ORIGINAL: RnRmaChine
Another question though: So if you own "cheap" converters you would technically gain by tracking at higher samples rates but if you coughed up a wad for real expensive converters it's not going to make enough of a difference to matter in a real world sense to compensate for the added grief of trackin/converting and the work that goes into taking the extra steps. I am sure we all agree there are MANY times when the extra steps taken in ANY process are what separates the novice from the pro.

Bingo! :)

ORIGINAL: RnRmaChine

Yea, I talked to E-mu and they said they are the only other company other then pro tools that has the "rights" to use them.

This of course is nonsense. Why would AKM limit their market like that? Some cards that use the AKM5394 AD chips:

Lynx 2
ESI Waveterminal 192X
Wamirack 192X
MOTU HD192
Audigy 4 Pro
Sound Blaster X-Fi Elite Pro

But much more importantly, the chips used are only a tiny bit of the equation. There is alot more involved in designing good converters than just dropping in a good chip: Analogue stages, analogue anti-aliasing filters, power supply, power supply filtering, op amps, grounding, circuit topology, RF rejection, clock etc. E-mu have been misleading with their advertising since the begining.


UnderTow

Good post Steve, just one thing:

ORIGINAL: SteveD

Finally, one should be aware that recording at high sample rates and then downsampling to 44.1khz for CD production without expensive hardware SRC can do just as much or more damage than tracking and mixing at 44.1k by an experienced engineer.

You don't need expensive hardware SRCs any more. On the contrary, doing the conversion in software allows the coders to through the whole might of the modern CPU at the problem and the algorithms don't have any real-time constraints.

The best SRCs are software these days. (Stuff like Weiss Saracon, iZotope 64 bit or R8brain Pro).

UnderTow

ORIGINAL: UnderTow


Good post Steve, just one thing:

ORIGINAL: SteveD

Finally, one should be aware that recording at high sample rates and then downsampling to 44.1khz for CD production without expensive hardware SRC can do just as much or more damage than tracking and mixing at 44.1k by an experienced engineer.

You don't need expensive hardware SRCs any more. On the contrary, doing the conversion in software allows the coders to through the whole might of the modern CPU at the problem and the algorithms don't have any real-time constraints.

The best SRCs are software these days. (Stuff like Weiss Saracon, iZotope 64 bit or R8brain Pro).

UnderTow

I had heard that, but wasn't sure it was true. Learned somethin' today.

Thanks UT.

r8brain pro with the ultra steep curve an min-phase is great.

Kind regards

Dave Rich

I heard Dan Lavry is of the opinion that 60K would be the optimum sampling rate, but unfortunately it is not one of the standards. What do you think of this?

ORIGINAL: mtl777

I heard Dan Lavry is of the opinion that 60K would be the optimum sampling rate, but unfortunately it is not one of the standards. What do you think of this?

Correct. That is Dan's position. That would make 88.2K more than sufficient to satisfy Dan's OPTIMUM sampling rate for DAW efficiency and plugin performance. Dan Lavry is quick to state that this doesn't alter the Nyquist-Shannon principal and is related to conversion artifacts from ADC and plugins that find their way into the audible range. That's why plugins sound different (better?) at higher sample rates and why some plugins perform their own internal up and down sampling while processing.

Dan states, and backs up with scientific data in white papers, that sample rates greater than 96k begin to move into diminishing returns in quality vs performance. Simply put, you just can't do any better. In fact, he states and attempts to prove that due to processor limitations (which is always changing) audio quality begins to decline due to distortion with a 192k sample rate.

With today's processors, the decline in quality at 192k may be getting addressed. However, I'm convinced the sound doesn't get any better when the sample rate is greater than 60k. I'm also convinced that quality ADC and high-end plugins sound excellent at 44.1k, and that in that scenerio, the improvement realized at 96/88.2k is very hard to hear on a 44.1/16 CD, and isn't worth the loss in track count, plugin count, and the time it takes for bounces and exports.

As this relates to the posters original topic... (44.1k vs 48k... remember?) 48k doesn't satisfy Dan's 60K threshold for optimum sound and performance, but your converters may sound better with that little extra headroom because the aliasing artifacts will be above 22k and out of the audible range. But then you'd have to be sure you're using excellent SRC when downsampling to a 44.1/16 CD if you want to be sure you're not doing more harm than good. If your target is DVD... then the reverse is true and 48k is the way to go.

I do drum tracks for clients at 44/48/88/96 sample rates. I hear absolutely no difference between 44.1k and 48k projects when recording in my studio.

YMMV.

ORIGINAL: SteveD
Dan Lavry is quick to state that this doesn't alter the Nyquist-Shannon principal and is related to conversion artifacts from ADC and plugins that find their way into the audible range.

This is the key. It's not about the frequencies in and of themselves (arguments of human hearing ability aside), it's all the various external variables.

That's why plugins sound different (better?) at higher sample rates and why some plugins perform their own internal up and down sampling while processing.

But again, if you have a "high quality" plugin that upsamples for processing then downsamples for output, and the SRC that it uses is "less than perfect", then you're better off driving the plugin at the higher rate to begin with. Of course, the only way to test that is for each such plugin to have a special bypass mode that still goes through SRC, which for some would only be airing dirty laundry...

ORIGINAL: keith

That's why plugins sound different (better?) at higher sample rates and why some plugins perform their own internal up and down sampling while processing.

But again, if you have a "high quality" plugin that upsamples for processing then downsamples for output, and the SRC that it uses is "less than perfect", then you're better off driving the plugin at the higher rate to begin with. Of course, the only way to test that is for each such plugin to have a special bypass mode that still goes through SRC, which for some would only be airing dirty laundry...

That's a good point... and I've often thought about that. The only thing I can offer is that if it's a high quality plugin, the up/down sampling is the best and most efficient way to deliver the quality sound the plugin is known for. For example... UAD-1's Pultec eq upsamples to 192k. Does it need to go that high? Wouldn't 88.2k or 96k be enough? I can only assume that there were enough artifacts in the upper range when modeling that exquisite piece of hardware that a steeper slope conversion was too resource intensive or maybe would have made the product too expensive. Maybe it's just plain marketing . However, UAD-1's Pultec eq IS in fact in incredible sounding plugin. I use it... and trust those guys have their reasons. It is, by the way, one of those plugins that do sound different at various sample rates. But I like the way it sounds... always.

Good post.

ORIGINAL: SteveD
For example... UAD-1's Pultec eq upsamples to 192k. Does it need to go that high? Wouldn't 88.2k or 96k be enough? I can only assume that there were enough artifacts in the upper range when modeling that exquisite piece of hardware that a steeper slope conversion was too resource intensive or maybe would have made the product too expensive.

Right. At some point, though, as a musician you just have to stop think about it so much, turn the knobs till it sounds good, and trust your ears... otherwise you go crazy doubting and double-self-double-doubting what you think you might be hearing versus what you think you shouldn't and vice versa...

Of course, if you make a living as an engineer or studio owner, then ultimate sonic clarity should certainly be something you strive for. When the CD comes back and it sounds like crap, at last you know it's not your stuff that's to blame...

Maybe it's just plain marketing . However, UAD-1's Pultec eq IS in fact in incredible sounding plugin. I use it... and trust those guys have their reasons. It is, by the way, one of those plugins that do sound different at various sample rates. But I like the way it sounds... always.

Interesting about something like UAD, they can bury the overhead of the higher-quality SRC in DSP. Whatever overhead is incurred is considered the cost of running the plug. And maybe a signficant cost of some of the hungrier UAD plugs is higher quality SRC and filtering for a more sonically pure result (i.e, not just more accurately modeled components, but better input and output stages as well).

Anyway, regarding plugins in general... one thing that shouldn't be lost in all this is that most plugins also have a quality setting -- good, better, best. I think UAD plugs have that as well (not sure, don't own a UAD). In any event, if you like the plugs you use or you use the plugs that you can afford, at least don't forget to do your offline bouncing with the high quality setting enabled!

ORIGINAL: SteveD

...(snip)...

I'm also convinced that quality ADC and high-end plugins sound excellent at 44.1k, and that in that scenerio, the improvement realized at 96/88.2k is very hard to hear on a 44.1/16 CD, and isn't worth the loss in track count, plugin count, and the time it takes for bounces and exports.

As this relates to the posters original topic... (44.1k vs 48k... remember?) 48k doesn't satisfy Dan's 60K threshold for optimum sound and performance, but your converters may sound better with that little extra headroom because the aliasing artifacts will be above 22k and out of the audible range. But then you'd have to be sure you're using excellent SRC when downsampling to a 44.1/16 CD if you want to be sure you're not doing more harm than good. If your target is DVD... then the reverse is true and 48k is the way to go.

I do drum tracks for clients at 44/48/88/96 sample rates. I hear absolutely no difference between 44.1k and 48k projects when recording in my studio.

YMMV.

After reading this entire thread I have come to the conclusion that from now on I will be recording my projects at 48K. 88.2K and higher are out of the question because of the severe hit on disk and CPU resources, track count, plugin count, etc.

I am using a Mytek Stereo96 ADC and maybe the conversion will not be much better at 48K than 44.1K considering the quality of my converter, but I think the additional headroom will be of significant benefit and yet the penalty on resources isn't that much. Current SRC's are already very good (e.g., r8brain pro) and they will only get better in the future, so I'm not worried about downsampling the final mix to 44.1K for CD. For me and my needs, 48K represents the best balance considering the current available technology.