Page 1 of 2
#1
Hi,

Nowadays we can get audio interfaces that go way above 48khz sample rate. My one is an old Apogee Jam that goes up to 48khz, so I was wondering if an upgrade would be worth?

--[ KenA ]--
--[ ]--
Last edited by KenACwb at Jan 1, 2016,
#2
Hi,

I don't think so. Most people record in 41kHz anyway.
But I might be wrong.
#3
Actually I've seen some AI going up to 192khz ... I do note differences between 44100khz and 48000khz and sometimes I'm playing/recording at 48khz and it sudenly drops to 44100khz ... not sure it's an audio interface issue or my DAW (Ableton Live).
#4
For a direct guitar?
44.1k are more than enough, no need to spend money in something else.
Name's Luca.

Quote by OliOsbourne
I don't know anything about this topic, but I just clicked on this thread because of your username :O
Quote by Cajundaddy
Clue: amplifiers amplify so don't turn it on if you need quiet.
Quote by chrismendiola
I guess spambots are now capable of reading minds.
#5
So let's talk about sampling rates:

44.1kHz is enough. 48kHz is not really any different but it's the video standard because of time-stamp reasons.

Anything higher than 48kHz is pointless unless you're going to pitch/time correct the shit out of it.
#6
If you are running a home project studio, 44.1khz/24bit is fine.

Pro mastering lab? 88.2/24bit
"Your sound is in your hands as much as anything. It's the way you pick, and the way you hold the guitar, more than it is the amp or the guitar you use." -- Stevie Ray Vaughan

"Anybody can play. The note is only 20 percent. The attitude of the motherfucker who plays it is 80 percent." -- Miles Davis

Guthrie on tone: https://www.youtube.com/watch?v=zmohdG9lLqY
#7
I prefer 44.1. I don't really understand why people record in higher resolution than that really if you will have to downsample it and introduce dithering. I guess there must be a good reason, but for my purposes, I can't see the purpose of recording at a higher sample rate.
Last edited by fingrpikingood at Dec 28, 2015,
#8
Eventually most projects end up in a media that compresses and greatly lowers the quality whether it's a CD, MP3, MP4, whatever. Unless you can distribute your recordings in a media like FLAC it's all just academic and a waste of resources to sample at higher rates besides "almost" no one will hear any difference. I had to say "almost" because there will always be someone who will swear they can tell the difference between 44.1 and 48 or 96. I don't have supersonic hearing so 44.1 and 48 are fine with me. It's better than my old standard 15 I.P.S. @ 59db s/n ratio on a 10 1/2 inch reel.
Attachments:
sony_tc-854-4.jpg
Yes I am guitarded also, nice to meet you.
Last edited by Rickholly74 at Dec 28, 2015,
#9
As new technologies such as High Res Audio come to the fore, people will begin to substantially notice the difference between lossy compression schemes such MP3 and MP4, as well as the harshness of the CD 16 bit 44.1kHz format. For professionals, I would always suggest recording in the highest bit depth possible (except for DSD 1 bit and 2.8224 mHz sample rate, simply the best in digital), so 24 bit 192kHz is definitely the way to go, and your master should reflect any format as needed from radio to web distribution. You can even use Endless Analog's CLASP which allows you to get all of that beautiful harmonic distortion and tape compression with digital's superior dynamic range. For hobbyists, you can do whatever floats your boat.
#10
Most of my studio work was at 24bit/48khz, every once in a while a project came in or was requested to be at higher res, usually 96khz or 192khz. Honestly, once you get into a rock mix you would barely hear any difference or any at all. For symphonic music you get a bit more detail. The deeper resolution places a further burden on your system and file storage that most of the time doesn't really warrant it.
I work at 24/44 in my home studio and am quite happy with it for the moment and it sounds good. In the end it is all mastered into 16/44 so I don't find the need to up the resolution.
#11
I wonder... Do these higher resolutions reduce listener's fatigue? I know most people can't hear the difference above 44.1khz, but for long term listening (say, an entire album) a higher resolution might reduce it. I know that lower resolutions get pretty hard to listen to, so maybe the higher it goes, the more "pleasant" the sound would be even if it's not consciously perceivable.

Personally, if I think what I'm recording is important, I might as well get the best possible recording I can for archival purposes. I currently top out at 24/96 so that's what I go with.

