| Author
|
FL Producer 5 Cubase SX 2 comparison.
|
H2O
IsraTrance Junior Member
Started Topics :
16
Posts :
352
Posted : Dec 27, 2004 21:14
|
Spindrift ear comparison test will was original plan also, we would us it. I agree that musicians should trust their ears but some very small noise as we discover in LF Producer is not so heard by ear. These cahin of tests is for musicians, developers and sound designers at once.
We do not ask muscians to leave this conversation, we just want objective data, objective means you should create exactly the same sequence and compare by ear. I assume noone realy done it, if yes then post a link for us all to compare signals. Different sequences can sound different. |
|
|
Spindrift
Spindrift
Started Topics :
33
Posts :
1560
Posted : Dec 27, 2004 21:29
|
Quote:
| I agree that musicians should trust their ears but some very small noise as we discover in LF Producer is not so heard by ear. |
|
It boils down to subjective listening test weather or not it's from a producer or developer angle.
If it's not audible in blind test it don't need to bother anyone.
Quote:
| We do not ask muscians to leave this conversation, we just want objective data, objective means you should create exactly the same sequence and compare by ear. I assume noone realy done it, if yes then post a link for us all to compare signals. |
|
I said before repeated times I would be willing to formulate such test and carry them out.
You are the first to say that you also would like a test like that in this thread.
Since I had no response for that suggestion before and had to endure so much crap from the people that would be involved I don't have much feeling right now to carry out that work.
And I would probably not be suitable candidate to be involved with this community considering the problems I have understanding the individuals involved.
But please do test like that as well instead of all different versions of all software on all bit resolutions.
If nothing else it would shut me up 
 (``·.¸(``·.¸(``·.¸¸.·`´)¸.·`´)¸.·`´)
« .....www.ResonantEarth.com..... »
(¸.·`´(¸.·`´(¸.·`´``·.¸)``·.¸)``·.¸)
http://www.myspace.com/spindriftsounds
http://www.myspace.com/resonantearth |
|
|
|
ZilDoggo
Started Topics :
4
Posts :
663
Posted : Dec 27, 2004 21:32
|
"Sorry but that comment is an arrogant lousy argument about truncated 16 bit vs. 24 bit. "
uuhm, FYI ALL samples are somehow truncated.
if a sample would not be truncated then you would need infinite bitdepth to encode it !!. this is the very basis (and one of the two basic limitations) of digital sampling.
so there is nothing bad about truncation except that you lose resolution (bit depth).
but the truncating itself has very little artifacts (basically some rounding off errors at the LSB).
not more than a level change.
in fact, level changes can under some circumstances lead to bigger errors than truncation alone.
and when you cut the level then you are also truncating.
so truncation is trivial.
"You look as stupid to me as someone who think they can tell how a piece of hardware sounds by looking at the schematics of the unit. "
well to me you look stupid because we are dealing with digital information and software here.
schematics wont tell me everything that's going on in an electronic system (unless its mostly digital) but software is much more pragmatic.
you can see what is being done to the data under what circumstances.
whats more important tho is that we test the sequencers OUTPUTS. which means we are not concerned with algorithms or processes.
we just look at what gets sent out towards your monitors.
i cant imagine a fairer test then that.
you can actually deduce quite a lot from tests like these so i dont understand why you are against us doing these tests.
why are you afraid of it? why do you go through great lenght to say that you dont like this? why are you still here?
cos you have made your point several times now and noone seems interested.
and you got awnsers to ALL your questions and still you complain!.,
WTF., people take time to awnser you and you give them the same sheit over and over?
you quickly jump to the next thing when someone proves you wrong in one point.
either that or you repeat yourself without noticing you already got an awnser but didnt read and respond to it.,
move along PLEASE.
"You might think you understand to 99.9% how exactly how all sequencers on the market work,"
i never claimed that, you twist my words.
i dont know how 99.9% of all sequencers work.
i DO know how most of the insides work but that's not 99.9%.
READ , otherwise this discussion is pointless.
and dont put words in my mouth again.
"But stop asking people to leave the thread when they think a developers discussion is out of place here, because it is."
if you mean that "here" is this thread then yes, if you are not interested and think this is stupid and have nothing constructive to say then please close your mouth and leave.
because if you think this is not the place for you then it propably is not.
|
|
|
|
Spindrift
Spindrift
Started Topics :
33
Posts :
1560
Posted : Dec 27, 2004 21:52
|
Quote:
| schematics wont tell me everything that's going on in an electronic system (unless its mostly digital) but software is much more pragmatic.
you can see what is being done to the data under what circumstances.
|
|
Sure...if the software is open source and you are a good coder.
Quote:
| and you got awnsers to ALL your questions |
|
Not in a comprehensible way until UnderTow and H20 explained their views.
I said I am satisfied with the answers and even if I don't agree 100%, and I wish them good luck.
Many other people seem too busy being rude to formulate a proper reply to a couple of simple questions.
(``·.¸(``·.¸(``·.¸¸.·`´)¸.·`´)¸.·`´)
« .....www.ResonantEarth.com..... »
(¸.·`´(¸.·`´(¸.·`´``·.¸)``·.¸)``·.¸)
http://www.myspace.com/spindriftsounds
http://www.myspace.com/resonantearth |
|
|
|
H2O
IsraTrance Junior Member
Started Topics :
16
Posts :
352
Posted : Dec 27, 2004 21:54
|
|
Spindrift it was exactly was me and Cane done alone, we don't have a web space to upload 30 channels sequences of Cubase and FL Producer + Export files + Samples. We need 10mb webspace non sepearated one, our host allows only 800kb files. |
|
|
|
H2O
IsraTrance Junior Member
Started Topics :
16
Posts :
352
Posted : Dec 27, 2004 22:07
|
|
Digital Performer, Samplitude, Orion, Project 5 and other not mentioned here sequencers users, please join us. |
|
|
|
ZilDoggo
Started Topics :
4
Posts :
663
Posted : Dec 27, 2004 22:12
|
"It boils down to subjective listening test weather or not it's from a producer or developer angle.
If it's not audible in blind test it don't need to bother anyone."
these kind of tests are very difficult to do because they rely heavily on acoustics and monitoring chain.
it is much much much more difficult to do proper listening tests than a mathematic output comparison.
it is therefore much more difficult to make any asumptions based on listening tests.
furthermore, listening tests can hide certain aspects of the sound that can become important for for instance PA playback.
one such example is MP3.
some ppl say mp3's are always ok because under controlled tests at normal listening levels most ppl cant hear the ifference.
but that changes radically when you play mp3 over a big system with a less then optimal frequency response. then some mp3 errors start shining through.
ears are differently sensitive at different listening levels so you would actually have to conduct listening tests at different levels at different frequencies to get meaningfull results from such tests.
you would need calibrated equipment to make all tests equal.
and everyone taking the est should use identical equipment for the monitoring chain.
and you propably want to test every persons opinion on several monitoring chanins and thake a mean value from that.
the beauty of digital systems is that they are already calibrated (usually to 0db with a straight frequency and phase response)
this is why doing the test like it is now being done is so easy and effective.
you can readily compare the output of any sequencer and have meaningfull results.
now try that with the output of your brain as heared by your ears as played by your monitors as given by your amp as converted by your d-a convertor,.
see, it gets complicated very easily.,
trust me, a LOT more care must go into a proper listening test before you can give any conclusion.
big audio equipment manufacturers have such setups.
but this is mostly because they are leftovers from the analog times.
if you design someting analog then it is very ificult to predict the exact output.
temerature, humidity, component quality, etc, etc all have an influence.
so you are kindof forced to test your design in the real world.
you KNEW that the design will not behave exacly like you designed it (most of the times) so you had to make sure that whatever is left is adequate enough.
enter computers.
computers are digital, which means that they deal with digital information which means they are ABSOLUTE.
what you hear comming out of your soundcard is quite litterary the same thing that came out of your sequencer.
so you can leave out the whole listening part and just compare the outputs of the sequencers.
then you will already know some things that happened to the audio and you can actually predict to a certain point how it will sound.
one thing is easy to understand.
if you compare two sequencers and the outputs are the same then you wont hear a difference.
this is already an interesting thing to know for sure.
"I said before repeated times I would be willing to formulate such test and carry them out."
i asked you several times to formulate it.
but you still havent done so.
are you accusing others of the things you do yourself??.,
"You are the first to say that you also would like a test like that in this thread."
no, i said it before too. you just didnt see.
"Since I had no response for that suggestion before"
yes you had.,
from me at least 2 times.,
stop whining.,
"But please do test like that as well instead of all different versions of all software on all bit resolutions.
If nothing else it would shut me up "
YES!.,
because a listening test is also important! 
so be patient and constructive.
