• Register
  • Help
Results 1 to 9 of 9

Topic: Stupid questions about Mixtreme and Audio in general

Share/Bookmark
  1. #1

    Stupid questions about Mixtreme and Audio in general

    Please, can somebody tell me something about this card so that I could actually understand the terms...

    The card has no analog in/outs so you have to buy $600 rack unit just to get sounds out of it (...to plug RCAs to my amplifier)?

    You can\'t get any sounds off it in normal Win9x environment (no mp3 playing and your normal audio editing session would be without sound)?

    I need over $1000 to use one of these?

    Is these any other soundcards which could do Directx plugins in realtime while I play with GS? I mean, is there any cheaper one [ahem]?

    Could somebody please point me a way to read info about DIGITAL in/out ports, TDIF etc or shortly explain the basics of them?

    Thanks.

  2. #2

    Re: Stupid questions about Mixtreme and Audio in general

    hi sami,

    you are absolutely right - the mixtreame card alone is cheap but only provides TDIF in/out, just usable for tascam users or with some digital desks. I you want to know why the card doen´t use ADAT ports: Because they sell the TDIF-ADAT convertors for the same price as the whole card! Additional you have to pay extra for a AES/spdif port. The TC plugins are expensive, too, and when you habe spent over 2500$ you have only 1 DSP, which calculates 1 reverb and a simple 16 channel mixer. And you will need a direct X catd as SB live additional e.g. to run reactor or so.

    Better wait for the pulsar/scope integration, they told me August the work is done. For almost the same money you get 5 times of DSP power, driverst from GS over ASIO to dirX and MM, multiclient, free configuration and lots of plugins and synths. I guess they are ready with the integration when GS 1.6 is shipping!

  3. #3

    Re: Stupid questions about Mixtreme and Audio in general

    Can you tell me more about the Pulsar/Scope integration? How much it costs and is there a web page for specs and more info?

  4. #4

    Re: Stupid questions about Mixtreme and Audio in general

    Gigasampler drivers are ready end of August, they told me. Information:

    Manufacturer: www.creamware.de or *.com
    Users site: www.planetz.com/pulsar
    High end Plugins: www.sonictimeworks.com


  5. #5

    Re: Stupid questions about Mixtreme and Audio in general

    Hi Holger and Sami,

    Just to put you straight on the Mixtreme pricing and features as there seems to be more than a little confusion.

    Mixtreme is US$549 list (16 channel digital)
    Mixtreme & SS8IO-3 (8 channel analogue I/O) bundle is US$999 list.
    Mixtreme & SS8IO-2 (TDIF-ADAT) bundle is US$749 list (ie. US$200 for SS8IO-2).

    TC Reverb is US$599 list

    Check with retailers for deals etc.

    At US$1598 list you get a top quality REAL-TIME reverb from TC\'s M5000 (not to be confused with the cut down TC Native reverb), 8 ins and outs (approx 100dB S/N ratio) and an additional TDIF port (that you can add more analogue I/O\'s to later).

    The Mixtreme Mixer is not a simple one, you have 16 digital busses, Surround mixing and panning, as many pre or post sends & fully parametric EQs as you need. There are 128 positions for mono/stereo or surround mixer strips and multiple cards can be used.

    Mixtreme is also multi-client, so you can use the GSIF, ASIO and/or MME (working with Reality, Rebirth etc.) simultaneously. The drivers are low latency (9mS with the GSIF) and below 25mS with ASIO (VST\'s limitation). MME drivers go down to under 10mS too, but as the application sets the buffer size, most applications don\'t have latency this low due to Windows buffering, task switching and user level programming.

    This low latency with GSIF and ASIO, plus the real-tim Mixer means that you can easily record the GS output into Cubase VST or any other application via the Mixtreme mixer.

    DirectX effects will be of no use with Gigasampler as the latency for DirectX would mean that the samples appear several tens of milliseconds late (could be up to several hundreds of milliseconds too). This would make the Gigapiano completely unusable. With the TC Reverb, it is effectively real-time, with similar negligible latency to an outboard digital effects unit - plus the TC Reverb sound is fantastic, as for their high end units.

    Mixtreme has one Motorola 56301 DSP running at 80MHz with 240Mips peak and on-board memory for the reverb and other plug-ins. By comparison a Pulsar has 4 Sharc DSPs at 60MHz, peaking at 480Mflops, but has not enough on-board memory for reverb, so has to use the slower PC memory for this. This results in a much less efficient use of the DSPs and we estimate an actual 25% - 30% more usable DSP performance than for a Mixtreme, but Mixtreme is less than half the price of a Pulsar. The latency delivered by the Pulsar\'s drivers so far is way higher than for Mixtreme\'s ASIO and MME drivers and there\'s no indication that this would be better for the GSIF or for Scope.

    Pulsar requires an enormous amount of PC resources (256MB RAM, CPU cycles, PCI Bus activity etc.) that means that you need a monster PC just to run the GUI for Pulsar. Mixtreme on the other hand requires very little PC resources (and even the mixer metering updates at background level) to allow the other applications the maximum available CPU activity, so that updating the PC is necessary for Mixtreme.

