.1, I would think 88.2 would be the next logical sample rate to use because you'd think it would convert to 44.1 more cleanly. Although I could be wrong.


I think that at one point this was the case, however with the improvement of technology down sampling is done in a much cleaner way. I'm pretty sure I read something about this is Lavry's Sampling Theory paper..


some intersting points raised here guys! thanks for the input and interesting dialogue.
 
I don't think I can hear anything above about 15 khz anyway...
 
Back
Top