It's 2012- Why are we still pushing DSP processors?

[Irish916]

I came back to Pro Tools with installment 9 because Avid opened up the software to be used with just about any interface. One of things that always kept me away from Avid was this obnoxious marriage of the software to DSP processing systems. The software was designed to work in tandem with these hardware elements to maximize the ability to use plug ins. I think years ago, when most systems had single core CPUs and 2GB of RAM, it made sense.

What I don't understand is why Avid is still pushing newer, faster HD DSP processing in an age where most folks are using significantly more powerful computers. Is it really about the money? Every other DAW I've used in the last year can take full advantage of multicore CPUs and a 64 bit computing environment without the need for expensive HD boxes. Why not PT? Sure version 9 and 10 still allow you to use PT without the expensive crutches, but PT still has some significant shortcomings when try to use larger, more memory intensive virtual instruments - EVEN those from the approved list. It's just frustrating to me as a user of so many other DAWs because I want to work in the PT environment for the sake of web collaboration, but every time I do it feels like I'm stepping inside a silver Delorean and heading back to 1994!

[EGS]

Originally Posted by Irish916

...PT still has some significant shortcomings when try to use larger, more memory intensive virtual instruments ...

Agree that this needs to be fixed. As a workaround, many here use Reaper or Vienna Pro as a VI host rewired to Pro-Tools. Lots of pressure on Avid to get-it-all-done-right in version 11. My 2 cents.

[Chris Lambrechts]

Originally Posted by Irish916

What I don't understand is why Avid is still pushing newer, faster HD DSP processing in an age where most folks are using significantly more powerful computers. Is it really about the money? Every other DAW I've used in the last year can take full advantage of multicore CPUs and a 64 bit computing environment without the need for expensive HD boxes.

I've posted something similar in another thread. There is nothing wrong with developing a workflow that allows you to work in a native environment. Yes, CPU's are very capable of doing that for a lot of workflows.

However, no matter how powerful your computer is, there are still very specific reasons and workflow environments that demand DSP based systems.

I use both DSP and Native systems but I can assure you that I will choose my DSP based system over any native system at any given point in time .... for the things I have to do with it that is.

More so ... I would much rather have a slower cpu and run a lot of dedicated DSP then run the fasted CPU available on the market today without dedicated DSP.

Why ?

Well for starters because I don't ever want to worry about typical things inherent to any native system. Low latency monitoring for musicians in the recording room for example.

I never want to worry about having to disable native plugins because I need to do an overdub with the mixing process already well on it's way.

For some this may sound like an unnecessary luxury and I understand the reasons why. But please realize that there is still a very big market that demands dedicated DSP as well.

Yes it is 2012, and you want to know my truly honest opinion about CPU's / OS's and computers in 2012 ? For over 90 % of what we do they have become SLOWER then they were 20 years ago. Back under windows 3.1 ... which was arguably the first user friendly OS on the market there was Microsoft office which had word and excel in it. You would double click the icons for the app and the app would open up in a split second. You could not blink your eye or your app was ready for use.
You tell me how long it takes to open up on your system ?

Sure ... it looks fancier but at the end of the day ... what do you use it for ?

I know ... a lot of added features for sure but overall slowness at the cost of making things look graphically more appealing as opposed to creating a faster and more secure rock solid environment.

Oh and talk about rock solid .... I couldn't even make that 386 DX computer crash if I wanted to ....

There's a reason why slower Xeon processors are being used in server applications as opposed to much higher clocked i7's. The main reason is reliability. You may wonder why people spend fortunes on those processors ? well ... because when properly installed and configured you can run them for ages without requiring a restart. It wouldn't even make a difference.

How many native users can look me in the eye and say that a restart doesn;t make a difference after 5 or 6 hours of work. You can't because clearing up all that memory cache from time to time feels good.

Well ... Think in that line with DSP based systems. I can run mine for ages and it won't make a difference wether I restart it or not. It will just as happily purr along.

Yes, it is 2012 and yes for many things native can stand next to DSP. And that was still very different 5 years ago. It's a good thing. But there still is a difference ... at least in my opinion there is.

Chris

[lennieh]

... because for some users it's not enough to have a really fast cpu. Post-Production guys are running with hundreds of tracks, tons of plugs, too much for a purely native rig....

For music a native rig is fine now unless you need hundreds and hundreds of heavily processed tracks....

So they are selling HDX to three groups of people:

1. Those that really need it.

2. Those that have to have the biggest and best whether they need it or not.

3. Those that don't do the maths and think that because they currently have a TDM system they need to upgrade to HDX, when actually all they need is HD|Native and a modern PC/Mac.

[Irish916]

Originally Posted by Chris Lambrechts

However, no matter how powerful your computer is, there are still very specific reasons and workflow environments that demand DSP based systems.

I use both DSP and Native systems but I can assure you that I will choose my DSP based system over any native system at any given point in time .... for the things I have to do with it that is.

I appreciate your insights on this Chris. Good stuff. I wouldn't argue that DSP based systems work better when compared to their native counterparts. I would assume a PT10 with HD DSP works much better than PT10 Native. My opinion is that this is because of the way PT10 is designed. It's designed in such a way that you really need DSP to maximize your workflow and do more with the product. By contrast there are many other DAWs out on the market today that don't offer a DSP option and don't have issues with latency, large file size or memory issues. If the consumer has a marginally capable system and interface, they have all they need. So by the very nature of the PT design, a design many folks would say is significantly antiquated, the consumers are "nudged" to open their wallets for the more powerful and expensive DSP systems.

Seems to me that you could rewrite the base code for PT to take advantage of multicore CPU by reserving cores JUST for DSP processing. Realize if you did this, you wouldn't make millions on DSP boxes or expanded HD licenses on the DAW software. I think the reality is that is something Avid isn't willing to give up.

[Chris Lambrechts]

Originally Posted by lennieh

For music a native rig is fine now unless you need hundreds and hundreds of heavily processed tracks....

see that's where I would like to add a couple remarks.

First off, I hardly ever do post and when I do it is for commercials and not film mixing with the hundreds of tracks that you are referring to. Those guys more often then not will even run multi-user / multi-systems linked together.

I primarily do music. Recording live musicians and my typical session track count will be anything from maybe 16 tracks to say 80 or so. So not that high track counts in this day and age.

Then there's native and native. Lets split it into the 'normal' native and HD native. Both do plugin processing on the CPU (hence Native) but HD native will allow for a DSP based like, no latency I/O workflow where you - at least during tracking where you will likely not use any plugins yet - have up to 64 I/O available at no latency, no matter how many tracks you run.

HD native is definitely something I could live with and yes, a lot of music facilities can opt these days for a HD native system which will provide them with enough I/O to do both tracking and hybrid workflows with hardware inserts. And the plugin processing is done on a powerful cpu and for your average music mix that will suffice quite nicely thank you.

But at the end of the day I will still choose a DSP based HD system. Not because I need the extra I/O it can give me because I don't. But for 2 - for me - very important reasons.

1. Inline workflow with my controller where I can use the Mixer engine of my DSP based system and set it up as a very big digital desk. A console !!

2. Because I want to have the luxury of not having to worry about any processing already in place throughout any part of the process, from recording through editing through mixing up to the very minute before printing my final mix. I want the luxury of not having to analyze my session at that point in order to do a quick last minute overdub.

Does that happen a lot ? well , you'd be surprised. I'm not talking about actually adding a musical part. I'm talking about simple things like a quick re-amping of a guitar solo for example. And the luxury my DSP system gives me at that point that I can leave all my processing in place and hear the result of that re-amp while turning the knobs of that amp. Silly once every 3 moons stuff like that.

Details. Of course they are but I'm also a very firm believer that a good song and a good performance is far more important than the best mic pre and the most expensive mic. I'll record everything with a behringer pre amp and an sm 57 if I have to as long as the performance is provoking something emotional.

traditionally studios are an environment with the primary purpose of creating a comfort zone for the performer. As a musician in an ideal world you never have to worry about a technical aspect that can possibly throw you out of your concentration.

There are a lot of things that can help you with that. A nice room / good acoustics / a nice ambiance / good gear of course and in my opinion a DSP based system as opposed to a native based system.

And then a setup where I have everything at my fingertips which I can possibly offer and need.

5 musicians with each their own mix, sure ... here you go ...
Not something like ok, can we have a 15 minute break while I figure out the extra busses and routing and see if I can still give you the low latency monitoring that you need to be able to play.

In my preferred DSP world : low latency ? huh ... what's that ?

For a lot of people that sortof stuff is a luxury that is not worth the extra investment and believe me, I can fully understand that. Especially the current economic environment in the music industry and luckily there are options available to set up a very professional native setup.

But there is definitely still room for dedicated DSP luxury as well, and luckily as well !!!

Chris

[spearson]

Originally Posted by Chris Lambrechts

see that's where I would like to add a couple remarks.

How many native users can look me in the eye and say that a restart doesn;t make a difference after 5 or 6 hours of work. You can't because clearing up all that memory cache from time to time feels good.

Well ... Think in that line with DSP based systems. I can run mine for ages and it won't make a difference wether I restart it or not. It will just as happily purr along.
Chris

Chris,

I completely get your reasons for running a DSP based system, and if I had an eye to look at, I'd admit that rebooting a computer after 5-6 hours is a good thing anyway. Even humans need a rest after 5-6 hours. But can you honestly admit that your DSP chips on your HD system don't do funky things. Say for example, you drop a Re-vibe plugin on a channel then remove it and the DSP hasn't refreshed itself so the only way to fix it is to quit protools or trick the cards into refreshing by changing your DSP allocation. That happens a lot, you might not notice it because you're only running PT's for music but I have to fix this daily. It's not just my system either, it's common place. Once you start routing 5.1 channels into stereo down mixes X 7, and maxing out your cards, those DSP cards start acting really silly.

[Chris Lambrechts]

Originally Posted by spearson

But can you honestly admit that your DSP chips on your HD system don't do funky things. Say for example, you drop a Re-vibe plugin on a channel then remove it and the DSP hasn't refreshed itself so the only way to fix it is to quit protools or trick the cards into refreshing by changing your DSP allocation. That happens a lot, you might not notice it because you're only running PT's for music but I have to fix this daily. It's not just my system either, it's common place. Once you start routing 5.1 channels into stereo down mixes X 7, and maxing out your cards, those DSP cards start acting really silly.

DSP caching enabled vs DSP caching disabled.
I prefer to leave it disabled most of the time and let the system take it's time when switching from one session to the next. If I do need to go 'faster' and switch a lot between sessions I will enable it but my general rule of thumb is to disable it.

Yes I have seen funky things on DSP systems of course. I installed my personal HDx system only a couple of days ago. Of course it is too early for me to stick my neck out at this point and go yelling from rooftops that it doesn't ever crash. I've only ran one small tracking session on it and some intensive test sessions.

I have to say though ... and this is purely me relying on my years of experience with TDM systems (which go all the way back to the dark and evil days of the Nubus monster ) ... just going by feeling ... well, it's been a while since I sat behind a system that felt this good. There's been a few versions where I really could sit for weeks if not months on end behind my system and never ever worry about a single hickup.

Back in the Mix days under OS 9, probably something like version 5.x or maybe the early 6 versions (those still ran OS9 right ?) ... ROCK solid combo with a Pro Control setup. LOVED it.

And then there was 7.2 (or was it 7.4 ... the one with the big automation overhaul). Prior to Pro Tools 10 THE BEST version EVER !!!!!!

I've said this before. A lot of the stability of course comes from the software. I've been shaking, twisting and turning my previous TDM system up and down and back and forth and PT 10 for me personally was a HUGE deal because of speed / fluidity and general good feeling while using it. It feels way better to me then 9 and that is saying a lot because 9 served it's days being extremely loyal to my needs.

Adding HDx I do not expect to run into a lot of unexpected stuff.

System 5 channelstrip DSP ... yeah baby !!!!!

Chris

[Jeff E]

I think it has to do with our propensity to use what ever is available to us. The more the better in many cases. Its what we do as Americans.
I have a friend who works at a major animation company and they do feature films and are always maxed out with what they do. even when they get the next system that is 5 times as powerful it is quickly maxed out. Its the cycle we are in. Go from HD1 to HD3 and you will run out of DSP and so on and so on. AS long as there is a market, they will make it. Obviously their research shows that there are purchases.

Even after all the complaining that everyone did about the latest round of new stuff, people loved to be sold and love to buy more. I have a client who was ready to dump avid after all this. However after visiting a post house with all the new toys he call up and said, "time to upgrade."

So, the day people stop buying. is the day they will stop selling but thats not going to happen soon.

[spenner]

Originally Posted by Jeff E

So, the day people stop buying. is the day they will stop selling but thats not going to happen soon.

Yep. When what HDX does is no longer preferred by audio engineers, then whatever is better will sell more and be increasingly seen as the main system in pro environments. After all the debates this still has not happened. Every five years it's said that native is the future and will take over HD, in five years. Those who buy HDX have their reasons. Everyone else uses native. Big deal.