    I hope this clears up any confusion.

    Patrick Van Dyke
    Soundscape Development Team

  6. #6

    Re: Stupid questions about Mixtreme and Audio in general

    The gentleman from Soundscape has stated his case well. I have a feeling that all he claims about current latency and the MFLPS comparison is correct. However, even if the pulsar onboard memory is not enough (I would like to see a cut sheet comparison), Dram is getting pretty cheap and runs on a 100 mhz bus while the pulsar chips runs at 60 mhz. How can the computer\'s memory be slower? Please educate me!

  7. #7

    Re: Stupid questions about Mixtreme and Audio in general

    hi patrick & bailey,

    of course onboard memory is always faster because the acess to PC´s memory is only possible ocer the \"slow\" 33 MHz PCI bus and the main problem here is the timing. E.G. a bad graphics card or other DMA drivers can slow down or delay PCI performance for other tasks like pulsar or mixtream. Due to a nice price creamware only gives minimum of RAM on their pulsars, the much more expensive scope has 32 Mb on board which allows much more delay-based plugins (reverbs). I found a pulsar + srb system (8 DSP) in the current version is limited to 3-4 4080L reverb plugins, but you can still have the 32 chan mixer with 128 EQ bands and many plugs for dynamics etc or synths. So a reverb plugin uses approx. 50% DSP, but also 25 % of RAM resources.

    But be careful with the spec comparisons: The sharks calculate audio native in 32 bit floating point, which gives you a lot of dynamic headroom for high end signal processing. Especially when I do scoring, in many passages I input mp or pp sounds on approx. -20 db and find that the floating point algorithms working smooth and very analogue-like

    Motorola calculates 24 bit linear, so quantizing effects and internal distortion/overflows might be a problem, especially when I stream in 24 bit audio there is no headroom left. Then the internal resolution has to be increased to 48 bit linear, which cuts down the DSP performance to 50%. 48 Bit, am I right? And 48 bit is always recommended for professional sound processing, so you see all pro tools plugs use this - and therefore one of the expenive Mix Plus Farms in the end only performs a handful of plugs in real life...

    patrick, can you tell us a working example of a real life setup, e.g. 8 channel stereo mixer, TC verb, 6 stereo aux sends (1 into the TC), 8 out busses - how many DSP power then is left for EQ´s and dynamics? Are the Plugs you used to 24 or 48 bit internal resolution?

    thanx!


  8. #8

    Re: Stupid questions about Mixtreme and Audio in general

    Hi Holger,

    Ahh.. the never ending floating point vs fixed point argument. Here\'s a few facts :-

    1. No-one ever accused Lexicon or TC 24 bit fixed point reverbs for their studio boxes of a lack of smoothnesss or detail.

    2. Internally the Motorola DSP has 56-bit accumulators, 48-bit for full 24-bit multiplies and 8-bit headroom. Carefully designed fixed-point algorithms can be as accurate as floating point without a speed penalty.

    3. 32-bit floating point has only a 24-bit mantissa. Multiplying keeps the full resolution but addition reduces the resolution to the largest of the 2 numbers. So even with floating point you have to be carefull not to lose precious precision.

    4. Floating point also comes with the disadvantage of a \'moving\' noise floor.

    5. The moment that the DSP does not have direct access to the memory with \'zero wait state\', it is incurring delays and ends up kicking it\'s heels waiting for the data. This can occur even with on-board memory that is shared by several devices if two try to access the memory at the same time, something that is difficult to prevent, especially with a plug-in sofware architecture that means that the developers of the board do not know what algorithms the DSPs are running at any time. This could account for the well known problems that Pulsar has, in reliably loading a saved project where it cannot seem to ensure the same DSP load as when the project was created.

    Accessing memory across the PCI bus is *very, very, very* inefficient in DSP terms and is used for \'low end\' cards like the SBLive as a way to save money on memory. PC memory is cheap, but DSPs are not. Sacrificing DSP performance as a way to reduce the cost of a board is a dubious calculation, you may just end up paying for DSPs that are used so inefficiently that it would have been better to have fewer DSPs and put the money towards on-board memory. This is the approach that Mixtreme takes.

    As with all these DSP issues, it comes down to the actual implementation of the algorithm in determining the errors, headroom and DSP efficiency actually achieved. To a large extent this also depends upon the skill and knowledge of the actual developer and the time he is able to spend upon developing any particular function, which in turn affects the cost of the product. Hence a poor sounding reverb for a sequencer that you get for free is extremely unlikely to sound as good as one from a company that has invested heavily in the time and people to develope their algorithms and implement them efficiently.

    Regarding the DSP percentages, to hear are a few figures.

    16 Channel (8 x Stereo) Mixer with Track Insert 2 B EQ & Send per channel, TC Reverb & Master Fader
    P = 83.0%, M = 61.2%

    As you can configure exactly what is active in each channel strip (and whether it takes DSP activity at all), you can dynamically allocate the DSP power for each mixer used for each project. (ie. no send used...use the DSP for something else).

    To follow are the figures for the every element in Mixtreme - you can use these as a guide for creating mixers. Sorry for the length, but you asked....