If I had the choice of 32-bit float over 24-bit, I would definitely go with 32 no matter what though. And yes, I know it's not actually 32-bit, but there is an improvement in dynamic range, which is arguably more important than any other aspect of sound as far as reproducing reality goes. Eventually, I'm sure, there will be 32-bit D/A converters, thus the aforementioned desire for a high quality archive. Any possible improvement in playback quality aside, there are also benefits to using it during processing (namely, for all practical purposes, an unlimited dynamic range).

But then again, it depends on what you're going for. Everyone seems to want to compress the hell out of everything and pin it to just below clipping these days anyway.

As far as a need to upgrade, if you're happy with the sound quality, why bother?
"Practice doesn't make perfect. Perfect practice makes perfect." -some dude
#12
Why record music at 48kHz? Have fun with that when getting people to listen to the MP3 on an MP3 player, or burning it to a disk.
Quote by Watterboy
Do you have any dilithium crystals or fresh warm dumps for sale
#13
Scorpions recorded Animal Magnetism and Blackout on Sony digital tape in early 80s much more limited format than today's digital, essentially ADAT, and listen to how those alkbums sound. It is just about knowing what medium you work on and getting the best out of it.
#14
If I know I'm gonna be doing a lot of time compression/stretching I'll do 88.2 or 96. Beyond that is utterly pointless.

Quote by Prime2515102
I wonder... Do these higher resolutions reduce listener's fatigue? I know most people can't hear the difference above 44.1khz, but for long term listening (say, an entire album) a higher resolution might reduce it. I know that lower resolutions get pretty hard to listen to, so maybe the higher it goes, the more "pleasant" the sound would be even if it's not consciously perceivable.


This might be relevant for lossy filetypes (higher bit rate = less psychoacoustic fuckery going on) but certainly not for lossless types.
Current Gear:
LTD MH-400 with Gotoh GE1996T (EMG 85/60)
PRS SE Custom 24 (Suhr SSH+/SSV)
Ibanez RG3120 Prestige (Dimarzio Titans)
Squier Vintage Modified 70s Jazz V
Audient iD22 interface
Peavey Revalver 4, UAD Friedman BE100/DS40
Adam S3A monitors
Quote by Anonden
You CAN play anything with anything....but some guitars sound right for some things, and not for others. Single coils sound retarded for metal, though those who are apeshit about harpsichord probably beg to differ.
Last edited by oneblackened at Jan 1, 2016,
#15
Quote by Prime2515102
I wonder... Do these higher resolutions reduce listener's fatigue?
No, it doesn't have anything to do with that.

One may argue recording some stuff at a sampling rate higher than 44.1kHz could "reduce listening fatigue", but it's because of a different reason.
Quote by Prime2515102
If I had the choice of 32-bit float over 24-bit, I would definitely go with 32 no matter what though. And yes, I know it's not actually 32-bit, but there is an improvement in dynamic range
You can't build a floating point converter tho.
I mean, theoretically you can, but in good practice you can't really with today's tech, and even 32bit int converters would need a lot more precision than we have now.

Now there you make it sound like 32bit float doesn't have the same resolution as actual 32bit audio stream, so I'm guessing you mean you would choice 32bit float audio files over 24bit int files, all to be passed through a 24bit DAC, but there you lose all of the audio quality advantages.

The only situation in which a 32bit float stream would be more efficient than a 24bit int stream would be if you made music entirely ITB and then you exported the track into a 32bit float file.
Audio quality tho would arguably be reduced since that way you'd leave the quantization process to the audio player, which might not be as good quality as the DAW's.
Quote by Prime2515102
Any possible improvement in playback quality aside, there are also benefits to using it during processing (namely, for all practical purposes, an unlimited dynamic range).
This is the reason why most DAW's audio engines use 32bit audio, but not much of them offer the possibility of exporting 32bit audio.

I reckon it's because 32bit DACs simply aren't used for this application.
Quote by Prime2515102
As far as a need to upgrade, if you're happy with the sound quality, why bother?
Somebody may hear what you don't hear.
Quote by the chemist
Why record music at 48kHz? Have fun with that when getting people to listen to the MP3 on an MP3 player, or burning it to a disk.
It's because of aliasing.
http://recording.org/threads/oversampling-explained.48087/
Name's Luca.

Quote by OliOsbourne
I don't know anything about this topic, but I just clicked on this thread because of your username :O
Quote by Cajundaddy
Clue: amplifiers amplify so don't turn it on if you need quiet.
Quote by chrismendiola
I guess spambots are now capable of reading minds.
#16
I mean, I totally understand the theory behind it, but when everything is going to be turned into a shit MP3 anyways...
Quote by Watterboy
Do you have any dilithium crystals or fresh warm dumps for sale
#17
If you record something at 44.1kHz and you get aliasing distortion there's no turning back, you have it.
If you record the same stuff at, say, 88.2kHz and you don't get aliasing distortion, you convert the audio stream to 44.1kHz, then you don't get the aliasing distortion you would have gotten if you recorded the source at 44.1kHz.

If you think that there's no point in producing a better quality source material 'cause someone's not gonna listen to it in full quality you might as well record everything in 16bit to save space and bandwidth and not worry if some part of your chain clips the audio here and there.
Name's Luca.

Quote by OliOsbourne
I don't know anything about this topic, but I just clicked on this thread because of your username :O
Quote by Cajundaddy
Clue: amplifiers amplify so don't turn it on if you need quiet.
Quote by chrismendiola
I guess spambots are now capable of reading minds.
#19
It's distortion caused by trying to sample a signal that has a frequency higher than half of the sampling frequency.

It's still all written here - http://recording.org/threads/oversampling-explained.48087/
Name's Luca.

Quote by OliOsbourne
I don't know anything about this topic, but I just clicked on this thread because of your username :O
Quote by Cajundaddy
Clue: amplifiers amplify so don't turn it on if you need quiet.
Quote by chrismendiola
I guess spambots are now capable of reading minds.
#20
You must mean digital masking.
A lot has to do with A/D converters.
I remember why the studio I worked at dumped the Digidesign A/D, even at 192k it was nasty, Apogee beat it at all tests. Well designed A/D will take care of this even at 44k.
#21
Quote by oneblackened
If I know I'm gonna be doing a lot of time compression/stretching I'll do 88.2 or 96.


This is actually a good point. For fixing timing issues, a higher sampling rate could be pretty useful. I find 44.1 doesn't do time stretching so well. But if you record with good timing in the first place you should be good. Still though, if you improvise some sweet solo, and goof one note on the timing, it's useful to be able to fix it invisibly.
#22
Quote by Spambot_2
If you record something at 44.1kHz and you get aliasing distortion there's no turning back, you have it.
If you record the same stuff at, say, 88.2kHz and you don't get aliasing distortion, you convert the audio stream to 44.1kHz, then you don't get the aliasing distortion you would have gotten if you recorded the source at 44.1kHz.

If you think that there's no point in producing a better quality source material 'cause someone's not gonna listen to it in full quality you might as well record everything in 16bit to save space and bandwidth and not worry if some part of your chain clips the audio here and there.


That's an interesting point, but to be honest, I'm not sure if I've ever had issues with aliasing before, which doesn't mean I haven't, I mean, maybe I just didn't notice, which is very possible, but it seems to me like I couldn't avoid dithering, which is a guaranteed introduction if I'm going to downsample.

I'd like to hear examples of aliasing distortion though.
#23
I will never record above 44.1kHz becuase honestly, I never do Film work (48kHz), and ho little to no TCXE.

This debate is like all the engineers I knew years ago yelling and yammering for 32-bit resolution gear.

Great concept, does help dynamics.

But honestly, no one notices who doesn't work in the industry. Why do I say that? We can record at 192kHz, 32 bit res all day, but ultimately our mixes are going to get Red Book any way, so we loose 16 bits of resolution and get cut to 44.1kHz at press.

So that's why I record in 44.1; if it sounds good when recorded, then it will sound good on CD.

We can say that CDs are dead all we want, they're still the industry standard medium. Therefor, I record to that medium. I may go to 24 bit depth when recording, but that's mostly laziness (don't care enough to mess around with my settings)
Quote by Watterboy
Do you have any dilithium crystals or fresh warm dumps for sale
#24
One of the most overlooked aspects of digital recording is clocking. When you have a great word clock and all of your digital devices are slaved to it, you can eliminate jitter and really hear the difference between the higher sample rates. The whole purpose of digital recording was to surpass the dynamic range of analog tape. Antelope Audio has really good master clock solutions such as their flagship 10MX Rubidium Atomic Clock. This combined with good mic preamplifiers will also let you hear the difference. Also, one must consider the quality of the ad/da converters. Just because the converters say they go to high def sample rates and bit depth, the quality of the ad/da chipset does matter more than you think. The build quality of your converter's components are crucial to taking advantage of these higher bit depth and sample rates.
The quality of your cabling matters as well. Nothing saps tone quicker than cheap cables going into good converters. Anything worth doing is worth doing right, and an unimpeded signal path is worth its weight in gold tipped Mogami cable.
Digital audio is something that has to be dealt with from many angles, and simply plugging directly into a converter with a cheap 1/4"guitar cable is going to yield mixed results most of the time. Professionals understand that it is the combination of a great analog front end combined with rock solid digital stabilized by a really good master clock are key to making sound recordings that have the sound staging, dynamics, and clarity that makes the most out of high definition audio.
Anyone who says that they cannot tell the difference between 16 bit 44.1, 24 bit 44.1, 24/96, or 24/192 might want to examine the quality of the components, or what is lacking in their systems.
#25
According the the Nyquist sampling theorem to capture all of the data you trying to sample you need to sample at greater than 2 times the bandwidth to be able to fully reproduce the original signal. Since the audible range is roughly 20Hz to 20KHz the band with of the signal you are recording is 20KHz max. 2 * 20K = 40k. Providing the recording device is designed by someone who knows what they are doing they will place an anti-alias filter before the ADC to prevent signals higher than 20K from making their way into the recording.

Now say that you want to have a higher sampling rate after the recording is finished, if you had an anti-alias filter, you can use some fairly simple DSP to up sample the recording though it probably wont do much for you sonicly. In the process of typing this i forgot what my point was but thank goodness for my background in electrical engineering.

Dont get me started on possible analog to digital conversion types
Last edited by crupez5 at Jan 7, 2016,
#26
Quote by crupez5
Since the audible range is roughly 20Hz to 20KHz the band with of the signal you are recording is 20KHz max. 2 * 20K = 40k.
No, stuff produces harmonics at frequencies higher than that.
Quote by crupez5
Providing the recording device is designed by someone who knows what they are doing they will place an anti-alias filter before the ADC to prevent signals higher than 20K from making their way into the recording.
That generates a phase shift in the higher frequency region, which is something that people who know what they're doing really want to avoid.

There's no way to know if analog filters have been put there, and I'd argue it would make more sense not to put filters there in cheap equipment to save money and not to put it there in high end equipment so not to get phase distortion, since if you're using high end stuff you can probably afford enough disk space to record at a higher sampling frequency.
Quote by crupez5
Dont get me started on possible analog to digital conversion types
All audio converters being used for this kinda application are delta-sigma converters, so I don't really see why you would argue about that nor why it would make any difference if the conversion method was different.

Care to school me?
Name's Luca.

Quote by OliOsbourne
I don't know anything about this topic, but I just clicked on this thread because of your username :O
Quote by Cajundaddy
Clue: amplifiers amplify so don't turn it on if you need quiet.
Quote by chrismendiola
I guess spambots are now capable of reading minds.
#27
Still there's Scorpions' back catalogue with excellent sonics recorded on 16 bit 44k ADAT killing most of the new stuff that's recorded on higher sampling rates, which makes me think it is more of a foofoo argument.
The human hearing range peters out at 15k for older adults so this is one of the reasons most tape decks went up to 18k. Not sure the hyper frequencies above that would make much of a difference.
Last edited by diabolical at Jan 7, 2016,
#28
The harmonic overtone series is not always heard, but rather perceived in the human auditory system. What usually separates great sound recordings from the chaff is the accuracy in which these frequencies are reproduced up and down the series. This is where having the proper components in your audio chain come to bear, whether it is analog or digital. What happens below 20Hz and above 15kHz is usually felt not heard, and our perception of what happens there shapes how we feel about the sound that we actually hear. I will say this once again, if you cannot tell the difference between a truncated 16 bit 44.1 kHz CD, 24/96, and DSD, you should not be in the audio business in any kind of professional capacity.
#29
Lets realistically look at what instruments will tickle back your hearing from possible overtones being above 20k. In modern rock band scenario you'll most likely only have drum cymbals.
Guitar speakers usually start cutting out at 11k:
#30
Quote by diabolical
Lets realistically look at what instruments will tickle back your hearing from possible overtones being above 20k. In modern rock band scenario you'll most likely only have drum cymbals.
You may well have acoustic guitars, female vocals, keyboards and reverbs as well, but why only limiting ourselves to a modern rock scenario?
Name's Luca.

