Love em or Hate Waves finally puts DLS bundles on sale

[upscaps]

Originally Posted by propower

Assuming 96kHz and zero additional latency plugins ...

An Avid HDN system at
96/64 = 1.8ms Latency
96/128 = 3.12ms
96/256 = 5.8ms
96/512 = 11.2ms

An HDX system at 96/128 (the lowest setting it has ) with only dsp plug ins on record enabled channels (because that's what its designed to do)
96/128 = 0.7ms
96/256 = 0.7ms
96/512 = 0.7ms

There are many scenarios where being able to increase the buffer and not look back is preferable from a workflow POV. There are also a ton of scenarios where a Native system either has acceptable latency for the user (a personal opinion and tolerance level) or can be quickly "arranged" to go back to the lowest latency number.

Doesn't HD Native do some things via it's PCIe to keep latency low? What about a regular native system w/ "regular pro tools", motu interface, etc. at 44.1/48?

[propower]

Originally Posted by upscaps

Doesn't HD Native do some things via it's PCIe to keep latency low? What about a regular native system w/ "regular pro tools", motu interface, etc. at 44.1/48?

Not sure what you mean about PCIe. FWIW for HDN 1.8ms is the lowest latency THROUGH ProTools Native (with plugins) system. 0.47ms for A/D D/a and 1.33ms for (2*64)/96kHz. Sure there is Low Latency monitoring that avoides the PT buffer (but that bypasses plugins (and the protools mixer!)). LLM = 0.47ms RTL.

All other FW or USB interfaces will be significantly higher latency through PT because of the i/o bufering needed to get on and off the USB/FW bus. Other PCI solutions can certainly approach the 1.8ms number and theoretically can be even faster if the driver is very efficient (necessary since PT driver is near zero latency) and the D/A A/D is faster. Still - marginally faster is all that is possible - maybe 1.6ms.

At 44.1 for an AVID interface A/D D/A time is 1.9ms. Many other interfaces are faster at this - some are slower. With regards to latency THROUGH ProTools - The A/D + D/A time + (i/o time to get on and off whatever bus is used - FW, USB, Ethermet, PCI, TB) + 2*(PtoTools buffer Setting)/(Sample Rate) = Round Trip Latency (RTL)

[upscaps]

Originally Posted by propower

Not sure what you mean about PCIe. FWIW for HDN 1.8ms is the lowest latency THROUGH ProTools Native (with plugins) system. 0.47ms for A/D D/a and 1.33ms for (2*64)/96kHz. Sure there is Low Latency monitoring that avoides the PT buffer (but that bypasses plugins (and the protools mixer!)). LLM = 0.47ms RTL.

All other FW or USB interfaces will be significantly higher latency through PT because of the i/o bufering needed to get on and off the USB/FW bus. Other PCI solutions can certainly approach the 1.8ms number and theoretically can be even faster if the driver is very efficient (necessary since PT driver is near zero latency) and the D/A A/D is faster. Still - marginally faster is all that is possible - maybe 1.6ms.

At 44.1 for an AVID interface A/D D/A time is 1.9ms. Many other interfaces are faster at this - some are slower. With regards to latency THROUGH ProTools - The A/D + D/A time + (i/o time to get on and off whatever bus is used - FW, USB, Ethermet, PCI, TB) + 2*(PtoTools buffer Setting)/(Sample Rate) = Round Trip Latency (RTL)

PCIe card the hd native system comes with. we are talking about recording thru plugins correct? don't native systems recording thru plugins have to account for daw buffers and in/out trip or am I wrong?

[propower]

Originally Posted by upscaps

PCIe card the hd native system comes with. we are talking about recording thru plugins correct? don't native systems recording thru plugins have to account for daw buffers and in/out trip or am I wrong?

You are correct - in the equation in paragraph 3 of my last post "(2*ProTools Buffers)/SampleRate" is exactly this part (daw buffers in and out - thats why there is a 2 multiplier).

For complete accuracy one may have to also add in any samples of delay the particular plugins may also add...

[upscaps]

Originally Posted by propower

You are correct - in the equation in paragraph 3 of my last post "(2*ProTools Buffers)/SampleRate" is exactly this part (daw buffers in and out - thats why there is a 2 multiplier).