    Patrick Van Dyke
    Soundscape Development Team


    Mixtreme Performance
    ---------------------------------

    Worst Case DSP Power for Mixing and Effects (P = DSP Processing Cycles, M = DSP Memory)
    Note : this does not require PC CPU activity (all updating is at background level).

    Tested at 44.1kHz with Logic Audio 3.6 runnning 24 tracks mixed to the 16 track inserts in the Mixtreme mixer using the MME drivers.
    Also tested VST24 V3.6 and ASIO 24 bit drivers Latency set to 28mS - 24 tracks again).
    PC was PII 233MHz 64MB Ram, IDE HD (bus mastering drivers)

    16 channels audio streaming P = 1012%, M = 0% (not required when just using the mixing capabilities)
    2 Band Parametric EQ (mono) P = 1.9%, M = 1.2%
    2 Band Parametric EQ (Stereo) P = 3.7%, M = 2.2%
    Mono Dynamics (from Audio Toolbox) P = 2.3%, M = 1.1% (Gate/Compression/Expander/Limiter)
    Stereo Dynamics (from Audio Toolbox) P = 2.8%, M = 1.4% (Gate/Compression/Expander/Limiter)

    TC Dynamizer (from Finalyzer) P = 40.4%, M = 22.6%
    (3 band compressor/expander/limiter/soft clipping/look ahead peak detection/DC Removal filter)
    TC Reverb (from M5000) P = 29.1%, M = 42.0%

    Stereo Chorus/Flanger (from Audio Toolbox) P = 7.6%, M = 3.7%
    2 Tap Mono Delay (Short) (from ATB) P = 2.2%, M = 9.7%
    2 Tap Mono Delay (Medium) (from ATB) P = 2.2%, M = 19.5%
    2 Tap Mono Delay (Long) (from ATB) P = 4.6%, M = 39.0%

    2 Tap Stereo Delay (Shortl) (from ATB) P = 4.6%, M = 19.5%
    2 Tap Stereo Delay (Medium) (from ATB) P = 4.6%, M = 39.0%
    2 Tap Stereo Delay (Long) (from ATB) P = 4.6%, M = 78.0%

    Wave Mechanics Reverb (ex Eventide) P = 42.7%, M = 56.4%

    32 Channel Mixer with 2 Band Parametric EQ per channel, 16 external inputs mixed with 16 channels from Logic/VST
    P = 89.1%, M = 71.6%

    Mono in Mono Out Channel Strip P = 0.7%, M = 1.4% (with /fader/input mute/output mute/solo)
    16 x Mono in Mono Out Channel Strip P = 11.9%, M = 23.3% (with /fader/input mute/output mute/solo)
    16 Channel (mono) Mixer with 2 Band EQ per Channel P = 43.4%, M = 42.6%
    16 Channel (mono) Mixer with Dynamics per Channel P = 48.6%, M = 40.9%
    16 Channel (mono) Mixer with 2 Band EQ & Dynamics per Channel P = 80.1%, M = 60.2%

    Mono in Stereo Out Channel Strip P = 1.2%, M = 1.1% (with pan/fader/input mute/output mute/solo)
    16 x Mono in Stereo Out Channel Strips P = 16.0%, M = 15.6% (with pan/fader/input mute/output mute/solo)
    16 Channel (mono in stereo out) Mixer with 2 Band EQ per Channel P = 47.5%, M = 34.6%
    16 Channel (mono in stereo out) Mixer with Dynamics per Channel P = 52.6%, M = 32.9%
    16 Channel (mono in stereo out) Mixer with 2 Band EQ & Dynamics per Channel P = 84.2%, M = 52.3%
    16 Channel (mono in stereo out) Mixer with Track Insert, 2B EQ & Dynamics per Channel P = 91.3%, M = 64.8%

    Stereo Channel Strip P = 1.4%, M = 1.4% (with pan/fader/input mute/output mute/solo)
    16 Channel (8 x Stereo) Strips P = 11.0%, M = 11.3% (with pan/fader/input mute/output mute/solo)
    16 Channel (8 x Stereo) Mixer with 2 Band EQ per Channel P= 40.8%, M = 28.7%
    16 Channel (8 x Stereo) Mixer with Dynamics per Channel P= 33.1%, M = 22.1%
    16 Channel (8 x Stereo) Mixer with 2 Band EQ and Dynamics per channel P = 62.9%, M = 39.5%
    16 Channel (8 x Stereo) Mixer with Track Insert, 2 Band EQ & Dynamics per channel P = 69.0%, M = 52.3%
    16 Channel (8 x Stereo) Mixer with 2 Band EQ & Send per channel, TC Reverb & Master Fader P = 76.9%, M = 58.3%
    16 Channel (8 x Stereo) Mixer with Track Insert 2 B EQ & Send per channel, TC Reverb & Master Fader
    P = 83.0%, M = 61.2%

    MS Decoder Mix P = 10.1%, M = 12.8%
    MS Encoder Mix P = 10.2%, M = 14.3%



  9. #9

    Re: Stupid questions about Mixtreme and Audio in general

    Hello all,

    now where is my MXR phaser pedal?


Go Back to forum

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •