KVR Audio

Funkybot · Post by **Funkybot** » Mon Feb 21, 2005 2:20 am

Ok, I saw the one on the other forum but aside from the obvious flaws a proper test would need a few things. Actual live audio recorded through the same pre's and same converter at 44.1 and 96.6, as well as VSTi's. After that it would also require two identical mixes to be done in the same host for each of these projects, with things like EQs, reverbs, compressors, etc. Settings would have to be kept the same on both projects, which means volume and pan settings, EQ and compression settings, etc. Then the 96khz mix should be brought down to 44.1 using the best samplerate conversion settings available in said host, and the two resulting files should be posted somewhere without any indication of which is which. This would be the only way to test 44.1 vs. 96khz in any way worth listening to IMO. Individual audio/synth tracks may not be enough as if there's just a slightly better top and bottom end at 96k it may be more apparent in the real world context of a full mix. Same thing with using EQs and whatnot. We should be looking to as close of a real world test as possible, which should include live audio, VSTi's, VSTfx's, and a full mix. If anyone is up to doing this, I'd be highly interested.

The only downside is, I imagine this would need two computers with the same audio interface and an A/B/Y box of some sort to create two identical paths from the same source for the live instruments (otherwise the test would be bunk). Any takers?

C00kie · Post by **C00kie** » Mon Feb 21, 2005 10:03 am

No need for this test, we all know the difference by now. And the test you propose is very hard to set up.

For true audio input (from outside the PC) it is sufficient to record at the same rate as the intended end products sample rate. That will be 44.1kHz for CD. Catching the higher frequencies eats up harddisk space and CPU cycles. Most people say its not worthwile, since we're no bats.

For VSTi's (maybe some effects also) the quality may differ between sample rates. Some synths that have no internal oversampling will benefit from higher rates. Some other synths/effects on the other hand will sound different (fixed filter coefficients) or will even malfunction.

Too much variables in the equasion... If you are happy with lofi gritty recording at 8kHz, then do that. If recording at 192kHz puts a smile on your face, then do that. We all have the choice.

ericj23 · Post by **ericj23** » Mon Feb 21, 2005 10:15 am

surely youcould just use a sample - ok they will only come at 44.1 but that can easily be upsampled to 96 - then you wil get a quick way of checking whether processing at a higher sample frequency and then sampling down actually makes it better/different

the thing i notice at 96 is the reverb tails - smmooooth

so could any test include eq, reverb and .......

UltraJv · Post by **UltraJv** » Mon Feb 21, 2005 11:00 am

C00kie - Its surprising how many people think the change from 44.1 is radical. Its easy to be fooled by the fact that VSTis can sound different due to oversights on the programmers part and FX within the plug change and add an unintended different quality or level change. In trying to establish how wonderful going above 44.1 is, its easy to find stumbling blocks which confuse the issue. This came about on another thread :

http://www.kvraudio.com/forum/viewtopic ... sc&start=0

ericj23 · Post by **ericj23** » Mon Feb 21, 2005 11:06 am

im sitting listening to a 96khz dvd-a of lee perry

there is no fooling involved - the increased clarity actually means i can understand the words on songs which are mastered from 30 year old 7" singles

and there are several other examples that I own- 96/192/sacd uniformly leads to a smoother top end and better instrument seperation

this has nothing to do with differnces in vst performance

but i'm still not convinced that if your going to master at 44.1 the benefits of the tiny resulting difference between 44.1/96 once downsampled are really worth the hassle

UltraJv · Post by **UltraJv** » Mon Feb 21, 2005 11:08 am

Yes, acoustics are one area that do benefit - My point is VSTi may not. On the subject of remastering, its possible to make something sound sweeter within the realms of 44.1 without going higher. I wonder how the sound of the SACD sounds when tranferred to CD?

ericj23 · Post by **ericj23** » Mon Feb 21, 2005 11:10 am

UltraJv wrote:Yes, acoustics are one area that do benefit - My point is VSTi may not.

yeah of those 4 instrument comparisons yesterady only zeta didn't seem to have some wierd effects going on - double delay - double vibrato

this does make me wonder if these same synths then perform differently at 48 Khz tho

scary thought

Spaceman Sounds · Post by **Spaceman Sounds** » Mon Feb 21, 2005 11:11 am

My test wasn't really about recording .wavs or d/a converters in fact some 96khz d/a converters are noisier at 96 than 44.

My point was just that the extra resolution to the vsti engine working at 96khz as apposed to 44khz gave vsti's and vsti fx like reverbs, flanges and phasers and ring modulation a much cleaner deeper sound.

I really must stress that my second test (as was my first), really was a straight export of the same midi file and same vsti .... but exported at different sample rates...then converted to 44.1khz/16bit and placed end to end in soundforge then posted on this forum. If one came out louder than the other then I have no idea why...or if the delays were far louder on one than the othe...then this is a problem with the vsti maker.

I do know that a lot of vsti's don't compensate for 96khz i.e the lfo's are over twice as quick as are some adsr's. There is a genuine difference whether you think its a good one or bad one is down to taste, but simply exporting a vsti at 96khz with the sequencer engine set to 96k makes a difference.

Having the sequencer engine at 44.1khz and exporting at 96khz makes no real difference that I can think of which is a shame.

Also no one has yet replied to my point that my Uas-1 card which make some of the best mastering plugins ever...3 of those top plugins operate at 192khz and resample back down to 44.khs or whatever sample rate you are working at. Uad say the gains from upsampling far outway the losses from having to downsample again to the project samplerate. These guy's really know thei stuff. If you do an eq boost on the Cambridge eq at 44k then the same boost at 96k the difference is black and white...Oh and my Creamware synths upsample to 96k as well...I'll post some examples later if I get time..got to finish this track first

ericj23 · Post by **ericj23** » Mon Feb 21, 2005 11:17 am

UltraJv wrote:Yes, acoustics are one area that do benefit - My point is VSTi may not. On the subject of remastering, its possible to make something sound sweeter within the realms of 44.1 without going higher. I wonder how the sound of the SACD sounds when tranferred to CD?

lots of recent rolling stones cds are mastered in this way - original tapes into a sacd recorder - then converted down

don't own any of them although let it bleed is quite tempting but i have heard people speak very highly of them

The remastering -- performed with Direct Stream Digital (DSD) encoding -- is a drastic improvement, leaping out of the speaker yet still sounding like the original albums. This is noticeable on the standard CD layer

still it's all rather irrelevant - just play the sacd layer !

A Proper 44.1 Vs 96khz Test Would Need...