"Sure...if the software is open source and you are a good coder. "
i dont agree because most audio processes are pretty straight forward linear algorithms.
you can clearly see their working (or not) on the outputs.
you do need to know how it is usually done then you can recognise it easier.
and in this case, there is nothing magical about a summing algorithm. 1+1=2
this is what most sequencers do and what i expect from a sequencer.
we know that most sequencers do this because we compared them and they were the same as mathematically adding the numbers.
greetos.,
aka., |
|
|
|
Spindrift
Spindrift
Started Topics :
33
Posts :
1560
Posted : Dec 27, 2004 22:35
|
How do you think they test dithering algorithms, standards for bit resolutions and sample frequency...etc??
They just analyze the code over and over again?
How many parts of they way a software is coded is dependent on listening test at some point?
I am aware of how listening test SHOULD be conducted and that there is issues involved.
But if you try to be a bit less theoretical I think you realize that if people could compare the output of the sequencers online and make a judgement on the output there would be some quite conclusive results even if not 100% accurate.
About making a comprehensive test I can find that you said "maybe it should be the the next thing?".
In your last few posts you seem to want to test all versions of all software on both 16 and 24 bit.
I envy the amount of time the people involved in this test must have.
(``·.¸(``·.¸(``·.¸¸.·`´)¸.·`´)¸.·`´)
« .....www.ResonantEarth.com..... »
(¸.·`´(¸.·`´(¸.·`´``·.¸)``·.¸)``·.¸)
http://www.myspace.com/spindriftsounds
http://www.myspace.com/resonantearth |
|
|
|
H2O
IsraTrance Junior Member
Started Topics :
16
Posts :
352
Posted : Dec 27, 2004 22:50
|
to both of you, ear comparison test features you discuss are perfect situation which never seems to happen in real life, moreover on this subjective forum.
Spindrift, let's assume we will post here output signal for users to compare, how possible the accurate result when most of here users are emotinal, means know the results before the tests which leads us to subjective results. By emotions i mean pre test emotions not after test ones.
example,
FL Producer sequence heard by Darkpsy will be never identical to Cubase sequence since he is sure "Cubase sounds better than FL Producer, has better sound engine and it is amateur".
I think you should get my point.
|
|
|
|
ZilDoggo
Started Topics :
4
Posts :
663
Posted : Dec 28, 2004 00:37
|
" How do you think they test dithering algorithms, standards for bit resolutions and sample frequency...etc?? "
well., any dithering algorithm in existance was invented on paper.
it was furthermore proved on paper that it would work.
in fact most digital technologies are invented loooong before they can be implemented properly.
this is only now changig as computer power becomes more and more redundant.
but a lot of things (like sampling and dithering and digital synthesis) were mathematical inventions mostly done by the army for signal processing applications.
so, yes, testing of dithering algo's happens on paper in the heads (and computers) of mathematicians.
btw, dithering was proposed in a paper in 1964 or so, possibly earlier.,
and we all know there was not much sampling going on then.
also, we have a pretty good standard data on how ppl hear stuff so we can predict a lot of things.
Philips was one of the companies that did xtensive testing in this field and with that data they were able to build a psychoacoustical model of human hearing.
which gave us MP3 compression ratio's of 1/12 or so without significant loss of quality at optimal listening conditions.
but the mp3 algorithm was mostly invented on paper with use of the psychoacoustical model.., no humans were involved untill it was ready.,
in fact, for something like mp3, you cannot just do listening tests while you invent it because it requires a series of fully working bits to hear anything at all.,
so before you hear anything it should be 80~90% ready programmed!
testing is done , but it has no big part in the inventing itself. invention requires briliance of thought.
you do not invent dithering by listening to undithered stuff and a friend says "would be fun to add noise, huh?".
same goes for most design things unless a human is part of a feedback system (like a user interface).
and even then, there is so much data about how humans experience various stuff that you can just use statistical data instead of doing real tests ., and the beauty is that it works!
"How many parts of they way a software is coded is dependent on listening test at some point? "
of sequencer software i think it is not so much.,
what is far more important to test is the user interface as this will be directly related to usage.
also, any built in effects and processors should be tested with a listening test (one bad example i think was cubase VST .,., it had realy unusable eq's. they were not musical in any way)
but as for the rest of the sequencer itself, well, its more or less only about mixing the different channels.
and since that is a well defined mathematical function you already KNOW how it will sound when you implement it.
summing has been around for a long time now.
so if you just sum then you dont have to worry about sounding different.
"But if you try to be a bit less theoretical I think you realize that if people could compare the output of the sequencers online and make a judgement on the output there would be some quite conclusive results even if not 100% accurate. "
that's what i am hoping for too.
but would you not like to know WHY those results are the way they are ?