Quote by OliOsbourne
I don't know anything about this topic, but I just clicked on this thread because of your username :O
Quote by Cajundaddy
Clue: amplifiers amplify so don't turn it on if you need quiet.
Quote by chrismendiola
I guess spambots are now capable of reading minds.
#32
Quote by kendall jones
I will say this once again, if you cannot tell the difference between a truncated 16 bit 44.1 kHz CD, 24/96, and DSD, you should not be in the audio business in any kind of professional capacity.

Bullshit
RIP Gooze

cats
#33
My thoughts:
1)Moderns streaming applications like Spotify Premium can send high quality signals, so not EVERYTHING goes into a crappy MP3. And there are ways to save MP3's at higher bit-rates, anyways.

2) Crappy speakers may be the limiting factor in hearing the difference. If you have a pair of $8 skull candy headphones you won't hear the difference between 8K and 192K (because you won't hear anything but tinny noise :P). Get a nice pair of studio monitors that are designed to play every frequency correctly, then maybe you'll hear the difference if the rest of your equipment is good. Mine is cheap so I can't say.

3) I use 192K for a really strange reason. I amp my guitars without using a DAW-- I use VSThost, which does nothing but, well, host VST's. It's great to just double click on the icon and immediately have my amp up. But since I don't have an ASIO device I have to run it through MME. For some reason, MME can process WAY faster in higher sample rates (because it can... sample... faster..). It sounds stupid but it's true. It's a great way to set up an amp for screwing around because it's fast and simple (no clunky DAW controls) but I wouldn't recommend it for recording anything but ideas and rough tracks.
#34
Quote by josephbburg
My thoughts:
1)Moderns streaming applications like Spotify Premium can send high quality signals, so not EVERYTHING goes into a crappy MP3. And there are ways to save MP3's at higher bit-rates, anyways.

2) Crappy speakers may be the limiting factor in hearing the difference. If you have a pair of $8 skull candy headphones you won't hear the difference between 8K and 192K (because you won't hear anything but tinny noise :P). Get a nice pair of studio monitors that are designed to play every frequency correctly, then maybe you'll hear the difference if the rest of your equipment is good. Mine is cheap so I can't say.

3) I use 192K for a really strange reason. I amp my guitars without using a DAW-- I use VSThost, which does nothing but, well, host VST's. It's great to just double click on the icon and immediately have my amp up. But since I don't have an ASIO device I have to run it through MME. For some reason, MME can process WAY faster in higher sample rates (because it can... sample... faster..). It sounds stupid but it's true. It's a great way to set up an amp for screwing around because it's fast and simple (no clunky DAW controls) but I wouldn't recommend it for recording anything but ideas and rough tracks.


1) Streaming in general compresses the audio signal. Not in terms of dynamics, but bandwidth. Hell, even radio does, and for some reason we all still look at 'Will this sound good on radio' as a valid mix test. MP3/not-MP3 is moot for streaming. I don't know of ANY streaming service that streams at 192/32; hell, even some Dante systems (professional audio BTW) don't always make use of 192/32.

2) Monitoring is one small part. I don;t know who said it, but accurate clocking is just as important, as is accurate A/D D/A conversion, properly captured source material, processing, personal technique, etc.

3) I don't really have any serious experience with running with an MME engine (I'm using all ASIO, except the brief period I ran RADAR) and haven't found that ASIO can't sample that high. The studio I track in has a RedNet that samples 60 channels of I/O just fine at 192/32. A lot of this comes down to personal preference; some people use MixCubes for reference, some use MS10s, the worst speakers you will ever hear in terms of polish.

This debate is more about preference. Have I recorded above 44.1/16? Yes, for sure. I did, and still do, a large number of engineering sessions at 88.2/24 and 88.2/32-float.

From experience I've only found a massive, and I mean MASSIVE, difference between 88.2 and 192 when working with 'classical' instruments, such as strings, horns and woodwinds. I think it has to do with the very rich but subtle overtone structure that these instruments have compared to the brute force overtone structure of distorted guitars or loud rock drums.

A lot of this also will be impacted by your gear. If you have an analog console (for example, an AWS924+), you can somewhat get away with lower sample rates because of analog signal path. You can find the videos on youtube of James Lugo, who is a well respected engineer in Nashville, still using 44.1kHz sample rate, and his work sounds incredible. Keep in mind, he's using NanoSyncs (incredible clocks) and an SSL E/G+ series console, but that shows that a good enough engineer can overcome the known and inherent limitations of lower sample and bit rates.

