Which samplerate for your DAW? [a sort of poll]
-
Mr Arkadin
- Posts: 3283
- Joined: Thu May 24, 2001 4:00 pm
of course it is (basically) true for ANY board, I just picked the EMU for it's extreme position in budget versus pretended (advertized) capabilities.Lima wrote:Astro, is this also true for our board? If the answer is "yes", we teorethically don't have any advantage using 24 bit resolution, right?So to say (just an example) an EMU 192khz 24bit card CANNOT produce precise results according to the physics it's based on.
If you don't believe it, simply calculate the required clock stability and the voltage difference between 2 bits...
but your conclusion puts the (original) statement out of context
Given there would be a perfect signal generator, then I am convinced that not a single 24 bit converter exists, which will be able to produce 2 identical sampling runs in the lower 4 bits because the signal interval is simply too low to be physically measured.
Ideally the samples would 'oscillate' around the 'real physical' values, making it a kind of dithering, which may even 'sound' very good.
but it is in no way precise
recently numeric precision was frequently associated with reproduction quality and 'good sound' - I don't think it's that simple.
But in any case I wanted to get away from specs focussing instead of listening.
'Our' cards sound good because they are designed in a certain way, driven by a specific software - and no question that a 24 bit headroom is simply a matter of convenience
cheers, Tom
you drive me crazy with your bits and byts;
noone here using 16 bit as me ?!
that makes me wonder;
have my ears grown old so that i dont hear the higher quality of the more bits?
lets see;
i spare the sampling rate question,
everyone hears synths sounding better at higher sampling rates why . . .
maybe some can enlighten me primitiv on the technical background;
but for audiorecordings which end in 44 anyway? test to come . .
so just have a look at the bits for now
here we go:
powerup PC; starting with a cleen project (no devs/no ins/no outs) plane project window;
sampling rate is on 44
i load ControlRoom device; load sts3k; load another sts3k; load a 3rd sts3k;
connecting the test from controlRoom to the LrecIN from sts3k (1/2/3) so all sts3ks are connected;
i took the presetted -20DB 1000Hz Sin as signalsource; have a look at the signal incomming to sts3ks;
i see -13.1 DB and wonder whats going on; stsbug or what?
i load meterbridge connect to controlroomTestOut and sts3ks recIN level is -20DB;
took a channelstrip for compare = -20DB;
took a dynamixer -20DB IN and -23DB out (as expected the -3DB for sum)
took a stm16/32 -20DB IN and -23.1DB out
took a stm24/48 -20DB in and -23.1DB out
took a stm48/96 -23.1DB IN and -23.1DB out whats that
pushing the PAN button to the left puts the level on -20DB for In and OUT!!
a PANBUG ?? on the 48/96;lucky i am to never used this mixer!!
!!NO: stm48/96 is the only one with presetted InPlace VUmode!!
back to sts3k i realised the presetted +DB gain and feel better for now;
you know; get no sleep last night . ..
here we go again:
making new progr./new keygroup and new wave on all sts3ks the 1. = 16Bit, 2nd = 24Bit, 3rd = 32Bit;
i check with windows Explorer the files created 16/24/32Bit files are ok. so the rec could start:
pushing the DB and rec button on all sts3ks ready for signal;
pushig the test sin gen.button to create signal in controlroomn;
record a few seconds and pressed stop on all sts3ks + yes for the replaceFile dialog;
check again with explorer; what the f*ck is that??
result is 3Files with 16Bit???????
my ears are old, my eys are old i feel old and very sad;
but you know, i man am a fighter stand up and shout:
you crazy litte germans sitting inside my DSPs, i am the master of the power
button to turn you off;
said and done;
powerup PC; starting with a new plane project;
controlroom only 2 sts3ks one at 24 the other at 32Bit;
to make it short result is 2 files with 24Bit !!!!!
