|
|
Post by Bob Olhsson on Oct 24, 2014 11:42:50 GMT -6
Exactly, every time you crunch the numbers you lose precision. Ideally one would just throw them all up in the air as one grand calculation.
There's no point in recording from a converter at more than 24 bits but the minute you calculate DSP, you want the resulting file to be written in float.
|
|
|
|
Post by tonycamphd on Oct 24, 2014 11:49:55 GMT -6
so please correct me if i'm wrong, this is verification that technically, the most "accurate"(loosely stated) replication you can achieve would be to simply use ur converter/interface as a multi track tape machine I/O to a console, and don't touch it ITB?
|
|
|
|
Post by popmann on Oct 24, 2014 11:57:43 GMT -6
32bit float is a proprietary format--that should be noted. It will not carry from one DAW to another....so, no never RECORD at 32bit FLOAT (I keep with the caps because it's NOT literally 8 empty bits "more").
What I do in Cubase is record at 24 bit....and then flip the project to 32bit format when I go to mix. That way any internal renders I do during the mix process stay in "mixer format". In reality, that's not going to change a DSP process fucking up a signal--just CUBASE fucking up the signal....A Waves plug running double precision is still needing it's own method to reduce back to the mixer's rez.
Regardless--you don't want to record in 32bit float. Not only is there no advantage (unless you print plugs?)....there's the detriment of your track not being readable by any other DAW--the OPPOSITE of progress. That's how it used to ALWAYS be....now that we have solidly established timestamped WAV/AIFF please don't throw that compatibility away...
|
|
|
|
Post by scumbum on Oct 24, 2014 13:11:56 GMT -6
so please correct me if i'm wrong, this is verification that technically, the most "accurate"(loosely stated) replication you can achieve would be to simply use ur converter/interface as a multi track tape machine I/O to a console, and don't touch it ITB? From what I've gathered from all these discussions is for the best possible sound , use the daw as a tape machine and don't touch ANYTHING , no fader movements , no panning , nothing.......and come out your converters to an analog console . Treat the computer like a tape machine and you'll solve all your issues with digital and you'll get very close to the old school sound we all grew up with . I'd believe editing wouldn't affect the original audio files.......but what about cross fades ? Now your messing with the gain . Well I guess if you just processes the small edit points the rest of the track will be unaffected ? |
|
|
|
Post by Bob Olhsson on Oct 24, 2014 14:17:13 GMT -6
32 float is very universal. It began on Samplitude and is used all over the place. I've heard logic doesn't support it but am not sure if that's true. Pro Tools finally supported it with version 10. The problem is that 32 bit float files can easily clip a TDM system so they held off on going there.
All signal processing causes inaccuracy and properly dithered digital can be more accurate than analog.
|
|
|
|
Post by tonycamphd on Oct 24, 2014 14:23:44 GMT -6
so please correct me if i'm wrong, this is verification that technically, the most "accurate"(loosely stated) replication you can achieve would be to simply use ur converter/interface as a multi track tape machine I/O to a console, and don't touch it ITB? From what I've gathered from all these discussions is for the best possible sound , use the daw as a tape machine and don't touch ANYTHING , no fader movements , no panning , nothing.......and come out your converters to an analog console . Treat the computer like a tape machine and you'll solve all your issues with digital and you'll get very close to the old school sound we all grew up with . I'd believe editing wouldn't affect the original audio files.......but what about cross fades ? Now your messing with the gain . Well I guess if you just processes the small edit points the rest of the track will be unaffected ? This has been