PlanetZ forums - scopeusers.com

I decided to start a new thread because the other has too much data to you can help.
I have find that the problem is not on the STS, is on the midi. I tried this: used the Terratec midi in on Cubase, and sent it to Scope through "Sequencer Midi Source 1", and conected to STS. The result was the same. 8 ms. So here the midi went to windows first, and then for the dsp.
So when I use the "PS Midi A Sourse", probably happens the same, and the midi don´t go directly to the DSP, and perhaps go first to the windows driver.
How can occur this? It´s supposed PS Midi A Sourse go directly to DSP, but the result claims no!

how do you test the sts latency?

Cubase is fairly notorious for bad MIDI timing. I don't want to start an argument about it here, but do be aware that there can be issues - make sure you do your test in another host.

When you test things like this, there are many factors at work - you should make sure that you test each factor before you make any conclusions.

afaik midi forwarding ('through') takes about 5ms - imho it's perfectly ok
of course not if it's directly connected to the the hardware midi source...

cheers, Tom

<font size=-1>[ This Message was edited by: astroman on 2006-04-07 15:59 ]</font>

First conclusion, is what easy is to make people believe. Just is told on Creamware page, that STS will have less latency than a hardware sampler, and everybody is pretty convinced of it, even without have tested it, no here, please, just read my other thread where I explained more detailed the test, and before you come to conclusion that I´m a pretty lost guy that is missing many factors, if you have a little time for the reading, you´ll come to conslusion that the test takes out some unknown issues.
Since now, nobody have read nothing I have post, and just conclude, I´m wrong. I would like you to understand the test first, then discuss, have to say, that is very simple.
At this moment I have found some issue. The latency of the STS, gives the same result, if I send the midi to windows, and againg to STS, that if I conect directly the midi in to STS, what this means?
Clearly, that Midi In A source, what is supposed is a physical input, really doesn´t go directly to the dsp, what would must do, instead (having account the results), it seems go to windows, and return to the card, what means a quite absurd increasing of latency.
Hope, there are anyone here who really think of it, and even can test it himself.

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-08 05:49 ]</font>

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-08 05:50 ]</font>

Did you do your test with a different sequencer yet? It's up to you to create some credibility for yourself if you're new to a forum - it's irresponsible to make speculative claims about hardware without taking proper precautions to eliminate all the different factors.

I´m really don´t know if better laugh or cry.
With all my repects, no sequencer is involved.
I use Cool edit Pro to record the signals, and measure just that the latency.
I´ll be direct: Routing from hardware midi in called "PS Midi A Source", to STS gives just the same result as sending first to sequencer and then to STS. So is absurd your question because the sequencer isn´t adding nothing bad, instead, I´m claming the bad of "PS Midi A Source", because would must be faster, and isn´t. The sequencer gives no bad result, so why ask for it?,
With all my respects, seems a new recently guy in this forum (me) have found a myth you all guys have been unknown.

Is supposed that STS sampler and Scop synhs, have very low latency, because of DSPs, while testing tells, really, the midi goes first to windows, and returning to dsp, instead going directly from midi in-to dsp, what everybody have been afirming lots of times. I´ll post 4 proyect files here, very simple to do really.

On 2006-04-07 15:48, astroman wrote:
afaik midi forwarding ('through') takes about 5ms - imho it's perfectly ok
of course not if it's directly connected to the the hardware midi source...

cheers, Tom

I´m afraid that it is, even worse, conecting directly to midi source, gives just the same result as from sequencer.(note: must use "non emulated" midi outputs, on sequencer)
Have a look at the 4 files I have added. 2 for STS ,2 for modular.
Seems, the hardware midi source, is not what it´s names tells.
I called them "midi hiden traveling"
http://www.planetz.com/forums/viewforum ... otal_forum


<font size=-1>[ This Message was edited by: lagoausente on 2006-04-08 07:17 ]</font>

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-08 07:18 ]</font>

On 2006-04-08 06:52, lagoausente wrote:
...With all my respects, seems a new recently guy in this forum (me) have found a myth you all guys have been unknown...

congrats
but I'd file it under irrelevant anyway, just like last weeks revelation that all Scope synths (except Flexor!) don't adjust their envelopes in realtime, but with the next keydown event.
Either way would have it's pro and cons - whatever decision - it would never be perfect and cover all cases.