NOTE: when you record with more than one sts equal waht bitdeep you set!
you finally record at the bits that you first set!! on all stses!!!!
so here we go again:
PowerDown/UP plane project only 1 sts3k recording the sin with 16Bit
extracting 1 full period saving;
power/down/up same for 24 Bit;
same for 32Bit
i had a sort of (Einschwingphase) to reach the full level which i wouldnt have expected in a digital application
but no problem;
Now i have 3 Files with one period of a sin at 1000Hz with 16/24/32Bit for further check;
I open wavlab ( i have 5.1) to compare the files;
at first to say the 32Bit doesnot open (not supported format bla bla)
so i compare 16 to 24 Bit and guess what they are completly the same!!
tricky steinacle ; i would have expected more for a 600Euro+ SW;
and i hope that the lazy developers have just used the same object for displaying different numbers
and not what it looks like !!
check the sampleeditor of sts 16Bit number o.k. but 24Bit are the same as 32Bit seems they using the same object too;
so i export to ascII text formated files to see the decimal numbers;
and guess what: the are completly the same numbers in all files !!!!!!
they are 16Bit!!!!!
so what ????
in winExplorer i have
16Bit = 280bytes size / 16384 size on disc ?
24 = 328 / 16384
32 = 374 / 16384
whats that?
not with me; now i want to know it
i wrote a small c++ programm to read the header + the samples in file and converts the hex to dec;
results (in hex of course) are stored in another file for coparison in wavlab to see if the sin is still a sin
at first the header:
16Bit = 272Bytes filesize / 45 samples in file (oK. i cutted one sample to much)
24 = 320 / 46
32 = 366 / 46
so that would fit to winExplorer exept the missing 8 Bytes;
45samples*2byte+32byte for the header = 122byte / makes 150byt for the CW copyright chunk
46*3+32=170 / makes 150 byte for Cw copyright
46*4+32=216 / 150 again for the copyright chunk
c++ programm works;
now the 1. sample:
16 Bit = 1103
24 = 282449
32 = 72292960
the second one is:
16 Bit = 2107
24 = 539516
32 = 138102588
---------------------
i had to say it was clear befor the test starts that 24Bit is 16Bit*256 and so on but
282449/256 = 1103.31640625 and
72292960/65536= 1103.10302734375
------------
539516/256 = 2107.484375
138102588/65536= 2107.27825927734375
so what again ? does this mean that 32Bit is better course of nearer to the final ending 16Bit than 24Bit
so this brings me nowhere and i had to look from another side;
16Bit is 256*256=65536 steps to have a range from 0 to 32767 for the positiv phase
that is 0 to 100DB right? so i have 327.67 steps for 1DB
and having 0.00305185094759971......DB of measurable level differeces
i guess noone could hear but i know there are always freaks out there who just *feel* it;
for them im afraid that there are no speakers available to support it but who knows;
24Bit is 256*256*256=16777216 steps / 0 to 8388608 ->
0.000011920930....DB
i man better man record at 32Bit
no i man better man record at 32 float
i man best man wait for 64Bit . . . .creazy!!!!
so to say the ADCs deliver 24Bit so why not recording 24Bit; thats an argument;
but see that its finaly the same than 16Bit;
the argument of further processing is in theory ok. for some as symbiote told us already;
but i learned:
the more effects you had to use the worser was the recording
and you cant be shure in what you get when dealing with higher numbers;
the difference is minimal!
better conzentrat in the recording itseft than pumping it up afterwards;
by the way the sin sounds like a sin equal what Bitdeep
sorry for long posting
good vibes
noone here using 16 bit as me ?!
that makes me wonder;
have my ears grown old so that i dont hear the higher quality of the more bits?
lets see;
i spare the sampling rate question,
everyone hears synths sounding better at higher sampling rates why . . .
maybe some can enlighten me primitiv on the technical background;
but for audiorecordings which end in 44 anyway? test to come . .
so just have a look at the bits for now
here we go:
powerup PC; starting with a cleen project (no devs/no ins/no outs) plane project window;
sampling rate is on 44
i load ControlRoom device; load sts3k; load another sts3k; load a 3rd sts3k;
connecting the test from controlRoom to the LrecIN from sts3k (1/2/3) so all sts3ks are connected;
i took the presetted -20DB 1000Hz Sin as signalsource; have a look at the signal incomming to sts3ks;
i see -13.1 DB and wonder whats going on; stsbug or what?
i load meterbridge connect to controlroomTestOut and sts3ks recIN level is -20DB;
took a channelstrip for compare = -20DB;
took a dynamixer -20DB IN and -23DB out (as expected the -3DB for sum)
took a stm16/32 -20DB IN and -23.1DB out
took a stm24/48 -20DB in and -23.1DB out
took a stm48/96 -23.1DB IN and -23.1DB out whats that
pushing the PAN button to the left puts the level on -20DB for In and OUT!!
a PANBUG ?? on the 48/96;lucky i am to never used this mixer!!
!!NO: stm48/96 is the only one with presetted InPlace VUmode!!
back to sts3k i realised the presetted +DB gain and feel better for now;
you know; get no sleep last night . ..
here we go again:
making new progr./new keygroup and new wave on all sts3ks the 1. = 16Bit, 2nd = 24Bit, 3rd = 32Bit;
i check with windows Explorer the files created 16/24/32Bit files are ok. so the rec could start:
pushing the DB and rec button on all sts3ks ready for signal;
pushig the test sin gen.button to create signal in controlroomn;
record a few seconds and pressed stop on all sts3ks + yes for the replaceFile dialog;
check again with explorer; what the f*ck is that??
result is 3Files with 16Bit???????
my ears are old, my eys are old i feel old and very sad;
but you know, i man am a fighter stand up and shout:
you crazy litte germans sitting inside my DSPs, i am the master of the power
button to turn you off;
said and done;
powerup PC; starting with a new plane project;
controlroom only 2 sts3ks one at 24 the other at 32Bit;
to make it short result is 2 files with 24Bit !!!!!
NOTE: when you record with more than one sts equal waht bitdeep you set!
you finally record at the bits that you first set!! on all stses!!!!
so here we go again:
PowerDown/UP plane project only 1 sts3k recording the sin with 16Bit
extracting 1 full period saving;
power/down/up same for 24 Bit;
same for 32Bit
i had a sort of (Einschwingphase) to reach the full level which i wouldnt have expected in a digital application
but no problem;
Now i have 3 Files with one period of a sin at 1000Hz with 16/24/32Bit for further check;
I open wavlab ( i have 5.1) to compare the files;
at first to say the 32Bit doesnot open (not supported format bla bla)
so i compare 16 to 24 Bit and guess what they are completly the same!!
tricky steinacle ; i would have expected more for a 600Euro+ SW;
and i hope that the lazy developers have just used the same object for displaying different numbers
and not what it looks like !!
check the sampleeditor of sts 16Bit number o.k. but 24Bit are the same as 32Bit seems they using the same object too;
so i export to ascII text formated files to see the decimal numbers;
and guess what: the are completly the same numbers in all files !!!!!!
they are 16Bit!!!!!
so what ????
in winExplorer i have
16Bit = 280bytes size / 16384 size on disc ?
24 = 328 / 16384
32 = 374 / 16384
whats that?
not with me; now i want to know it
i wrote a small c++ programm to read the header + the samples in file and converts the hex to dec;
results (in hex of course) are stored in another file for coparison in wavlab to see if the sin is still a sin
at first the header:
16Bit = 272Bytes filesize / 45 samples in file (oK. i cutted one sample to much)
24 = 320 / 46
32 = 366 / 46
so that would fit to winExplorer exept the missing 8 Bytes;
45samples*2byte+32byte for the header = 122byte / makes 150byt for the CW copyright chunk
46*3+32=170 / makes 150 byte for Cw copyright
46*4+32=216 / 150 again for the copyright chunk
c++ programm works;
now the 1. sample:
16 Bit = 1103
24 = 282449
32 = 72292960
the second one is:
16 Bit = 2107
24 = 539516
32 = 138102588
---------------------
i had to say it was clear befor the test starts that 24Bit is 16Bit*256 and so on but
282449/256 = 1103.31640625 and
72292960/65536= 1103.10302734375
------------
539516/256 = 2107.484375
138102588/65536= 2107.27825927734375
so what again ? does this mean that 32Bit is better course of nearer to the final ending 16Bit than 24Bit
so this brings me nowhere and i had to look from another side;
16Bit is 256*256=65536 steps to have a range from 0 to 32767 for the positiv phase
that is 0 to 100DB right? so i have 327.67 steps for 1DB
and having 0.00305185094759971......DB of measurable level differeces
i guess noone could hear but i know there are always freaks out there who just *feel* it;
for them im afraid that there are no speakers available to support it but who knows;
24Bit is 256*256*256=16777216 steps / 0 to 8388608 ->
0.000011920930....DB
i man better man record at 32Bit
no i man better man record at 32 float
i man best man wait for 64Bit . . . .creazy!!!!
so to say the ADCs deliver 24Bit so why not recording 24Bit; thats an argument;
but see that its finaly the same than 16Bit;
the argument of further processing is in theory ok. for some as symbiote told us already;
but i learned:
the more effects you had to use the worser was the recording
and you cant be shure in what you get when dealing with higher numbers;
the difference is minimal!
better conzentrat in the recording itseft than pumping it up afterwards;
by the way the sin sounds like a sin equal what Bitdeep
sorry for long posting
good vibes
Very very interesting post! Thank you for the efford you put in making it!
I also use Asio2 16 bit @ 44100.
Now I know that the difference is small. But what about the DSP usage?
Passing from 16 bit to 24 bit to 32 bit is there a great increase of the demand or not?
Finally, can we quantify in terms of amount of things we can do with a low number of dsp (eg.: with a scope home you can mix 10 tracks if you use 16 bit instead of 8 tracks if you use 24 bit)?
I also use Asio2 16 bit @ 44100.
Now I know that the difference is small. But what about the DSP usage?
Passing from 16 bit to 24 bit to 32 bit is there a great increase of the demand or not?
Finally, can we quantify in terms of amount of things we can do with a low number of dsp (eg.: with a scope home you can mix 10 tracks if you use 16 bit instead of 8 tracks if you use 24 bit)?
Welcome to the dawning of a new empire
and of course, this is the reason that 16/44.1k was chosen for cd audio, file size/resource usage vs. hearable frequency range/capability of playback hardware.
as astro pointed out, recording at a slightly higher bitdepth just adds a bit of headroom in maintaining good bitdepth in a sample, as i said, reducing the need for a compressor to keep the signal from falling into that 8bit and below level where things sound terrible.. either way, a good recording of something worth hearing is gonna be just as good to hear at 16bit or 32bit so tgstgs' suggestion to pay more attention to the recording than the bit depth is a point well taken.
file size/resource usage vs. hearable frequency range/capability of playback hardware.as astro pointed out, recording at a slightly higher bitdepth just adds a bit of headroom in maintaining good bitdepth in a sample, as i said, reducing the need for a compressor to keep the signal from falling into that 8bit and below level where things sound terrible.. either way, a good recording of something worth hearing is gonna be just as good to hear at 16bit or 32bit so tgstgs' suggestion to pay more attention to the recording than the bit depth is a point well taken.
-
Counterparts
- Posts: 1963
- Joined: Tue Aug 19, 2003 4:00 pm
- Location: Bath, England
i vote for the current minority : 24 bit 48khz
i like the sound enhancement at 48khz too much to do without it.
but: my mixes and "masterings" are not yet as good as they can be. i had quite a development in the sound of my mixes this year not because of the samplerate, but because of training my ears (what's remaining of them
) and the better realisation of knowledge. that did a lot more to the sound than moving from 44.1 to 48.
16bit for kicks, basses and all the loud stuff with little dynamic
pro 48: you can render straight to mp3
: 24 bit 48khz i like the sound enhancement at 48khz too much to do without it.
but: my mixes and "masterings" are not yet as good as they can be. i had quite a development in the sound of my mixes this year not because of the samplerate, but because of training my ears (what's remaining of them
) and the better realisation of knowledge. that did a lot more to the sound than moving from 44.1 to 48.16bit for kicks, basses and all the loud stuff with little dynamic
pro 48: you can render straight to mp3
i meant that you don't need to convert to 44.1/16 when you render to mp3.
but i'm not sure if it makes sense to produce in 48khz all the way, when it's only meant for mp3 (who does that?), because with mp3 there's not much going on in the high ends.
and with a good algo there shouldn't be a loss from 48 to 44.1 compared to the producing at 44.1 from the beginning.
i just like the little difference from 44.1 to 48.
there are only two reasons/situations where it makes sense to me but that's enough for now:
- listening to the music on a computer (wav not mp3) at (somebodies) home.
- playing a "laptop-live-gig" (which i not yet have )
and i'm on 6dsps only atm, so there is no way to place everything into one project anyway...
i'm really surprised how few people use 48khz here!
and some are even satisfied with 16bit! wow!
maybe i don't get it, or the priorities are different...but i'm sure it's not only esoteric
but i'm not sure if it makes sense to produce in 48khz all the way, when it's only meant for mp3 (who does that?), because with mp3 there's not much going on in the high ends.
and with a good algo there shouldn't be a loss from 48 to 44.1 compared to the producing at 44.1 from the beginning.
i just like the little difference from 44.1 to 48.
there are only two reasons/situations where it makes sense to me but that's enough for now:
- listening to the music on a computer (wav not mp3) at (somebodies) home.
- playing a "laptop-live-gig" (which i not yet have
)and i'm on 6dsps only atm, so there is no way to place everything into one project anyway...
i'm really surprised how few people use 48khz here!
and some are even satisfied with 16bit! wow!
maybe i don't get it, or the priorities are different...but i'm sure it's not only esoteric
Last edited by pling on Mon Oct 23, 2006 3:56 pm, edited 1 time in total.
-
Herr Voigt
- Posts: 624
- Joined: Sun Jan 06, 2002 4:00 pm
- Location: germany, east
yes but having creamware cards is supposed to be pro, no matter if 44.1 masterning and mixing is ok.
so the question is:
what is the samplerate used in big professional productions?
i mean madonna, micheal jackson and that kind of perfect sound.
ok, if we talk about thriller(best selling album of all times) i suppose it was all analog.
anyway the above artists are just example cause we might talk about many more reaching top quality in their sound including classical recordings, or jazz.
anyway what is the samplerate mostly used in big professional productions?
so the question is:
what is the samplerate used in big professional productions?
i mean madonna, micheal jackson and that kind of perfect sound.
ok, if we talk about thriller(best selling album of all times) i suppose it was all analog.
anyway the above artists are just example cause we might talk about many more reaching top quality in their sound including classical recordings, or jazz.
anyway what is the samplerate mostly used in big professional productions?
it really depends on the producer and engineer. five years ago most worked at 44.1 or 48k. these days, people with unlimited budgets use the most exotic gear available. it's a point worth discussing if sound quality has improved all that much. the best engineers and producers got great sounds 10 years ago without 192k. the current ones often can't live without 192k. most of this "best sound" is people talking out their butt.ARCADIOS wrote: anyway what is the samplerate mostly used in big professional productions?
the sample rate is just one element. above 44.1k it's the smallest element. the cableing used is much more important(and expensive)......