Hello,

I'm sorry for this beginner question, but I'm looking to buy a new drive to use Synthogy Ivory and need a bit of advice. Since my budget is strained, I'd have to install Windows 7 + Ivory + a few video games on an SSD. I'm aware that this isn't the ideal solution, but do you think I will run into problems using this method? What I want to avoid are sound problems like crackling noises and dropouts due to a "slow disk".

Or would I be better off installing just Ivory on an HDD? My main focus is on playing the piano, a faster computer start-up etc. are secondary. My computer's specs: Intel 2500k, 8 GB RAM, Asus Xonar DX. Any help would be appreciated.

Thanks in advance,
Patrick

How about one SSD for the OS, apps, and games, and another SSD for Ivory and any other music software you purchase in the future?

Cheers,
James
x

With that system you will not have any problems with Ivory regardless of whether it's on an SSD or HDD. It has streaming tech that stores the first part of each note in memory (a few hundred megabytes, most likely, and probably a tunable parameter). It fetches the tail end of the note while you are holding the key down. It always has it by the time it needs it.

I don't own Ivory but I run Galaxy on an old core 2 duo, underclocked, with a 5400 rpm hard drive and 2 GB ram and I have no issues whatsoever.

Stick Ivory wherever you have room. It will be fine. Problems like crackling are typically due to bad audio drivers these days, not slow hard drives.

I did a double take when I saw your question because I just upgraded to an SSD specifically for the Ivory II American Concert D last week. I see you are PC user - and as a Mac user I make a point of not advising PC owners on hardware setup issues - because I don't know which information that applies to Macs will or won't also apply to PCs due to hardware, OS, and application software differences. So ordinarily I wouldn't respond to your question. But since I just went through HDD vs SSD comparisons on the Ivory II last week I will share my observations as they apply to Macs.

i.e. You may want to ignore the rest of my comments - they may not apply at all to your situation. But I will post these comments because a Mac user will someday have your same question and find the title of this thread in a search.

So with that HUGE caveat, here is the situation I found on a Mac.

I previously ran Ivory II GP's standalone with a 128 sample buffer (44.1 kHz) and a 96 voice limit set on a 3.06 GHz Core 2 Duo iMac with the samples on a 7200 rpm Firewire 800 external disc drive. I never had "Slow Disk" messages, or the accompanying dropouts. When I added the Ivory II American Concert D, which includes a new version standalone app and AU Plug-ins, I started getting "Slow Disk" messages when the number of actual voices reached around 80 in number. It is very easy to generate 80 or more voices with this program. The samples are longer (5-8 secs in the program readout depending on the notes - the undamped top strings are the longest) on the American D than the previous Ivory II pianos, so more notes overlap even when released and there may be other differences in the newer version app as well. Bottom line, I had to increase my buffer to 256 samples (or reduce the number of voices) to eliminate the "Slow Disk" messages and accompanying audio artifacts, neither of which I wanted to do. According to my Mac disk activity monitor the Firewire 800 HD was hitting about 30 MB/s (with 80 voices) when the "Slow Disk" messages appeared, but the maximum transfer rate of the Firewire 800 HD measured well over 80 MB/s (which is expected from FW 800 transfers). So I concluded the problem was not the FW800 transfer limit, but instead the random access time of the mechanical hard drive.

So I purchased a 256 GB SSD with an external FW800 (and USB 3) interface since the random access time of the SSD is negligible, approximately 0.1 mS. Once I moved the samples to the SSD it solved the issue entirely. I was able to run up over 250 voices according to the Ivory II readout (using the sustain pedal and pressing an unrealistic number of keys) before getting the "Slow Disk" message and artifacts using a 128 sample buffer. I bought the SSD with a USB 3 interface (which is about 5 times faster than FW800) - just in case the SSD limited by FW800 didn't solve the problem.

For my normal playing with the SSD I set the audio buffer to 64 samples (below that I could generate occasional ticks and pops from maxing out the CPU) and the maximum voices at 128. At 80-100 voices, which my normal playing generated, the transfer rate measured a maximum around 60 MB/s with no "Slow Disk" messages. The SSD still measured over 80 MB/s max due to Firewire 800, but Ivory II simply didn't need to go faster with my playing.

