|
Post by Johnkenn on May 15, 2017 8:27:15 GMT -6
As compared to 44.1? I realize this thread could devolve quickly - so, please - let's keep it civil...Anyway, for some odd reason, I've always recorded at 48kHz...and I get a TON of sessions through here in 48. But is there any point in dealing with the extra conversion for any gain you might be achieving. Should it just be 88.2 or 96 if you're not doing 44.1? I use my ADAT connection for my headphone system, so my headphones get funky because I lose some channels...so I've just avoided the hassle. Any thoughts/opinions?
|
|
|
Post by drsax on May 15, 2017 8:47:48 GMT -6
In my opinion it's not worth it unless for video. If I'm gonna go higher than 44.1kHz, then I'd be using 88.1 or higher. Although I rarely have heard benefits reaped that high except for very open and exposed, sparse recordings. Although I like the sound of 48kHz while working, I do not like the sound of SRC. I'm always disappointed after SRC. In the end I get better masters from staying at 44.1kHz all the way. If you work in tv and film, I believe most in that part of the industry work at 48k for video. Over here for music I'm at 24-bit 44.1khz and the masters sound excellent. That's my preferred sample rate.
|
|
|
Post by jcoutu1 on May 15, 2017 9:11:23 GMT -6
As compared to 44.1? I realize this thread could devolve quickly - so, please - let's keep it civil...Anyway, for some odd reason, I've always recorded at 48kHz...and I get a TON of sessions through here in 48. But is there any point in dealing with the extra conversion for any gain you might be achieving. Should it just be 88.2 or 96 if you're not doing 44.1? I use my ADAT connection for my headphone system, so my headphones get funky because I lose some channels...so I've just avoided the hassle. Any thoughts/opinions? Video. 48 is the standard for video. I run at 48 for the video compatibility (since EVERYTHING is video these days and CD is dying), not for a sonic improvement over 44.1.
|
|
|
Post by Martin John Butler on May 15, 2017 9:20:03 GMT -6
I agree with Jcoutou. I'm planning on trying 96 kHz next time I begin any new sessions. I've heard plug-ins can sound better at that sample rate.
I'm unfamiliar with higher sample rates when bouncing. When I bounced my first 96 kHz session, it automatically converted to 48, is that normal?
|
|
|
Post by Ward on May 15, 2017 9:23:17 GMT -6
If you're asking because of frequency response, the supposed argument that 48k will translate into reproduction of 24khz frequencies, then no. If that were the case, we could just go to 24 bit 24khz since most people over 20 years old can't hear above 12K anyhow.
But there's more to it than that. Jesse's point about video is a strong one. The other points to be made are about how plugins work with the different sampling frequencies and hardware compatibility across the board. 48 or 96 seem to be preferred and advised by most software and convertor developers.
|
|
|
Post by nick8801 on May 15, 2017 9:53:56 GMT -6
I like 96 for acoustic music. With my DAV pres the high end is crystal clear and the music feels very relaxed and comfortable on playback. For anything else 44.1 works just fine.
|
|
|
Post by adamjbrass on May 15, 2017 10:10:37 GMT -6
with 48 compared to 44.1, you get a smidgen more top end, as it just pushes that filter a little higher up. Though honestly most days I think that more information is bad. Unless its beautiful information, then I skip past all those rates and use 192K or DSD256
|
|
|
Post by schmalzy on May 15, 2017 10:32:27 GMT -6
I work at 48kHz unless forced into something else.
Video is so much happier with 48kHz and that's a legitimate revenue stream for artists. So I accommodate.
I have two interfaces - both Focusrite from the same era - and two available computers, I should mult a stereo pair out of the patch bay, record at 48kHz and 88.2kHz, downconvert to 44.1kHz, and see if they null.
Maybe when I have free time.
What's "free time" like?
|
|
|
Post by drsax on May 15, 2017 10:37:07 GMT -6
with 48 compared to 44.1, you get a smidgen more top end, as it just pushes that filter a little higher up. Though honestly most days I think that more information is bad. Unless its beautiful information, than I skip past all those rates and use 192K or DSD256 Agreed
|
|
|
Post by EmRR on May 15, 2017 10:38:42 GMT -6
Converter loop measurements I've taken suggest greatly improved phase response with higher rates, due to cutoff filter design. There could be a flaw in the measurement for all I know, but it would appear anyone looping out and back to analog at 44.1 would only need a couple passes for the top end to be 180 degrees out from the bottom end. But they say people can't hear that. Until you go on GS and people argue about the phase effects of LPF/HPF usage. But it would relate to the interplay between various channels tracked in the same bleed environment, if some get loop processing and some don't.
|
|
|
Post by svart on May 15, 2017 10:44:36 GMT -6
Other than compatibility needs, I'd say moving up to 88.2K is your best bet. It's more than double the data usage, but the difference in fidelity is readily apparent and in theory has better samplerate down-conversion without truncation.
IMHO Going to 48K from 44.1K is barely measurable, let alone noticeable.
|
|
|
Post by avgatzeblouz on May 15, 2017 11:42:15 GMT -6
How in sound history did this sample rate come to life anyway ?
|
|
|
Post by Johnkenn on May 15, 2017 12:09:48 GMT -6
Wish I could figure out how to fix my headphone system to operate the way I want it to in 88.2 or 96. All of the routing is different...
|
|
|
Post by jazznoise on May 15, 2017 12:53:26 GMT -6
How in sound history did this sample rate come to life anyway ? Sync. 44.1 doesn't divide well against most conventional framerates. So yeah, 48's for video. I actually use it sometimes. It's also slightly more forgiving for warp editing than 44.1, but you should be using 88.2 if you really intend to do dramatic editing.