my understanding for a while, but not so much from a technical understanding, primarily from the results i've heard from friends who adhere to this concept. My current plan is not as extreme as using converters as "tape machine only", i will be going hybrid, but with very little ITB action, and i will have to dither each track going D/A on its way to my console, if at this point i am still hearing the "digititus" i so despise, i will indeed turn my converters into a tape machine only rig, and if that still doesn't work, i will absolutely go back to using a tape machine, the DIY skills i've been collecting over the last few years will allow me to maintain a machine for fairly little cost, so whatever it takes. I simply will not stop until i'm satisfied with the sound coming from my rig, all that said, i am happy with the sound coming out of my PT10/BLA FM clocked 002... so i think i'll be good to go. |
|
|
|
Post by scumbum on Oct 24, 2014 16:08:38 GMT -6
32 float is very universal. It began on Samplitude and is used all over the place. I've heard logic doesn't support it but am not sure if that's true. Pro Tools finally supported it with version 10. The problem is that 32 bit float files can easily clip a TDM system so they held off on going there. All signal processing causes inaccuracy and properly dithered digital can be more accurate than analog. If you know all the technical , computer stuff then the Daw can be more accurate and sound great . But you gotta know what your doing because you can easily screw things up and cheat yourself out of great sounds by over processing digitally . Analog is idiot proof and I'm an idiot . That's why I'm trying to go more and more OTB . With analog there's no dither , floating point , or plugin delay , buss delay .......nothing to worry about except making sure the gain is set correctly so you don't clip . It seems today you gotta be a computer technician just to record and mix music . |
|
|
|
Post by tonycamphd on Oct 24, 2014 17:29:20 GMT -6
32 float is very universal. It began on Samplitude and is used all over the place. I've heard logic doesn't support it but am not sure if that's true. Pro Tools finally supported it with version 10. The problem is that 32 bit float files can easily clip a TDM system so they held off on going there. All signal processing causes inaccuracy and properly dithered digital can be more accurate than analog. If you know all the technical , computer stuff then the Daw can be more accurate and sound great . But you gotta know what your doing because you can easily screw things up and cheat yourself out of great sounds by over processing digitally . Analog is idiot proof and I'm an idiot . That's why I'm trying to go more and more OTB . With analog there's no dither , floating point , or plugin delay , buss delay .......nothing to worry about except making sure the gain is set correctly so you don't clip . It seems today you gotta be a computer technician just to record and mix music . and you almost don't have to worry about clipping 8) |
|
|
|
Post by drbill on Oct 24, 2014 19:15:10 GMT -6
Oct 24, 2014 15:08:38 GMT -7 scumbum said:
" With analog there's no dither , floating point , or plugin delay , buss delay .......nothing to worry about except making sure the gain is set correctly so you don't clip .
It seems today you gotta be a computer technician just to record and mix music . "
There's a whole ton of OTHER stuff to worry about with analog. But that's a whole nut her subject. No matter what the medium, if you want to turn out top quality stuff, you've got to rise above the noise floor of commonplace recording.
|
|
|
|
Post by Bob Olhsson on Oct 24, 2014 19:25:09 GMT -6
Digital is often like building a kit. It certainly could be done a lot better than it often is. The crunchiness is the lack of dither but the other problem is aliasing which often is an opaqueness in the lower mid-range. High sample rates are the easiest way around this although there are plug-ins that don't have that problem.
|
|
|
|
Post by scumbum on Oct 24, 2014 19:26:46 GMT -6
If you know all the technical , computer stuff then the Daw can be more accurate and sound great . But you gotta know what your doing because you can easily screw things up and cheat yourself out of great sounds by over processing digitally . Analog is idiot proof and I'm an idiot . That's why I'm trying to go more and more OTB . With analog there's no dither , floating point , or plugin delay , buss delay .......nothing to worry about except making sure the gain is set correctly so you don't clip . It seems today you gotta be a computer technician just to record and mix music . and you almost don't have to worry about clipping 8) Ha , yes , Raw Power by the Stooges comes to mind....... I've always only liked gear that makes it very hard to make things sound bad . I like the boxes that have one knob , run a signal through it , and no matter what it sounds good . I don't like a million knobs , flexibility and options . That just gives me more ways to screw things up . And that's the DAW , endless opportunity's to think your doing something good when your really making it worse . |
|
|
|
Post by tonycamphd on Oct 24, 2014 19:31:48 GMT -6
Digital is often like building a kit. It certainly could be done a lot better than it often is. The crunchiness is the lack of dither but the other problem is aliasing which often is an opaqueness in the lower mid-range. High sample rates are the easiest way around this although there are plug-ins that don't have that problem.Thanx Bob, would you mind recommending some of the good ones? |
|
|
|
Post by scumbum on Oct 24, 2014 19:37:28 GMT -6
Oct 24, 2014 15:08:38 GMT -7 scumbum said: " With analog there's no dither , floating point , or plugin delay , buss delay .......nothing to worry about except making sure the gain is set correctly so you don't clip . It seems today you gotta be a computer technician just to record and mix music . " There's a whole ton of OTHER stuff to worry about with analog. But that's a whole nut her subject. No matter what the medium, if you want to turn out top quality stuff, you've got to rise above the noise floor of commonplace recording. Yeah Tape to me is a nightmare . When I say going analog I mean using a digital device to record and edit , but then mixing it all OTB . Also I'm wanting to record rock , not Modern Pop . |
|
|
|
Post by Bob Olhsson on Oct 24, 2014 20:03:40 GMT -6
Sony Oxford is pretty transparent at low sample rates. DMG is another if you turn off all the extra effects. I use lots of other stuff too but when the audio is really clean they seem to do the least harm. Up at 96 there are lots more choices. What's interesting is that a model can sound wonderful on some things and not good at all on other things.
|
|
|
|
Post by jimwilliams on Oct 25, 2014 10:17:18 GMT -6
According to the folks at Metric Halo, their measurements showed minimum distortion with program peaks at -12. That seems unusual as all the top-o-line converter chipsets show decreasing THD with increasing levels. Most well designed transistor circuits will do that too. Sometimes it's the noise that is factored out at higher measurement levels, but IMD tests eliminate that out. Perhaps some designs like Burl with saturating transformers will show a right turn on THD vs amplitude measurements, but well designed transistor stages should not do that. Years ago I had some discussions with Steven Green at Cirrus Logic, the Crystal audio division head designer. I found their 'new' 5381 ADC slightly increased THD at -1 db full scale to .002% THD compared to -3 db at .001%. We both used the same test rig, Audio Precision. He was stretched to explain this, I thought the modulators were saturating slightly but he had no way of measuring that as it's embeded into the die and it's not accessable seperately. Later chipsets had the modulator outputs available apart from the converter so that problem was eventually solved, but not for that design. The PCM/DSD 4222 ADC from BurrBrown is my current choice, it does -117 db THD or about .0005% THD, not too bad. |
|
t3
Junior Member
I'm in the hi-fidelity first class travelling set.
Posts: 54
|
Post by t3 on Jan 22, 2015 21:35:40 GMT -6
Maybe ya'll are all used to it by now but I find myself listening to pre-digital/DAW music almost exclusively...I had expected that by now these artifacts would be eliminated but the music still sounds like crap to me compared to anything recorded before a computer touched it. Hey, I thought I was the only weirdo doing that... |
|
|
|
Post by Johnkenn on Jan 22, 2015 23:03:12 GMT -6
Where is scumbum? Hasn't been on since November... |
|
|
|
Post by tonycamphd on Jan 22, 2015 23:18:24 GMT -6
Where is scumbum? Hasn't been on since November... he's been known to take long hiatuses, he'll be back. You have his regular email because of the RM DAC deal you stole from me right? 8) U could always hit him up there and tell him to get his arse over here lol |
|
|
|
Post by Johnkenn on Jan 23, 2015 9:15:10 GMT -6
|
|
t3
Junior Member
I'm in the hi-fidelity first class travelling set.
Posts: 54
|
Post by t3 on Jan 23, 2015 12:00:10 GMT -6
Yes I agree the time spent chasing things like this takes away from actually making music and one of the reasons I've been staying away from the internet and forums . SOOOO much time wasted when I could have written or worked on a song . Where is scumbum? Hasn't been on since November... This might be why. Great thread, thanks to everyone who participated. |
|
|
|
Post by tonycamphd on Jan 23, 2015 13:02:49 GMT -6
this is the best thread ever! learned so much on this one, i think if you got JW, Bob O, and Pop in a room together, all the gear would run and hide 8) I do have one more Q though, does housekeeping editing of the wave forms constitute processing and a need to apply dither? I wouldn't think so, but i figured it'd be worth it to ask. Bob Olhssonthanx |
|
|
|
Post by Johnkenn on Jan 23, 2015 16:36:31 GMT -6
Not that I looked...but he might have been on another board here lately  Cheater! |
|
|
|
Post by swurveman on Jan 24, 2015 8:57:31 GMT -6
wow, this thread really turned south, Scum asked a legit Q, He couldn't get a master fader bounce and a - master fader bounce to null, does anyone have a good (not smart ass) answer for that? @jordanvoth, i appreciate you saying you weren't trying to dis, but the thing is, it's a thread that you are not required to participate in if you have nothing positive to contribute, right? So if you're bored by the content, then why bother? Honestly, reading guys saying "it's the performance/song/mic placement" or "gear doesn't matter" on a GEAR FORUM! is unbelievable? Ya know what else is tiring? when guys repeatedly try to protect their paradigm by call things they may not hear or don't understand BS, in the face of guys that do or have the audacity to inquire, this is just another form of defensive confirmation bias. I digress, there are a lot of seriously great engineers who realize the results of what they achieve are contingent upon the sum of thousands of small things, thats why they mix in beastly outfitted studios, or strive to improve their own gear until satisfied, they get that every single piece counts in conjunction with the skill of application(the song/performance/competent tracking in a great room are simply assumed). The undesired effect of what great engineers complain about when mixing ITB is commonly referred to as "digititus" (for lack of a technical answer, or better term), i've heard many of them say this, and most of them hybrid up, and use consoles and hardware because of it, I'm certainly not putting myself in their company, but I can honestly say i've NEVER been satisfied with a single ITB mix i've done, because i personally CAN HEAR that over edgy, disconnected "digititus" effect, i also hear that effect accumulate in direct correlation to more tracks and plugs getting used ITB, perhaps this is what scum is hearing? My dissatisfaction with this phenomenon along with latency is why i'm putting forth a long laborious effort, and obnoxious expense to move primarily OTB, but hey, i guess i could be full of BS 8) Here is a little vid comparison of ITB vs OTB summing, is the obviousness of these comparisons BS also? What's interesting to me about the itb vs SSL summed mix is the reverb/ambiance of the song. That above all imo is what makes the itb mix sound too bright. In comparison, the reverb/ambience seem smaller and shorter-like its in a smaller room- than the itb mix. I don't know why that is, but would be interested in hearing anybody's opinion. |
|
|
|
Post by Bob Olhsson on Jan 24, 2015 17:02:33 GMT -6
Anything that expands the bit depth needs dither such as gain changes.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 31, 2015 18:53:26 GMT -6
I got results , it sounds better and doesn't null , I did a DAW A sounds better than DAW B on the same hardware thread and got my ass handed to me by dozens of people who thought I was either a liar or a quack and the thread has tons of pages and keeps coming back up to haunt me years later. But the reality is, they didn't null, and it wasn't subtle either. (DP vs. PTHD both running on DAE engine / PTHD hardware.) I was METICULOUS about the testing protocol, and no one could fault me on that. I assume you were too. Right? Maybe something is broke in your DAW or you have a plugin that's acting whacky. I've been doing the PT thing for close to two decades and have never had the experience you have had. Something is whack. Edit: Lol the never ending DAW debate.. Anyway, I noticed with the master fader turned on. There's phase incoherency happening, chances are with your singular tracks going through line out with plugins on them and also a non latent version being re-routed to the master fader. There's latency between the two out signals, one thing I've noticed in Pro Tools using VTM and some heavier UAD plugs is that there latency compensation sucks.! I added VTM to the master fader, everything went out of time and sounded like a sonic mess. I use it to make sure I'm not clipping then turn the thing off.! Never had this problem in Logic. |
|
Cheater!
