| Author
|
SRC comparisons 96 to 44.1 kHz
|
shamantrixx
Started Topics :
7
Posts :
549
Posted : May 8, 2007 07:16
|
sorry but we had numerous debates about sampling rate and I don't want to argue again. Look out for old topics if you're interested or simply TRY exporting 4 or 8 bars of VST synth phrase on 44.1 and 96 kHz and hear the difference. Automate some cutoff and resonance in that 8 bars and let us know if you hear no difference.
  "It occurred to me by intuition, and music was the driving force behind that intuition. My discovery was the result of musical perception"
Albert Einstein, speaking about his theory of relativity |
|
|
|
Upavas
Upavas
Started Topics :
150
Posts :
3315
Posted : May 8, 2007 08:26
|
|
Spindrift
Spindrift
Started Topics :
33
Posts :
1560
Posted : May 8, 2007 11:38
|
Any resampling will involve a low-pass. Are you saying to add yet one more to make extra certain that there is no aliasing??
And I'm not talking about listening to a 96k signal, but actually doing sound generation and processing in a higher bitrate and then resample down to 44.1k. That will sound noticeably different unless the synth/fx has internal oversampling for oscillators and filters which is getting more and more common.
(``·.¸(``·.¸(``·.¸¸.·`´)¸.·`´)¸.·`´)
« .....www.ResonantEarth.com..... »
(¸.·`´(¸.·`´(¸.·`´``·.¸)``·.¸)``·.¸)
http://www.myspace.com/spindriftsounds
http://www.myspace.com/resonantearth |
|
|
|
texmex
Started Topics :
5
Posts :
189
Posted : May 9, 2007 17:23
|
@Upavas:
Only way to minimize aliasing is to use hq synths with interal oversampling or to use higher sampling rate and hq SRC to convert to the target khz.
If you have a synth that outputs aliased signal, the damage is already done. No filter/eq in the world will salvage it. |
|
|
|
psylevation
IsraTrance Full Member
Started Topics :
52
Posts :
841
Posted : May 10, 2007 03:45
|
so the top image is 44.1 and the bottom 96? Does it introduce a bunch of noise then? Is working at 48 better then converting the final to 44.1? or is it better to just work at 44.1 when working with mostly soft synths...?
~Airyck~
~Unoccupied Mind ~
Psyowa! |
|
|
|
shamantrixx
Started Topics :
7
Posts :
549
Posted : May 10, 2007 14:03
|
top and bottom image both represent results of conversion from 96 to 48 kHz. page is designed to select one software in upper window and another software in bottom window. Than pictures change and you can compare two softwares you wish to compare.
As well as changing the software you can scroll among few source signals (like sweep etc.)
As for the question of is it better to do in 44.1 when working with soft synths there are 2 major theories.
First is that our hearing is such that we don't need more than 44.1 kHz. I guess some of us don't hear any difference and most listeners at home would certainly have a hard time making any distinction.
Second theory is that with more "frames" (96000 frames per second compared to 44100 frames per second) you get smoother sound. The more "calculation points" the more accurate and less "digitalized" would sound be. This affects soft synths since we're exporting them. So instead of recording actual sound computer calculates the sound in a mathematical fashion.
When you zoom in a sample in sample editor on a huge scale you'll start noticing that waveform is made of squared lines looking like stairway rather than waveform. Each step is a calculation point and "motion" between two such points is a matter of our brain perceiving it like a constant sound rather than like individual steps played back in the speed of 44100 steps per second. Similar like static frames on the film that we perceive as motion.
So do we need a TV with 100 Hz instead of existing 50 Hz TV. We don't. We can watch a 50 Hz TV cause it's just fine. Is the picture better (smoother) on a 100 Hz TV? Of course it is. Does it affects viewing of VHS tape? Of course it doesn't. Does it affects viewing of DVD. Of course it does.
So in the end it depends from what you want. Are you "good enough" type or are you "it can allways be better" type 
"It occurred to me by intuition, and music was the driving force behind that intuition. My discovery was the result of musical perception"
Albert Einstein, speaking about his theory of relativity |
|
|
|
UnderTow
Started Topics :
9
Posts :
1448
Posted : May 10, 2007 15:03
|
Quote:
|
On 2007-05-10 14:03, shamantrixx wrote:
top and bottom image both represent results of conversion from 96 to 48 kHz.
|
|
44.1Khz actually but that was probably just a moment of inattention on your part.
Quote:
|
page is designed to select one software in upper window and another software in bottom window. Than pictures change and you can compare two softwares you wish to compare.
As well as changing the software you can scroll among few source signals (like sweep etc.)
|
|
The most revealing graph (at a fast glance) is the sweep graph. It shows the result of a sine wave sweep from 20Hz to 22Khz after conversion. For instance, load the "iZotope 64bit SRC steep, no aliasing" in the first window. It shows a nearly perfect sine sweep.