Keeping this in mind, I won't deny that my recordings at 44.1/16 aren't as crisp or as accurate as something done at 192/32, but I can get around that with careful technique. I also can say that I've started experimenting with higher rates in the last few months (96+) and may switch if I find it worth the effort to upgrade my clock (kinda need to anyway)
Quote by Watterboy
Do you have any dilithium crystals or fresh warm dumps for sale
#35
Quote by KenACwb
Hi,

Nowadays we can get audio interfaces that go way above 48khz sample rate. My one is an old Apogee Jam that goes up to 48khz, so I was wondering if an upgrade would be worth?

--[ KenA ]--
--[ ]--


44.1khz - stick with that and save yourself the hassle - the other sample rates don't sound any better - it's a bunch of marketing hype. The only real advantage you get a higher rates is lower latency, but that gets offset from the problems you'll have from a performance standpoint.
#36
Quote by the chemist
Monitoring is one small part. I don;t know who said it, but accurate clocking is just as important, as is accurate A/D D/A conversion, properly captured source material, processing, personal technique, etc.
Whoever it was, he was wrong.
The clock you find in a $50 audio interface won't make your high end equipment produce a sound nearly as bad as if you used a pair of low quality monitor speakers.
Quote by the chemist
a RedNet that samples 60 channels of I/O just fine at 192/32.
How fast and how much data you can manage/record/reproduce depends on the computer you're using more than anything - if you used the same rednet system on a $200 laptop with a 5400rpm HDD you wouldn't be able to record many tracks at once.
Quote by the chemist
This debate is more about preference.
If it was a matter of preference alone we wouldn't really be debating.
Quote by the chemist
a large number of engineering sessions at 88.2/24 and 88.2/32-float.
Floating point audio converters are so uncommon that you could say they don't exist, and there's no DAW that even lets you record floating point audio.
So no, people don't record stuff at 32bit float resolution.
Quote by the chemist
If you have an analog console (for example, an AWS924+), you can somewhat get away with lower sample rates because of analog signal path.
Meaning what exactly?
Quote by the chemist
that shows that a good enough engineer can overcome the known and inherent limitations of lower sample and bit rates.
Nobody's arguing that you can't make stuff sound better when recording it at 16/44.1, we're arguing that stuff sounds better when you record it at higher sample rates and higher resolution.
Quote by reverb66
44.1khz - stick with that and save yourself the hassle - the other sample rates don't sound any better - it's a bunch of marketing hype.
You should read the previous posts in this thread.
Name's Luca.

Quote by OliOsbourne
I don't know anything about this topic, but I just clicked on this thread because of your username :O
Quote by Cajundaddy
Clue: amplifiers amplify so don't turn it on if you need quiet.
Quote by chrismendiola
I guess spambots are now capable of reading minds.
#37
Quote by Spambot_2
Whoever it was, he was wrong.
The clock you find in a $50 audio interface won't make your high end equipment produce a sound nearly as bad as if you used a pair of low quality monitor speakers.


I understand that.

Most people on this forum won't be going off and buying ADAM/Focal/Barefoot monitors, or buying a nanoSync either.

And I never stated that monitoring wasn't important; it's 1 part of the process.

Quote by Spambot_2
How fast and how much data you can manage/record/reproduce depends on the computer you're using more than anything - if you used the same rednet system on a $200 laptop with a 5400rpm HDD you wouldn't be able to record many tracks at once.


Can't really use RedNet effectively on a laptop either. While you can use the built in Ethernet port, you'll get the same performance as your standard USB/FW interface. They work best with Digi or Dante-tested cards.

But, very fair point to make. I misunderstood what he was saying. Rereading his point I realize he stated that MME seems better with higher sample rates, not an ASIO/MME difference. I'll concede that.

Quote by Spambot_2
If it was a matter of preference alone we wouldn't really be debating


People debate preference all the time; gear recommendation threads a good example. It's validating your preference that moves that discussion.

Quote by Spambot_2
Floating point audio converters are so uncommon that you could say they don't exist, and there's no DAW that even lets you record floating point audio.
So no, people don't record stuff at 32bit float resolution.


In terms of hardware, true. 24-bit is all we see.