So I want to stress the original limitation was due to the random access time of the 7200 rpm disc (5400 rpm discs have even slower access times). That had limited the number of voices I could get with a 128 sample buffer and forced me to a 256 sample buffer. With the faster random access SSD, still using the same FW800 interface, I could reduce the buffer to 64 samples and increase the voices to 128 (that was an arbitrary choice over the max voices I needed and I'm not sure of the max number of voices I could have used with the 64 sample buffer, but I got over 250 voices with a 128 sample buffer).

----

The other observation I will make is that the Galaxy Vintage D - Kontakt program streams samples at a MUCH lower rate from the disc than Ivory II. It never exceeded about 7 MB/s from the disc in my playing when using 256 voices and a 128 sample buffer. (Voices are not counted the same way by the Ivory II program and Kontakt - Kontakt counts release samples and Ivory doesn't for instance.) The Vintage D samples are shorter and I think they may also be compressed. In any event, I never had any slow disc issues with the Vintage D or Alicia's Keys (also Kontakt) using the mechanical HD.

So hopefully this information will be useful to a Mac user considering an SSD, and the big point I want to make is that the random access time of the mechanical drive was the limiting factor and not a FW800 interface (on the other hand, a USB 2 interface would have been limiting for Ivory II - probably not for the Vintage D). Finally, I again stress this information is my experience with a Mac, and may not apply at all to a PC user because the hardware, OS, and virtual piano software is different.

------

BTW, I'm now happier with the American Concert D (I wasn't happy with increasing the latency or reducing voices before the SSD) and will reassess it vs the Vintage D (which I had previously favored).

Macy: Thanks for that detailed writeup. Just btw, does Ivory have any settings for the disk streaming? (i.e like Kontakt's pre-load size) I did some testing of Kontakt with different pre-load sizes, and it can affect polyphony. However, I was doing artificial performance testing - not playing normally.

Greg. (on Windows)

Wow, I did not realize there was such a difference in the effective requirements of Ivory vs Galaxy. Somehow I had the impression that they were on par. Thanks Macy.

Galaxy pianos are at 4 to 6 GB per piano (Vienna, Vintage D, and Steinway).

Ivory's samples were quite large in v1.5 ... over 60 GB for four pianos.
It seems the Ivory v2 samples are larger still.

I tried putting my Ivory II program on a 128 gig ssd...it worked great for about a month then my computer couldn't find the ssd any more, It was like it was completely dead, so I went back to putting my Ivory II program on a standard but seperate hard drive, which in my opinion is much more reliable, and it's working like a charm....

I have East West's Symphonic Orchestra on a different ssd and that's working fine.

So I'd say ssd's are a gamble.

Ivory II has a Small, Medium, Large setting for the RAM sample buffer (this is for the digital audio file samples not the audio sample buffer, i.e. it has no effect on latency) but the sizes are not specified anywhere that I can find. That's it.

On the Vintage D-Kontakt I use the maximum preload size, which is about 240 KB per sample, which results in a total preload of 720 MB. That's 3000 samples (720 MB/240 kB). (Alicia's Keys preloads about 1.26 GB of samples, which is almost twice as many samples as the Vintage D.) If you divide 3000 samples by 88 notes, that's about 34 samples per note. The Vintage D has up to 13 layers and it has soft pedal samples, sustain resonance samples and also release samples. But the total size of the samples stored on the disc is only 4.23 GB (not counting the Pads which are 1.12 GB). On the other hand, the Ivory II American Concert D has almost 49 GB of samples on the disc and doesn't use separate sustain resonance samples.

Yeah, about 49 GB alone for the American Concert D.

For Vintage D, I load the default of 197 MB.
Why do you load so much more (720 MB)?

There have been reports of reliability problems with some SSDs, particularly early models, and also there were a few wacky firmware problems in some models. I use Samsung SSDs on my computers and they have had a very good reliability track record. I used a new Samsung 840 256 GB SSD, just released, for the piano samples. It is a different technology (TLC, rather than MLC) than the previous Samsung 830's I use, so I suppose it is a bit of a gamble. But when it comes to technology I like living on the edge.

The more you load on memory, the less it has to stream from the HD. You can find a good compromise depending on the amount of RAM and speed of your HD.

For what is worth, I keep it at default and it loads about 257MB (when most of the 'extras' are turned on like repedaling, release and noises are loaded).

Macy's approach may work much nicer for systems with lots of available RAM. How much do you have Macy?

Rafa.

It reduces latency requirements when playing notes the first time and I have a lot of RAM, so why not?