Now select the "Gold Wave 5.18" graph in the second window. It is the same sine sweep after some pretty bad conversion. You can see pre and post echo, aliasing and a huge mess really. 
Now if you select the 1Khz tone graph you can see all the junk that Gold Wave adds. The iZotope one is near perfect all the way down to -175 dB FS or so.
Now select the transition graph. This represents the cut-off from the anti-aliasing filter used in the SRCs. The iZotope graph shows that the filter leaves the signal completely untouched up to arround 21.5Khz and then cuts sharply to end up at below -90 dB before reaching 22Khz.
In the Gold Wave graph you can see that the filter allready affects signals at 16Khz and then doesn't really cut much of the frequencies above 22Khz. This is what is causing all the aliasing.
If you look at the passband graph you will see that the Gold Wave filter allready affects signals starting at arround 2Khz!!! In other words, Gold Wave is a piece of shit.
Quote:
|
As for the question of is it better to do in 44.1 when working with soft synths there are 2 major theories.
|
|
No, there is one theorem and one load of complete and utter BS from people that do not understand digital audio.
Quote:
|
First is that our hearing is such that we don't need more than 44.1 kHz.
|
|
Actually, it goes more like: We don't need more than +- 20Khz of bandwidth as we can't hear frequencies above that BUT as you need to allow ALL the sound to pass below +- 20Khz but NONE of the sound above 22.05 Khz (Half of 44.1Khz) you need a very steep filter to achieve that. Using a slightly higher sampling rate than 44.1Khz allows the filters to be less steep which, depending on the filter design and available DSP resources, can give you less filter artefacts.
The top people in digital audio tend to agree that the ideal sampling rate would be arround 64Khz and this is being very generous. The roll-off needed for an anti-aliasing filter at 64Khz SR is close to the natural high-frequency roll-off of air and no one is complaining about what air does to sound.
Quote:
|
I guess some of us don't hear any difference and most listeners at home would certainly have a hard time making any distinction.
|
|
It isn't that easy. It depends on several factors. For instance your converters. Alot of converters are crap at 44.1Khz because their anti-aliasing filters are crap. When you switch those converters to 88.2 or 96 Khz, they do sound better. Not because 96Khz is inherently better but because the converter is crap.
When someone doesn't hear any difference between 44.1Khz and 96 Khz is could just be because they have good converters to start with.
Then there is the processing done within the DAW (or synthesis in case of VST instruments). Some coders are better at making things work at 44.1Khz than others. There are various methods, like oversampling, to achieve non-aliased sound but not all coders know enough or they decided in favour of reducing CPU cycles over sound quality.
Quote:
|
Second theory is that with more "frames" (96000 frames per second compared to 44100 frames per second) you get smoother sound. The more "calculation points" the more accurate and less "digitalized" would sound be. This affects soft synths since we're exporting them. So instead of recording actual sound computer calculates the sound in a mathematical fashion.
|
|
This is utter bullshit. You only need two sampling points to PERFECTLY represent a sine wave. The above "theory" comes from people that don't understand how digital audio works (and thus should STFU).
Quote:
|
When you zoom in a sample in sample editor on a huge scale you'll start noticing that waveform is made of squared lines looking like stairway rather than waveform.
|
|
This is just bad software or software that is designed to use less CPU to represent the waveform. What you see when you look at the waveform in such an editor is NOT what comes out of your DACs.
Look at the waveform in Adobe Audition (Cool Edit). It shows the RECONSTRUCTED wave form as it comes out of your soundcard. No staircases there!
This is an interesting little experiment: Create a sinewave in Audition, zoom all the way in and grab one sample point. Drag it down (or up) and see how the wavefrom changes.
Quote:
|
Each step is a calculation point and "motion" between two such points is a matter of our brain perceiving it like a constant sound rather than like individual steps played back in the speed of 44100 steps per second.
|
|
No. It is not about how our brains perceives things. It is about the reconstruction filter in a converter "smoothing out the steps". What comes out of a (properly designed) converter has no steps!
If you really want to learn more about this, check out this white paper by Dan Lavry: http://www.lavryengineering.com/documents/Sampling_Theory.pdf
Or get Nika Aldrich's book "Digital Audio Explained (for the audio engineer)". It is very understandable and doesn't go too deep into the maths.
Quote:
|
Similar like static frames on the film that we perceive as motion.
So do we need a TV with 100 Hz instead of existing 50 Hz TV. We don't. We can watch a 50 Hz TV cause it's just fine. Is the picture better (smoother) on a 100 Hz TV? Of course it is. Does it affects viewing of VHS tape? Of course it doesn't. Does it affects viewing of DVD. Of course it does.
|
|
You can not simply compare sound and video this way because with sound we have the full bandwidth and dynamic range of human beings covered with 44.1Khz/16bit. With video we don't so increasing the numbers does affect what we perceive with our eyes.
Audio <-> Video comparisions and analogies just confuse things.
Quote:
|
So in the end it depends from what you want. Are you "good enough" type or are you "it can allways be better" type
|
|
No. In the end it is about using tools designed and created by people that understand audio DSP.
For instance, DiscoDSP Discovery does not have proper anti-aliasing so increasing the sampling rate from 44.1Khz to 96Khz does improve its sound. On the other hand, a synth like Rob Papen's Predator (with 16x internal oversampling) or something like Cakewalk's Rapture (with anti-aliasing in the design) do not benefit from increasing the sampling rate. (And before you tell me that they sound better at 96Khz, consider the quality of your soundcard).
UnderTow |
|
|
|
sy000321
IsraTrance Junior Member
Started Topics :
46
Posts :
1142
Posted : May 10, 2007 15:29
|
great post undertow
roll a joint or STFU :) |
|
|
|
sy000321
IsraTrance Junior Member
Started Topics :
46
Posts :
1142
Posted : May 10, 2007 15:36
|
one thing i dont understand is why people making psytrance (supposedly) are so obcessive with this stuff.
as someone day before, some distortion will only add character to your sound, more harmonics
it's not like we are recording a jazz band or anything
roll a joint or STFU :) |
|
|
|
psylevation
IsraTrance Full Member
Started Topics :
52
Posts :
841
Posted : May 10, 2007 17:02
|
I always know I can get quality information from Undertow....I now have a much better understanding of the whole thing and feel alot better about what I have to work with!
Thank you very much Undertow 
~Airyck~
~Unoccupied Mind ~
Psyowa! |
|
|
|
shamantrixx
Started Topics :
7
Posts :
549
Posted : May 10, 2007 17:36
|
Quote:
|
On 2007-05-10 15:03, UnderTow wrote:
And before you tell me that they sound better at 96Khz, consider the quality of your soundcard.
UnderTow
|
|
I don't mind you arguing me on each "step". It is healthy to doubt everything and double check everything. And here we get to the part that I mind about you. Instead of simply let people to change sample rate, export 8 bars of synth and HEAR the difference or DON'T hear any you waste your time making sure that nobody even tries. I wonder what's in it for you? You're not making any sense.
Only way to learn anything is the way of having experience. But you spent countless hours making sure that every inhabitant of this virtual community believes in YOUR truth without even trying for them self. Weird!
Back to converters...
some people don't hear any difference because they can't hear it. Like everything else hearing can also be trained and with focus on it people in any "hearing" business can hear a lot more than people who use their hearing for watching football, making a phone call and other "day to day" activities.
So since I have creamware converters and I still hear the difference before you suggest that I must be crazy or that I'm doing 2 much drugs (or some shit like that) consider one thing. Reading books or any other method of learning is fine as long as you have some actual experience.
So try investing some of your "forum" time to really listen. Not to listen and make comments about how you would make a better bass line or something like the usual chattering. Listen carefully and with time you will hear better. Until than realize that your chattering and repetition will just increase your fun club with people who are as def as you are. If you don't hear the difference... it's not knowledge that failing... it's your hearing
"It occurred to me by intuition, and music was the driving force behind that intuition. My discovery was the result of musical perception"
Albert Einstein, speaking about his theory of relativity |
|
|
|
psylevation
IsraTrance Full Member
Started Topics :
52
Posts :
841
Posted : May 10, 2007 17:52
|
Well, I have noticed a difference before, but I've been using 3 particular synths (Zebra 2, Sylenth1, & Trilogy) lately and I don't notice the difference as much anymore, what I found useful is that people can code things in a way at the same rates to make them work better at those rates. So these things I was hearing had more or less disappeared, which made me think that maybe my ears were playing tricks on me. I know next to nothing about DSP programing, so this is welcome information for me.
I will always listen for myself, but it's nice to have some help along the way, the school of hard knocks teaches best but it's also one of the slowest ways... Learning to trust in others sometimes helps as well. (which i'm trying to learn atm).
I see no need in constantly learning from the ground up...Solid as that foundation may be, I'm too impatient for that...
Undertow has given me good info in the past, which I tested myself and found to be true and helpful(hence the beginning to trust what Undertow has to say)
I do not blindly follow, I take the information I receive and think about it while performing the process or action it is in regard to.
It helps me to see if those steps someone else has taught me are actually the truth, and if they are then I have moved through an opening in the wall of learning and can move forward quickly. If it's wrong then I know thats one less spot I have to look.
Just my perspective on it...I can understand yours as well Shamantrixx
~Airyck~
~Airyck~
~Unoccupied Mind ~
Psyowa! |
|
|
|
psylevation
IsraTrance Full Member
Started Topics :
52
Posts :
841
Posted : May 10, 2007 17:55
|
and to be honest, what you said above what undertow said, is what I thought about this topic up till this point, it's seems very logical, but I don't know enough to make that judgement...
~Airyck~
~Unoccupied Mind ~
Psyowa! |
|
|
|
FluoSamsara (Oxygen)
IsraTrance Full Member
Started Topics :
84
Posts :
1164
Posted : May 10, 2007 21:12
|
|
I do notice a huge diference when recording external instruments in 96 or 44 (never did AB with soft synths) and I believe most people musicians or non musitians can notice the diference (if they care is another story...) |
|
|
|
UnderTow
Started Topics :
9
Posts :
1448
Posted : May 11, 2007 02:25
|
Quote:
|
On 2007-05-10 17:36, shamantrixx wrote:
And here we get to the part that I mind about you. Instead of simply let people to change sample rate, export 8 bars of synth and HEAR the difference or DON'T hear any you waste your time making sure that nobody even tries.
|
|
Not at all. Did I give a list of VSTs that do or don't alias? Did I give a list of converters that do or don't sound good at 44.1Khz? No.
Knowing a bit of theory does not exempt anyone from doing their own practical research to apply that theory to real world situations so as to fully understand the theory nor does the theory tell anyone what devices or plugins have properly applied said theory.
It is always a good idea to test things for yourself.
The only thing I am doing is countering your rubbish with some real facts. I am also giving some extra information to people so that they can interpret the graphs on the Infinitewave site a bit better.
And have you downloaded Audtion to check what I wrote? Probably not...
Quote:
|
I wonder what's in it for you? You're not
making any sense.
|
|
It probably doesn't make any sense to you, no. But sense isn't something I have seen much of in your posts.
Anyway, what is in it for me is dispelling myths and none-sense and sharing some of the knowledge I have accumulated over the years.
Quote:
|
Only way to learn anything is the way of having experience.
|
|
And your assumption that I am in any way against people gaining experience for themselves is yet another example of your jumping to erronous conclusions based on lack of knowledge and faulty logic.
Quote:
|
But you spent countless hours making sure that every inhabitant of this virtual community believes in YOUR truth without even trying for them self. Weird!
|
|
Again your faulty assumptions at work here.
It is not MY truth I am talking about. It is basic sampling theory. The fact that you don't know it is completely besides the point except maybe as a demonstration of what an arse you are to write stuff about something that you do not know or understand.
NO ONE with any understanding of sampling theory believes in your bogus "steps" theory. Only other idiots that have not done the necessary research or do not have the intellect to comprehend things beyond the seemingly obvious will spout such rubbish.
Quote:
|
Back to converters...
some people don't hear any difference because they can't hear it. Like everything else hearing can also be trained and with focus on it people in any "hearing" business can hear a lot more than people who use their hearing for watching football, making a phone call and other "day to day" activities.
|
|
Agreed.
Quote:
|
So since I have creamware converters
|
|
Meaning? They are nothing to write home about...
Quote:
|
and I still hear the difference
|
|
In between what exactly? You need to be a hell of a lot less vague when trying to make a point.
I won't even mention placebo effect... Oops, too late.
Quote:
|
before you suggest that I must be crazy or that I'm doing 2 much drugs (or some shit like that) consider one thing. Reading books or any other method of learning is fine as long as you have some actual experience.
So try investing some of your "forum" time to really listen. Not to listen and make comments about how you would make a better bass line or something like the usual chattering. Listen carefully and with time you will hear better.
|
|
And there you go again jumping to stupid conclusions. Your problem is that you are way too arrogant for your very limited intellect and are too damned stupid to realise it. You are deluded.
FYI I am a freelance sound engineer that gets called by the studios to record/mix stuff, do sound design, mastering etc. Literaly millions of people hear my work. I havn't applied to any studios or whatever in several years yet they still call me so I must be doing something right.
I also offer some studio design consultancy and stuff like that. On average I spend about 10 hours a day listening to stuff. Usually on equipment and in rooms that most of us can only dream of.
Oh and btw, if you knew my real name, you would recognise it on the credits page of the Infinitewave SRC site which we have been talking about...
Quote:
|
Until than realize that your chattering and repetition will just increase your fun club with people who are as def as you are. If you don't hear the difference... it's not knowledge that failing... it's your hearing
|
|
Now it is really time to let us hear some of your music. It is the only possible way for you to redeem yourself and proove that you are not the complete and utter fool that you seem to be.
UnderTow |
|
|
|