At least the current method didn't disturb hardcore testers, even pros from magazines didn't care and for 7 years noone complained.
Looks like the decision was the right way.

Now there is this latency problem which is that the midiports should deviate, but don't - at least that's my understanding.

It's not that easy to read and honestly - if 4-5 ms drown a person in disappointment I'd rather skip the topic (unless curiosity drives me)
I respect your personal preferences about what to expect from your system - I hope you do the same with my conclusions about exactly that fact

I don't believe that many hardware samplers have 5 ms latency - 10 to 20ms are more likely (but that's just(!) a personal guess from the hardware usually found inside)

As mentioned the exact latency IS irrelevant because it doesn't matter if you compensate 4,5 or 8 ms - as my favourite quote goes: either it's sample accurate, or you can ignore it

I have a (asio) latency of 25ms, yet all my signals are nicely time-aligned BEFORE the monitor (and adjusted where required).
Of course I hear them much earlier than 25ms, but if that's 2, 8, or 12ms doesn't bother me at all.

There is not one single device in scope that's truely latency free in the technical sense of the word - as it would be impossible - there's always a buffer
In some cases (say you stack hi-hats) this has to be carefully compensated, in other cases it doesn't matter at all.

If you find an anomaly in some of the devices that is of course much appreciated, but don't forget that (given your finding is right) the context is rather exotic.
People usually sequence their stuff and even if you play live to that the 5 ms are just 2 steps in another direction - or 1.5 m on stage...

cheers, Tom

On 2006-04-08 09:21, astroman wrote:

On 2006-04-08 06:52, lagoausente wrote:

It's not that easy to read and honestly - if 4-5 ms drown a person in disappointment
cheers, Tom

The disappointment is not about. 4.5 ms, 8, or 20, is about the Scope is selled as a faster device, with lower latency. So really, I don´t think the vSTi is "better than STS because of 2 ms. But I think is dissapointment that what is called a "solution", real isn´t, even very little worse. But in definitive is untrue advertisement.
Regarding to the numbers, I agree, that 8 ms, is neglectable playing a keyboard, but also think, is not so neglectable when playing drums live.
Perhaps I looked for in the wrong place a real hardware sampler, but as well, I have read on many sites, that "is better than hardware sampler", while still is no better than sofware ones. I don´t like swindles.
Scope is very good device, but trues a trues, en lies are lies. Just that.

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-08 10:00 ]</font>

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-08 10:00 ]</font>

i think one of the best soft samplers is the emulator x, but the sts 5000 is still the better for timestretching and pitch shifting quality.

I think it needs some interface improvement and some feature such as direct from disk stream.
I use the sts 5000 just for realtime timestretching operations, otherwise the emu x is my sampler.

I do not agree about what you tell for live operation, because i've used the sts 5000 some year ago for my live sets and it was perfect and in sync.

I think your test is not reliable and there are many issues that interfere with the latency (roland hardware audio to cv latency, terratec latency).

On 2006-04-08 09:59, lagoausente wrote:
... I agree, that 8 ms, is neglectable playing a keyboard, but also think, is not so neglectable when playing drums live...

it IS neglectible even in this context because it is constant - either the player adopts automatically (Dark, does it work that way ? ) or you delay the mix/monitor for him appropriately.
As mentioned, usually there's more to time adjust anyway ...

regarding the advertisement I agree with you about the pure facts.
But who's relying on 'pure' facts today ?
Even more in business, advertisement, soccer...
the world is full of liers and if you look closely you'll even find them in monastries

don't get me started on that one or I'll write you a page of facts within 10 minutes that you either ignored, overlooked or tolerated - and in some cases even took advantage of, a couple of times in your life

cheers, Tom

I believe I read somewhere that the common "latency" for an analog guitar-amp-speaker setup is around 6-7ms a couple of meters away.. :/

Of course, this is related to how sound travels. DSP latency will be added to that.

live sound in the air is delayed 1ms per foot.

not really relevant, but....

for all intents and purposes, scope is "zero latency". ad/da and ANY calculation even in a hardware device MUST include some latency....unless it's a time machine.