I'm running 8 GB of RAM. The iMac is 2 years old and soon to be retired so I haven't added more RAM to it. I'll go to 16 GB in its replacement.

Even with 8 GB I have no trouble running several Ivory II pianos and Vintage D all at the same time in MainStage so I can instantly switch pianos, even without cutting off notes from the previous piano. But 16 GB will allow even more pianos to be loaded simultaneously.

Mine's at 830MB with the effects turned on


This ^^
With the price on RAM so ridiculously low nowadays, it has pretty much become a standard for your average desktop to have a 6-8 gigs of memory. That's enough to fit an entire library and then some, if only software would allow it

I also like living on the edge until I got burned with this one SSD. Now I'm back to a standard hard drive and happy as a clam with it's performance.

Thanks. Did you try changing the ram sample buffer in Ivory then? Mind you, if you were getting 30MB/s, that's already pretty good for a HDD (for the random nature of sample streaming) IMHO.

Given that you are using the maximum pre-load in Kontakt, that would of course help to reduce the data rate from the disk, because the larger the pre-load, the less likely it is that the disk is required, when playing in in a normal fashion. (i.e, not holding down lots of notes for a very long time) If the time-per-sample stored in memory equated to much larger time than Ivory, that might partly explain the large discrepancy in the data rates. I'm pretty sure you're right that Vintage D uses compressed samples though. (encrypted and compressed, I think, but I'm not 100% sure)

Greg.

I've always used the Large RAM buffer in Ivory. The smaller settings make the CPU work harder and may lead to needing a larger audio buffer and more latency.



What you say is true, but the pre-load size (even at maximum) produces very, very, short bursts of audio. You can listen to the pre-load length in Kontakt by disabling the samples library after the pre-load completes at startup. Then just play the notes. (With an external drive you can simply unmount it, disconnect it, or shut off its power. I suspect you could also just temporarily rename the sample library folder on an internal drive too, but I never tried that.) I haven't tried that with Ivory II. It's something I'll have to do.

Yeah, I'm not certain that the Vintage D samples are compressed, but I've been told that.

@Macy: On Windows, I found that the maximum pre-load in Kontakt = 480kBytes per sample per channel. I.e, a stereo sample would have 480*2 = 960kB stored in memory. For a 44.1kHz stereo 24-bit sample (not compressed), this is 3.7 seconds. I found that Kontakt waited quite fair percentage of this 3.7 seconds before accessing the disk. (obviously it can't wait until the very end, otherwise an underrun would occur. I was actually surprised that it waited as long as it did) For normal playing, a 2 or 3 seconds is quite a long time IMHO. Time permitting I'll repeat the test.

Greg.

Sorry, I made a mistake describing what happens on the Mac in Kontakt when the samples are unmounted after the pre-load. I was writing from my memory of hearing only short sounds and hadn't tried it in a while. I just went and tried it again. On the Mac if I unmount the external disk containing the samples after the preload, I only get a short pop when a key is pressed after the FIRST key press. It appears that Kontakt realizes that the sample library is missing after that first key press, and then it no longer plays the preloaded samples (or attempts to mute them with just a slight pop sound). However, each time you release a key, you get a short note which may be the release sample for each key. i.e. you get a pop when the key is pressed and then regardless of how long you hold the key you apparently get a portion of the release sample when you release the key. That must have been the short sounds I remembered earlier. Apparently this works differently on the PC version, but that is what happens on my Mac (I tried it several times in a couple different ways just now). Sorry for the earlier confusion.

Interestingly, Ivory II works exactly the opposite, and more like what you described for Kontakt on the PC. I just tried it (I hadn't tried it before). On Ivory II it indeed plays the preloaded sample portion for each key pressed after the disk containing the sample library is removed, but it doesn't play any portion of the release samples for each note when a key is released. Perhaps the release samples are not preloaded at all, or more likely after it can't load the remaining part of each sample it doesn't try to play the release sample. The latter seems likely because each key press produces a separate Error message in the Ivory II display.

Although I concede that a few seconds won't change the maximum - i.e - it won't affect the demand on the disk when playing big, long chords with the sustain pedal. I suppose it would mainly affect the average data rate, which isn't very interesting. Since you were using the maximum memory in Ivory, you were doing a fair comparison with Kontakt, using it's maximum pre-load.

