|
|
Post by schmalzy on Jun 20, 2022 11:51:53 GMT -6
I agree with Bob directly above. 48kHz is notably better than 44.1 to me.
I've toyed with going back to 48kHz but my bread-and-butter EQ/saturation/compression stuff just seems to sound a little better at 88.2khz. So I think that's where I'm staying. If I KNOW I'm going to be manipulating a recorded part a lot - time stretching, pitch shifting, other methods of audio mangling - then I definitely prefer 88.2kHz. It seems like the Reaper (my DAW) pitch/stretch algorithms sound significantly better with higher sample rate captures.
|
|
|
|
Post by Blackdawg on Jun 20, 2022 15:04:00 GMT -6
I agree with Bob directly above. 48kHz is notably better than 44.1 to me. I've toyed with going back to 48kHz but my bread-and-butter EQ/saturation/compression stuff just seems to sound a little better at 88.2khz. So I think that's where I'm staying. If I KNOW I'm going to be manipulating a recorded part a lot - time stretching, pitch shifting, other methods of audio mangling - then I definitely prefer 88.2kHz. It seems like the Reaper (my DAW) pitch/stretch algorithms sound significantly better with higher sample rate captures. The more samples you have the more data you have to manipulate. Very helpful for time stretching and noise reduction. |
|
|
|
Post by notneeson on Jun 20, 2022 15:37:34 GMT -6
Lavry says the optimal theoretical sampling rate would be 60k, so 88.2 and 48k are closest available FWIW.
|
|
|
|
Post by plinker on Jun 20, 2022 19:50:14 GMT -6
I don't know if this is still relevant ( svart), but I remember reading a decade, or so, ago that chips are either optimized for single or double rates, then scaled either direction depending on the selected rate. Anyone know if this ever was, or is still, a thing? |
|
|
|
Post by svart on Jun 20, 2022 21:29:48 GMT -6
I don't know if this is still relevant ( svart ), but I remember reading a decade, or so, ago that chips are either optimized for single or double rates, then scaled either direction depending on the selected rate. Anyone know if this ever was, or is still, a thing? Modern Delta-sigma converters are all oversampling. They oversample 128x to 512x times the audio rate. A master clock of 22.579Mhz would divide by 512x for 44.1k, 256x for 88.2k, 128x for 176.4k. a master of clock of 24.576Mhz would divide by 512x for 48k, 256x for 96k, 128x for 192k. Oversampling significantly lowers the antialiasing filter requirements and significantly lowers the quantization noise. There's almost no drawbacks this way. I should also add that it's much easier for manufacturers to build crystal oscillators in this frequency region as well. The size of the quartz crystal in the oscillator is relative to it's frequency, so the higher the frequency, the smaller the crystal can be. You can't make a crystal large enough for direct audio frequencies, and you can't make them smaller than what would equate to around 100Mhz. |
|
|
|
Post by therealkey on Jun 20, 2022 22:37:30 GMT -6
I record and mix at 96khz. My story on that is basically, I was recording one day and came across the great sample rate debate. Given that it’s fairly easy to change sample rates on my Apollo x6 in console I chose to experiment for myself (as we all should strive to do) after reading up on it in forums and the like. I ended up coming to the conclusion that I enjoyed the way tracking at 96khz sounded better. Consequently, since my files were recorded at 96khz I opted to not have to convert everything, so the mixing followed suit at 96khz as well.
|
|
|
|
Post by mcirish on Jun 21, 2022 8:27:35 GMT -6
I jump between 44.1 and 48K for all my projects. I mix ITB. There is zero possibility of me NOT having to bounce and/or freeze tracks if I were working at 96K. My personal belief is that with the quality of converters at this time, no end user will ever be able to tell if you recorded at 48K or 96K. It doesn't matter. I need to work quickly and I hate having to freeze parts of a mix so I can finish. I have a pretty powerful computer and I think my mixes come out well. If there are faults, it is 100% not the converters. I feel like a Neanderthal for saying it, but I don't want to waste the limited time I have on extremely small differences that no one in a blind A/B test can tell apart.
If I mixes with hardware, I might feel differently, as I would be going though converters multiple times. But, for ITB mixing, 48K is fine.
|
|
|
|
Post by OtisGreying on Oct 29, 2022 7:30:56 GMT -6
I agree with Bob directly above. 48kHz is notably better than 44.1 to me. I've toyed with going back to 48kHz but my bread-and-butter EQ/saturation/compression stuff just seems to sound a little better at 88.2khz. So I think that's where I'm staying. If I KNOW I'm going to be manipulating a recorded part a lot - time stretching, pitch shifting, other methods of audio mangling - then I definitely prefer 88.2kHz. It seems like the Reaper (my DAW) pitch/stretch algorithms sound significantly better with higher sample rate captures. The more samples you have the more data you have to manipulate. Very helpful for time stretching and noise reduction. This must also translate into more natural sounding pitch corrected vocals right? Autotune, Melodyne etc. Perhaps not night and day difference but helping in that regard? I can definitely hear my tracks take much better to transposing and time stretching in comparison to what I was used to hearing with 44.1
In any case 96/24 is where I've decided to work from now on
|
|
|
|
Post by teejay on Oct 29, 2022 10:07:37 GMT -6
I've gone round and round on this. I seem to have a converter that does make a difference in what I'm hearing when I track. Have been on 96/24 for a while, but recently switched to 88.2/24 as it seems to smooth my vocals a bit during tracking. That's what I hear, but keep in mind I'm also the guy who thinks he can hear a slight difference in the position of the Hakan pop filter.  |
|
|
|
Post by cyborgssc on Oct 29, 2022 11:25:56 GMT -6
My Rednet converters sound like garbage at 44.1. Probably something to do with how the LPF interacts with audible high frequencies near Nyquist. 48 is OK, but I've settled on 96/24 as the sweet spot. All of my mixing is OTB so the computer is just a glorified tape machine - no worries about CPU.
EDIT: Evidently that's my first post. Hi everyone!
|
|
|
|
Post by Bob Olhsson on Nov 1, 2022 12:07:42 GMT -6
Lavry says the optimal theoretical sampling rate would be 60k, so 88.2 and 48k are closest available FWIW. Jim Johnston, who was at Bell Labs, says the same. Since nothing is captured at low rates with modern converters, to me it's really a question of storage and minimizing unnecessary resampling cycles when applying DSP. |
|
|
|
Post by theshea on Oct 22, 2023 1:58:25 GMT -6
since i have a M1 apple now i can go up in SR. the question is: 88 vs 96khz.
spotify accepts 44khz and thats the main market today. so would it be best to work at 88khz - the double of 44khz?
should not be a problem with 48khz video, as converting down for video never makes a difference to me as video gets consumed most of the time on youtube with already crappy yt conversion and listened through inears or crappy headphones.
i cant HEAR a difference between 88khz and 96khz …
|
|
|
|
Post by nomatic on Oct 22, 2023 3:23:34 GMT -6
48K sounds great with my JCf Ad-8s and much of what I do is interlocked with Video.
|
|
|
|
Post by Blackdawg on Oct 22, 2023 23:59:40 GMT -6
since i have a M1 apple now i can go up in SR. the question is: 88 vs 96khz. spotify accepts 44khz and thats the main market today. so would it be best to work at 88khz - the double of 44khz? should not be a problem with 48khz video, as converting down for video never makes a difference to me as video gets consumed most of the time on youtube with already crappy yt conversion and listened through inears or crappy headphones. i cant HEAR a difference between 88khz and 96khz … Don't over think it. Computers are great at the conversion process now. I do 96k Minimum on all projects. |
|
|
|
Post by thehightenor on Oct 23, 2023 0:33:15 GMT -6
My new computer is so powerful that I’ve decided to switch to 48/24. Bob Katz insists the HEDD 192 sounds better at 48KHz - I personally can’t really hear any significant difference but I trust Bob Katz  |
|
|
|
Post by mcirish on Oct 23, 2023 5:44:01 GMT -6
Since my last post on this, every project has been at 24/48. Possibly 96k would sound different but I'm not entirely sure it sounds better. It's kind of a numbers game that always keeps you upgrading. Possibly someone 16 could hear some difference that would make 96k preferred, but my ears can't make that call, so I work at 24/48. Sounds fine to me. I wouldn't get all hung up on it. Honestly, if there were a big, easy to hear difference, there wouldn't be this discussion.
BTW, I'm on an Aurora (n) converter. Excellent
|
|
|
|
Post by chessparov on Oct 23, 2023 10:05:04 GMT -6
I can hear it! Since I'm 14. (With 50 years practice) Chris |
|
|
|
Post by lee on Oct 23, 2023 23:57:14 GMT -6
48/24 most of the time, but I hear plug-ins sounding a little better at 96 or 88.2. I don’t really mind either. About 6 other factors are more important in my book.
|
|
|
|
Post by christophert on Oct 24, 2023 0:10:10 GMT -6
96k predominantly - unless it is for film > then 48k / or if it is a punk rock band > 48k
|
|
|
|
Post by seawell on Oct 24, 2023 0:11:58 GMT -6
96kHz here. I feel like I hear a bit more clarity on cymbals in particular but it’s mostly about keeping tracking latency so low. I’ve worked on projects at every sample rate that I was very happy with the end result so it’s definitely not something I would stress over 👍🏻
|
|
|
|
Post by niklas1073 on Oct 24, 2023 0:14:54 GMT -6
48/24. Cost effective regarding i/o. Cannot really hear any difference in end product on 48 vs 96.
|
|
|
|
Post by thehightenor on Oct 24, 2023 0:36:44 GMT -6
As I said I’m now on 48/24. But when I think about those Sting and Peter Gabriel albums I love the sound of done on Sony Dash recorders at 44.1Khz … 16bit! Kinda proves the sample/ bit rate does not a hit make. Nice to have sonic options, but certainly not an essential ingredient. It appears talent shows through even at 44/16 |
|
|
|
Post by chessparov on Oct 24, 2023 2:10:50 GMT -6
James Taylor's "Hourglass" won two Grammys in 1997. Best Pop and Best Engineered. And was also recorded 16/44.1 Hey less is more. Hail the new Vintage!   Chris |
|
|
|
Post by Hudsonic on Oct 25, 2023 6:08:04 GMT -6
192kHz because phase response is optimized at this speed. With my SONOSAX recorder can achieve a 90kHz. response on the top end.
Of course I'm in the dusty old classical music recording business using hi-res mics.
|
|
|
|
Post by Ward on Oct 25, 2023 6:35:13 GMT -6
192kHz because phase response is optimized at this speed. With my SONOSAX recorder can achieve a 90kHz. response on the top end. Of course I'm in the dusty old classical music recording business using hi-res mics.Which is why we admire and respect you. |
|