|
|
|
Post by Johnkenn on May 15, 2017 14:11:54 GMT -6
How in sound history did this sample rate come to life anyway ? Sync. 44.1 doesn't divide well against most conventional framerates. So yeah, 48's for video. I actually use it sometimes. It's also slightly more forgiving for warp editing than 44.1, but you should be using 88.2 if you really intend to do dramatic editing. So warp works better at higher sample rates? Never thought about it, but that makes sense.
|
|
|
Post by swurveman on May 15, 2017 14:36:39 GMT -6
Since you're either upsampling to 48.1 from 44.1 for video, or downsampling from 48.1 to 44.1 for audio, is upsampling preferred over downsampling, or vice versa?
|
|
|
Post by ChaseUTB on May 15, 2017 15:50:42 GMT -6
Math is math... SRC doesn't have to think or work harder to convert down in non multiples ( 96khz to 44.1khz or 48k or 44.1khz )
Would love to have a dual system setup and split record the session into a 48khz and a 88.2 and then perform ABX... Most I feel would not pick correctly consistently ( including me ) And you definitely know the sample rate you choose to work at so objectivity is lost...
All video is done at 48khz and from my experience if someone is paying for a full mix and master more than 90% are shooting a music video to this song. If not they would just release the rough mix... 🤦♂️ It is what it is...
|
|
|
Post by viciousbliss on May 15, 2017 15:53:53 GMT -6
I've had a lot of forum discussions with Andy from Cytomic in that giant plugin analysis thread about plugins at higher sample rates. Based on that discussion, it really sounds like 88/96 keeps the integrity of the plugins to a much higher degree than 44/48. Someone uncovered that Metric Halo Channel Strip 3 adds rumble even when you're subtracting it. I tried doing the same test on it and got the same result. The EQ3 from Pro Tools has the same problem. When I took it out of my mixes and subbed in Pro Q2 to do the same thing, it was a better result. These rumble-adding plugs are 32 bit NF1 biquads I think the terminolgy is. Andy says there are a lot of em out there and I've been attempting to test more. Pro Q2 has that natural phase mode to correct the problems eqs have at 44, but then adds some huge amount of latency. Regarding SRC, I've read in multiple places that the SOX or whatever it is that Audacity uses nowadays is extremely accurate. It is possible to use 44-48 and not suffer the drawbacks if you don't crank compressors enough to get audible aliasing and have EQs with settings that enable one to avoid cramping. There's probably more to it than that. My guess is that advantages of 48 over 44 probably depend on what plugs you're using and how you're using them. There may be some advantages, but I couldn't say how audible they would be.
|
|
|
Post by BradM on May 15, 2017 15:59:31 GMT -6
I've recorded at 48k exclusively since the last time I did a comparison between 44.1k and 48k over ten years ago. My ears heard a much more open top end and less digital sound. When I think digital sound, I think of the sound of 44.1k. I did the comparison using my Mytek 8x192 converters at the time. I still use them. I do all my tracking to a DAW with Radar Classic converters these days (in a Radar 24)...and 48k is as high as she'll go. I use the Mytek as master clock and it sounds great. There's better resolution of micro dynamics compared to the Radar clocked internally IMHO. I prefer the Radar at 48k with Mytek clock to the sound of my Mytek at 88.2k or 96k actually.
When I need to do sample rate conversion for burning CD's I use Voxengo R8Brain Pro with the minimal phase setting. Otherwise I just let the mastering engineer handle it.
Brad
|
|
|
Post by Johnkenn on May 15, 2017 16:27:40 GMT -6
I haven't had to make a CD in about two years...lol I just print a 24/48 file and then do high quality mp3's in itunes. I'm really interested in trying 96 again, though.
|
|
|
Post by wiz on May 15, 2017 17:02:10 GMT -6
I have found it to be interface dependant to a large degree.... some sound good at one, some another. I run at 48Khz... video and some system limitations here being major reasons. My last album that Bob Olhsson mastered, had one track I accidenetly tracked at 44.1 instead of 48... I can't remember which one it was now... 8) cheers Wiz
|
|
|
Post by Bob Olhsson on May 15, 2017 18:29:53 GMT -6
44.1x16 is the limit of Sony's digital audio recorded to 3/4" video tape system which was used with Philips' video disk vinyl plant replication system. It was never considered a transparent enough format for production. 48x20 is the SMPTE production minimum based on extensive testing.
Looking at a recording from an investment standpoint, I think 48 has always made a lot more sense. It'll be around long after the CD has gone away.
|
|
|
Post by svart on May 15, 2017 18:37:44 GMT -6
Since you're either upsampling to 48.1 from 44.1 for video, or downsampling from 48.1 to 44.1 for audio, is upsampling preferred over downsampling, or vice versa? In a nutshell, upsampling is essentially educated guessing on the part of the algorithm. It analyzes the samples before and after a point in time (or a few points), then guesstimates the addition points you want it to add in between. It's creating fake audio. On the other hand, downsampling analyzes the stream and makes a decision as to what samples to throw away. It tries to find the samples that will make the least impact to the audio, but it's still throwing away audio data. I guess either one is just a fistfull of trade-offs, but my money is on downsampling having the more honest output.
|
|
|
Post by avgatzeblouz on May 15, 2017 19:43:11 GMT -6
When I need to do sample rate conversion for burning CD's I use Voxengo R8Brain Pro with the minimal phase setting. Otherwise I just let the mastering engineer handle it. Brad Here is a very useful link to check software SRC : src.infinitewave.ca/
|
|
|
Post by BradM on May 15, 2017 20:31:48 GMT -6
When I need to do sample rate conversion for burning CD's I use Voxengo R8Brain Pro with the minimal phase setting. Otherwise I just let the mastering engineer handle it. Brad Here is a very useful link to check software SRC : src.infinitewave.ca/Cool! Thanks for that!
|
|