The memory size can change the maximum throughput that a disk can support, though. At least, that's what I have observed in Kontakt. The reason is simple: as the pre-load size is increased, the amount of data that is read from each sample each time a sample is accessed also increases, and that makes the accesses more efficient. (because the disk is spending more time reading sequentially, which it can do very well, and less time moving the heads, which is very slow) However, even if Ivory is the same as Kontakt in this regard, since you were already using the maximum in Ivory, there was nothing more you could do to improve it.

[technical side note]
As I've said elsewhere, when doing repeat testing, it is important to understand that there is secondary caching occurring by Windows, over and above the sample pre-load. So, if you do a test after rebooting your machine, and then repeat the test using the same performance (or MIDI file), you may get a vastly better result, simply because the system performed extra caching of the samples that were used in your performance. Unfortunately reloading the instrument or even restarting Kontakt does NOT clear this cache, at least in Windows. So, in Windows, to avoid rebooting the machine every time, the utility
RamMap (command: Empty | Empty Standby List) can be used. Unload the instrument, run the utility, and then re-load the instrument. This procedure will allow different settings to be tried without rebooting the machine inbetween each test.

Greg.

I just wanted to thank you for the answers, especially Macy - wow! I'm very glad you decided to write all this down, and I'm sure anyone searching for this topic in the future will be too. Kawai James, I hadn't considered that since I've heard smaller SSD's tend to be slower than the bigger models, but I think I will take your advice and go with two SSD's.

I don't want to disrupt the discussion any further, it's very interesting to read what's going in detail. I've certainly learned a thing or two - I appreciate it.

Thank you all,
Patrick

Yeah, thanks for this thread folks...

Let me just mention, I too had the problem of pops and crackles with Ivory II, when pressing down on the sustain pedal...

I've managed to fix it (well, kind of), by using Elgato's Thunderbolt SSD 240GB.

Having said that, with the benefit of hindsight, and having watched reviews by other users/commentators (particularly on Youtube), I think it may be more advisable to go with e.g. a Samsung 840 series SSD, perhaps 500GB, with a Thunderbolt adapter, e.g. the one by Seagate. Overall, this configuration apparently performs vastly better than the Elgato Thunderbolt SSD product. I mean it apparently achieves read/write speeds in the region of 300MB/s; whereas the Elgato SSD achieves something like 100MB/s write and 220MB/s read.

To be fair, the Elgato SSD works just fine for my requirements, it's just that I've realised, in hindsight, I could have gotten a more powerful set-up for the same price. I mean, the Elgato cost me approx £400; had I gone for a Thunderbolt adapter (£100), Samsung SSD 500GB (£250) and a separate Thunderbolt cable (£50) the overall cost would have been roughly the same, with more memory and faster performance (apparently).

I would also mention that, when "testing", the Elgato performs quite comfortably with a 64 voice buffer (i.e., rapidly and repeatedly tapping on an unrealistic number of keys), with no pops/crackles/dropouts; I think perhaps under more "normal" playing conditions, I could probably manage quite easily with a 32 sample buffer (I mean, at the moment I'm trying to learn relatively simple stuff, by e.g. Einaudi).

So overall, yes, using an SSD for memory, with Thunderbolt for I/O, has, for all practical purposes, fixed the problem for me, although of course at a fairly hefty price.

pv88, are you related to my mate Richard?

Sorry James, just curious, was this a PW defect again dropping your post in a wrong thread, wasn't it - I can't see a post of PV88 anywhere to realate to on this topic?

Yes, it appears that pv88 deleted his post.

He originally asked if the user above, Vijay '82, was related to golfer Vijay Singh.

I am holding out on moving my Ivory II pianos to my macbook until Thunderbolt external drives are available at a more reasonable price.

Macy- you are using firewire for your external SSD on your imac? (by the way- you consider a two year old computer to be "retirable..?" man you like to spend money on tech hardware!) Macbooks won't support USB 3.0, right?

What's wrong with simply getting an internal SSD with enough space?

Unless you plan on installing all of the pianos a 256GB or even 128GB SSD should be fine.

i assume you are addressing my inquiry... i use the macbook for a lot of other stuff and i'm a believer in segregating the samples on a separate drive from my OS files etc. maybe that's not necessary with an SSD- but 128GB/256GB would be pushing it for me as far as not enough capacity. perhaps if i used it only for pianos (a future possibility- but i think i'd get an ultrabook for that purpose)- as i currently do with an older desktop - that would be a good strategy.

thanks..