Not in any current DAW versions, at least.

Pro-Tools 9 was completely 32-float
Pro-Tools 9 HD RTAS structure was 32-float
Logic 9 was 32-float
Reaper was 32-float not too long ago, but I don't really care for Reaper
Ableton 8 used 32-float

To be fair to your point, Cubase, Nuendo, Samplitude, and Sonor did use 32-float for effect processing and time/pitch correction.

But, with for a lot of applications, you wouldn't gain the benefits of 24 vs 32 bit depth anyways. I think the depth difference is something like 4 dBFS (from -140 to -144). So, a very negligible difference. Unnoticeable really, but some wad somewhere will say they can hear the difference.

Quote by Spambot_2
Meaning what exactly?


Some hardware can mask combing. I know that API and WA EQs do from experience when we dug up some old Pro-Tools 6 session audio to test with. I'd hazard to say that the SSL SuperAnalog sound does as well, but it depends on what the audio is. I've never liked putting cymbals or acoustics through SSL anyways (too brittle for my liking), so it very well may accent it on some sources.

Something I'll test out next time I get to an SSL.


Quote by Spambot_2
Nobody's arguing that you can't make stuff sound better when recording it at 16/44.1, we're arguing that stuff sounds better when you record it at higher sample rates and higher resolution.


Yes and no.

I could give some people the best gear in the world and still get a turd. Everything is tied together.

From my experiences, having better sample rates/resolution will never help or correct crap technique and inexperience.
Quote by Watterboy
Do you have any dilithium crystals or fresh warm dumps for sale
#38
Quote by the chemist
I understand that.

Most people on this forum won't be going off and buying ADAM/Focal/Barefoot monitors, or buying a nanoSync either.

And I never stated that monitoring wasn't important; it's 1 part of the process.
You tho wrote "Monitoring is one small part. I don;t know who said it, but accurate clocking is just as important", and now you tell me I'm right when I say that's wrong, but again you seem to reiterate what you said was wrong.

Moreover I don't really understand why having higher quality speakers has anything to do with what I wrote
Quote by the chemist
People debate preference all the time; gear recommendation threads a good example. It's validating your preference that moves that discussion.
What I meant is that we are debating that the advantages brought by higher resolution and sampling frequencies aren't a matter of preference, and if we agreed it was a matter of preference we wouldn't be debating anything technical.
Quote by the chemist
In terms of hardware, true. 24-bit is all we see.

Not in any current DAW versions, at least.

Pro-Tools 9 was completely 32-float
Pro-Tools 9 HD RTAS structure was 32-float
Logic 9 was 32-float
Reaper was 32-float not too long ago, but I don't really care for Reaper
Ableton 8 used 32-float
Couple things:
I reckon most of these numbers refer to the audio engine of the DAW, which means the audio you recorded is converted to 32bit float and then edited, which is done to avoid clipping when processing stuff.
Also are you sure the PT HD "structure" you're referring to is about audio and not about processing word length (32bit vs 64bit like we see in most modern processors)?

You may as well produce tracks using synth with 32bit float synth engines, but none of these DAW offer the possibility of exporting it as 32bit float audio, so the point is quite moot.

Quote by the chemist
But, with for a lot of applications, you wouldn't gain the benefits of 24 vs 32 bit depth anyways. I think the depth difference is something like 4 dBFS (from -140 to -144). So, a very negligible difference. Unnoticeable really, but some wad somewhere will say they can hear the difference.
Don't really know where you're getting that data from, but from the tech standpoint the headroom offered by 32bit int would be 192.66dB, while that one offered by 24bit int would be 144.49.

You are right saying more than 24bit would be pointless, tho that's because the best AD and DA converter circuits you see around have about 122dB and 130dB of headroom.
Quote by the chemist
Some hardware can mask combing.
Combing tho has nothing to do with sampling rate and resolution, so I don't see what you're on about here.
Quote by the chemist
I could give some people the best gear in the world and still get a turd. Everything is tied together.

From my experiences, having better sample rates/resolution will never help or correct crap technique and inexperience.
Yes, but we're not talking about performance and experience and stuff, we're talking about sound, we're arguing that if you record the same shit at a higher sampling rates and possibly higher resolution will still be shit, but it will sound better.

It won't be better as in it will make the performance better, it will be better as in it will record the same performance with better accuracy.
Name's Luca.