For complete accuracy one may have to also add in any samples of delay the particular plugins may also add...

Thanks for the explanation.

Maybe I should give "autotune live" another go. We are on hdx but keep an HD rig floating mainly just for clients who want to record thru autotune (tdm). No one here trusts "autotune live" will do the job and provide the stability of autotune tdm.

[Ghost In The Attic]

I called Waves about a DLS system, but found out you have to use them in the Studio Rack only. If you want to use individual plugins you have to stay Native. So, I won't be buying one. Waves has really made a bad decision to go rack only for the DLS in my opinion.

[nerd513]

yes auto tune live is the way to go. that is actually the only native plug in i know of that is actually designed to track with. I have used it extensively and i can say at a low buffer it is suitable. its not as good as auto tune TDM but it works well enough. It really is sad that people would spread misinformation and swear its true but i guess it makes them feel better somehow. for example my normal channel strip for my vocals would be auto tune tdm> digi 1 band eq tdm> digi delay tdm>sometimes rverb on an HD system. On a HDnative system i can only use auto tune live kind of sucks but at least antares was smart enough to make a native solution. if u add a eq and delay to the chain it will be unusable so i dont what those other jokers are referring to when the falsely claim that there are certain native plug ins u can track with no problems. a one band digi eq and a digi delay are about as low latency as u can get but it wont work on HDnative i have tried it would be nice but its not happening. I have never used auto tune live on an Mbox or 003 so i cant say if it will work but i have used it a lot on HDnative and it is fine at 64 or 128 buffer 256 is borderline but still workable 512 almost unusable

[Emcha_audio]

So to put the claim of one to rest, I thought I'd provide a bit of proof.

Disclaimer. I'm not using autotune live so I cannot say how much latency it adds, that's not what my artist need while tracking.

Also to note, while I'm using PT HD 11, I'm using a liquid saffire 56 as interface, so this is a pure native system that I'm showing.



Note that I have the session running at 96khz, and the LS56 buffer is at 256. Noticed that the record armed track is showing a 39 samples delay .. so that's 39th of 96000 of a second.

Eq1 (0 samples)
Pro compressor (0 samples)
Maag eq 4 (0 samples)
Pro limiter (the culprit with 39 samples)

As you can see the session was running, with over 50 tracks of music and vocals, including some VI's that are not shown in the pick.

No one that has been using my studio has complained about having problems.

So if you choose carefully the plugins you use, you can indeed track through them without problems. But on the other hand, if I was doing a bigger project and was further down the mix stage (which this song was not) then I'd definitely recommend a HDX system (which all of the plugins I am using in this recording chain are dsp too hehe).

[propower]

Not sure what we are trying to prove here - of course you can track at 96kHz 64buffer through plugins even - done it tons of times.

As far as what latency is acceptable - its always a personal choice. For me - more than 3.5ms is objectionable (comb filtering - not slap back) and more than 5ms very objectionable.

As far as the latency of the above signal path it is...

A/D + D/A time + any driver buffer latency to get things on and off the FW or USB bus (usually 1ms each) + 2*256/96000 + 39/96000. For the parts we know that is

A/D + D/A + Driver buffer + 5.74ms.... probably the total is ~7 to 8ms.

[Emcha_audio]

Originally Posted by propower

Not sure what we are trying to prove here - of course you can track at 96kHz 64buffer through plugins even - done it tons of times.

As far as what latency is acceptable - its always a personal choice. For me - more than 3.5ms is objectionable (comb filtering - not slap back) and more than 5ms very objectionable.

As far as the latency of the above signal path it is...

A/D + D/A time + any driver buffer latency to get things on and off the FW or USB bus (usually 1ms each) + 2*256/96000 + 39/96000. For the parts we know that is

A/D + D/A + Driver buffer + 5.74ms.... probably the total is ~7 to 8ms.

at 7 or 8 ms a finely trained ear might perceive the combfiltering, even more so if there's no delay or reverb used at the same time. But the average musician/signer wouldn't, at least that's been my experience ymmv.