<font size=-1>[ This Message was edited by: garyb on 2006-04-08 10:50 ]</font>

Well, I have been playing sometime, and relaxing. Really is not a bad sampler. For me has the advantage of no drops at all, giving the same performance as the VSi, since 2 ms is no real improvement.
But, have to say, many people here seems like religion fans. I just was measuring "real" and discussing about that.
So really, speech of phylosophy, loves and hates, have nothing to do with the thread.
Some of you don´t like disturb, and don´t fancy reading the hole test, but fancy to speak agaist of in favor of, withoug having understanding nothig, what some of you have admited.
Just thanks to the ones who have readed and underdanting what I´m speaking about.
I think is a good performance, at least, but only have the doubt on the air.
Why the hell, have the same latency routing the midi from sequencer to STS, than routing directly from in to STS.
And just one thing, measuring the Modular II, had 3.5 ms o using a analog drum patch, but had 25 ms, using a synth patch, (I don´t remember this moment the name). So..., my great conclusion is, much more important to believe that to know. If you believe, all works, because you believe.
So I´ll try to start sesions of faith added to technology data.

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-08 14:52 ]</font>

On 2006-04-08 14:51, lagoausente wrote:
...But, have to say, many people here seems like religion fans. ...

dunno if I should feel adressed, but since it's mentioned from time to time I'll take the opportunity...

I'm not being paid by CWA and money is the only thing that makes me defend a position or convince people - honestly.
Otherwise I really don't care.

You will hardly find a statement by me that doesn't come with a reason for my opinion - or the mentioning that it's my personal speculation
A fan would simply applaud to everything...

So really, speech of phylosophy, loves and hates, have nothing to do with the thread.

it was you, who introduced the aspect of lies in advertisement

Some of you don´t like disturb, and don´t fancy reading the hole test, but fancy to speak agaist of in favor of...

any serious test is a serious effort - I simply assumed you were right and put your numbers in some real world context, what's so bad about that ?

Why the hell, have the same latency routing the midi from sequencer to STS, than routing directly from in to STS.

yes, why ?
why would you complain about something that makes things easier ?
whatever they did to the midi - wouldn't it be nice if both pathes have an identical timing ? (it does in fact smell like delay compensation...)

After all it's software, there is no such thing as a midi cable to a DSP pin
the card does not have any routing at all when the PC is powered up, otherwise it could be used without the SFP software.
sidenote: yes, IF 'freezing' a certain setup on the card would work it would be a great feature - and it would help tremendously to transfer devices from card to card...

And just one thing, measuring the Modular II, had 3.5 ms o using a analog drum patch, but had 25 ms, using a synth patch

you can make your modular latency anything you like, depending on what modules are used - insert a delay and it can be 500ms or more
[sorry, forgot to mention] I'm in no way against you personally, your findings or whatever - I've just added my own thoughts to it

cheers, Tom

<font size=-1>[ This Message was edited by: astroman on 2006-04-09 02:57 ]</font>

yes. is there a problem? does it need to be solved?

Astroman! I was not refering to you at all when talked about people who speak without reading first.
I liked your opinions since it had something to add to the post. I just don´t like other post that only makes longer the thread, without adding nothing positive.
And about the issue, the midi in port, would be funny to know, if there is a physical imposibility of the midi going directly to the dsp, or is a driver building decision.
I really think second. I think that audio really goes directly to dsp when use effects, and seems extrange that midi data couldn´t be plysically routed the same way. I suspect that the driver developers could make it.
It´s sofware, but what the computer does is control the parameters. They are just chips, and can use at many ways.
Of course it would need a Creamware engineer to answer how really works, and for me would be funny to know about. But think also it will no easy to find the engineer who want to explain it, would be great even nothing would change, knowing is always a good think.

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-09 06:21 ]</font>

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-09 06:23 ]</font>

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-09 06:24 ]</font>

<font size=-1>[ This Message was edited by: lagoausente on 2006-04-09 06:25 ]</font>

Man, the way you obsess about these things, I'd say you need a vacation =P