Quote by OliOsbourne
I don't know anything about this topic, but I just clicked on this thread because of your username :O
Quote by Cajundaddy
Clue: amplifiers amplify so don't turn it on if you need quiet.
Quote by chrismendiola
I guess spambots are now capable of reading minds.
#39
I'm just playing devil's advocate at this point really.

You are completely right: Higher sample rate = more accurate representation of source material.

In reference to the PTHD, that was in relation to the Plug-In processing. I did more reading becuase even as I typed it, something was off. PTHD PROCESSED in 32 or 24, depending on the plugin (EQs, gates and reverbs did 32, compressors, modulations did 24)

Quote by Spambot_2
You tho wrote "Monitoring is one small part. I don;t know who said it, but accurate clocking is just as important", and now you tell me I'm right when I say that's wrong, but again you seem to reiterate what you said was wrong.

Moreover I don't really understand why having higher quality speakers has anything to do with what I wrote


I don't even remember anymore

As to the dBFS, I did the math wrong. You are correct again.

I meant aliasing, not comb filtering. Sometimes certain analog analog signal chains can mask the harmonic structure with their own. It's hit and miss, but some gear does a good job of it (Looking at something like an EQP-WA or old Brit comps)

I think at the end of the day I'm not as versed in the computer side of recording, and should probably just stop posting before going to bed...
Quote by Watterboy
Do you have any dilithium crystals or fresh warm dumps for sale
#40
So I've been pondering this recently also so thought I'd share some thoughts and ask some questions in this thread.

I get quantizing and aliasing (to some extent) but putting all that aside for a minute I'm wondering if there is good reason to record over 44.1khz.

It seems the general consensus is no. I understand that there is a point of diminishing returns when it comes to audio resolution. But theoretically speaking...(famous last words)

We need at least two samples per cycle to get an accurate idea of frequency. Which means that for 44.1k we can record frequencies up to 22050Hz. Human hearing doesn't quite go that high right. However, that is only to capture frequency and it doesn't capture any of the shape of the wave form at that frequency. To capture the shape of the wave form and not just it's frequency how many samples per cycle would you need? Three four...ten?? Now I get that anything around 20khz won't matter too much because you probably won't be able to distinguish the detail in the sound that the shape offers, your ear won't detect it.

But what about four per cycle...that might not be enough...one at the start of the sinusoid, one at the peak one where it crosses 0, and one at the trough. Is that going to be enough? Lets say it is...then anything over 11khz is going to start to loose some minor detail.

But what if to really capture the character of a sound you actually need double that at least 8 samples per cycle to capture the true shape and character of a wave form. This would mean that anything over 5.5khz starts to lose detail.

But is that detail important? If such changes in shape are too minute for our ears to really hear does it even matter? Does our high hat start to lose a little character when recorded digitally at 44.1khz or do our ears just gloss over those minute details anyway in favour of the broader sound.

Similarly with bit depth and processing. Does the sample rate and having a more accurate capture of the source sound result in cleaner results when put through a plug in with less errors compounded through quantization errors etc

I read somewhere that audio resolution is not a concern for most plug ins with the exception of convolution in which the higher resolution recordings actually have better results.

That brings me to a slightly different topic that is oft debated but maybe I'll start a new thread for that one.

Anyway from what I've read in all the double blind testing the difference between 16 bit and 24 bit recording is barely distinguishable, and between 24 and 32 you're really pushing it to validate a claim that there is a noticeable difference. Similarly the 44.1 vs 96khz debate starts to get 96khz on the wrong side of the point of diminishing returns.

Then there's the cost. Recording at higher resolution and working with multi-tracks of higher resolution audio gets much more CPU heavy. Add in some convolution reverb, a few other plug ins and a couple software tracks with their own sample loads and is the resolution that is barely noticeable worth it??

I understand the want to work with the best most pristine recordings but considering what was done with so much worse...it doesn't feel to me like working in 196/32 is going to be the deciding factor to gain more listeners, more appreciation for your work, or better results. The deciding factor is surely the talent of the people involved at each stage of the process as opposed to the quality of the gear or the resolution (within reason of course...it doesn't matter how good the people are if the mic simply doesn't work, or the soundcard will only handle 8bit recording)

So after all that, nevermind I guess by posting this I found out where I stand on the question.

Peace, sorry if I bored you gonna hit post anyway cause I'm still open to criticism
Si
Page 1 of 2