"About making a comprehensive test I can find that you said "maybe it should be the the next thing?". "
page 10, first post:
ears are not objective.
but i agree that objective listening should be part of any quality assessment,.
(and with objective listening i mean 'as objective as possible' of course )
i didnt suggest that it should be the next thing litterary
but i DO agree that is should be part of any assessment of any audio gear.
"In your last few posts you seem to want to test all versions of all software on both 16 and 24 bit. "
well, that is basically true, but not sure it will happen
i think its simple.,
16 bits will do.
24 bits will do too,
if someone has problems with only one of the two then we should do both and see the difference,.
i've done it and can tell you from experience tthat phasing problems have the same magnitude in both bit depths and rounding errors scale downwards with increased bit depth.
and this is all acording to theory.,
soo, because i know some theory i dont have to test both 16 and 24 bit files because i know where certain artifacts (if they exist) should show up. that is their nature.
if i find artifacts in other places then i know that its a new kind of artefact.
this kind of testing is almost impossible to do with just listening tests because:
a) you dont hear the artefacts with just one (or a few) tracks playing and no effects.
b) the kinds of effects we HEAR when there are artefacts on individual tracks in a busy mix are a bit vague.
they can be unclarity, super clarity, faded stereo image, lack of detail, compression etc etc etc.
in such a situation it is difficult to say wether it was the starting samples or the mixer or dac or the amp or the monitor.
btw, with artefacts i mean anything that is added by the engine that was not present in the original or the sum of the originals.,
not nessesarily negative
(i say this because artefact sounds a bit negative)
|
|
|
|
High Pulse
Darkpsy
Started Topics :
57
Posts :
1187
Posted : Dec 28, 2004 05:33
|
Quote:
|
On 2004-12-27 20:33, ZilDoggo wrote:
" why not test against all the last version from logic , cubase, nuendo, sonar , protools etc etc ????? "
who said we will not test them????
everything will be tested!! (BUAHAHAHAHAAH )
why fruity right now?
cause it was just released and ppl were wondering wether fruities engine got upgraded (because tests were done earlier with fl4.)
what i'm wondering about now is wether there is a difference between fl4 and fl5
btw. another reason why the test is not yet conducted on newer versions of logic and cubase is becase it is HIGHLY UNLIKELY they would change their engine dramatically.
they would have A LOT OF disstressed customers.
possibly the same thing goes for fruityloops users and why image line didnt change the engine.
and if you are tiered of it then why keep on reading??
greetos.,
aka.,
|
|
well we need to be ocupied ehhe :S
but why ppl indeed of beeing here syaing this is bether that do that and blabla .,,, start making a final test .. FL5 vs logic 7 pro vs cubase sx 3 , vs protools 65Le VS sonar (last version i dont know wich one is ) .. and here we see what is what and wich give bether result .. and etc ...
so who is doing that ., start making the final test .. and stop say that this is that and that is this ,, everybody is going to th same direction ( is that everybody is saying good from each host they'r work) so start making this TEST ..
"HIGH PULSE" AKA FUSION OF DARKPSY PROJECTS
http://www.myspace.com/highpulsemusic
http://www.beatbiz.net/artists/High-Pulse <- musiC FOR SELL. |
|
|
|
Dreamthief
Started Topics :
3
Posts :
47
Posted : Dec 28, 2004 06:19
|
|
I still don't understand how this test is meant to be done. What kind of parameters will be employed? What methods will be used? What are the expected results? How will you differentiate between expected results and the results achieved? If you are testing the output, will you be testing panning laws? dynamic range? what about latency issues? [srug] sorry I didn't come into the channel the other day, but I am trying to finish a new track. Maybe these questions could be answered?
|
|
|
|
Cane
Started Topics :
0
Posts :
47
Posted : Dec 28, 2004 14:52
|
Yes, Darkpsy our dear lord of darkness. We must obey.
Dreamthief, panning laws very difficult to test since every program has it's own panning laws so we do it blindly. Dynamic range, i'm not sure it has to do with sequencer. Latency, yes it is a part of the test. If you realy wish to be a part of it, join a forum where we discuss these tests and after we finnish will post them on site(not this one). |
|
|
|
orik
IsraTrance Junior Member
Started Topics :
45
Posts :
317
Posted : Dec 28, 2004 16:03
|
|
guys.....give it a rest...your going in circles....just zula. |
|
|
|
High Pulse
Darkpsy
Started Topics :
57
Posts :
1187
Posted : Dec 28, 2004 23:21
|
|
|
