switching to 32 float from 24 for working?

[EMRR]

I’m with HCmarkus on this.

No advantage other than processing prints and mixing prints. Those prints have to be bounces. If you mix analog and record back, no advantage past 24 bit. Downside to 32? The extra 25% storage. 24 unprocessed in a 32float mix, those extra 8 bits are transitory and repeatable, not stored. 32 bit in a 32float mix, it’s changed and still 32. Nothing gained that I can sense.

Parallel: I’ve a lot of 20 bit ADAT import files. Too bad there’s no 20 bit setting, I’m wasting 20% saving those extra 4 bits of zero’s.

32 recording is great for remotes you can’t redo, in a semi-dummyproof recorder with no input gain controls. If you’ve gainstaged 24 bit even remotely reasonably, there’s plenty to work with which will get handled at 32 float.

[Tidwells@aol.com]

Question: I'm on DP 9.12 with projects set to 24 bit and recently I've been freezing a lot of virtual MachFive3 tracks. To prevent overs (peaks) which I can see in the audio monitor window, I often have to insert a -10 or -20 db trim plugin before freezing, then put a +10 or +20 trim plugin on the frozen audio track to get the right volume level. If I switched my projects to 32 bit floating point, would this become a non-issue, or could this still be a problem?

Doug

[monkey man]

Subbed...

[mikehalloran]

Question: I'm on DP 9.12 with projects set to 24 bit and recently I've been freezing a lot of virtual MachFive3 tracks. To prevent overs (peaks) which I can see in the audio monitor window, I often have to insert a -10 or -20 db trim plugin before freezing, then put a +10 or +20 trim plugin on the frozen audio track to get the right volume level. If I switched my projects to 32 bit floating point, would this become a non-issue, or could this still be a problem?

Doug

The best way to find out is to try it. Duplicate a project folder, convert all the Soundbites to 32 bit float and compare. If you don’t know how, I can post screen shots—very easy.

Curious as to why anyone would remain on DP 9.12 and not update to 9.52 — many bug fixes and other improvements.


I’ll be posting a before and after on my YouTube channel of a project I did in 2020 where I’ve done exactly that. I hope to post it before Christmas if I can find the time.

[James Steele]

Curious as to why anyone would remain on DP 9.12 and not update to 9.52 — many bug fixes and other improvements.

True that! As I recall, DP 9.52 was one of those really solid versions.

[Tidwells@aol.com]

The reason I reverted to DP9.12 after trying 9.52 is because I was having difficulties doing vocal pitch editing in DP 9.52. I would try to grab a pitch segment with my mouse and use the option key to "smooth" the pitch variations in the segment and instead, the whole segment would highlight in yellow and "refuse to co-operate". Then I would have to select something else, then try to come back again and select the segment I wanted to edit. I would sometimes have to click on a segment 3 or 4 times before it would let me edit it. This was driving me crazy because I'm not a very talented vocalist and I spend a LOT of time pitch-correcting my vocals to make them more "bearable" for others to listen to Once I reverted to DP 9.12, the problem went away. I mentioned this problem in this forum, but no one else had a solution or seemed to have this problem. I would love to upgrade to DP 9.52 if I didn't have this issue.

Doug

[mikehalloran]

The reason I reverted to DP9.12 after trying 9.52 is because I was having difficulties doing vocal pitch editing in DP 9.52. I would try to grab a pitch segment with my mouse and use the option key to "smooth" the pitch variations in the segment and instead, the whole segment would highlight in yellow and "refuse to co-operate". Then I would have to select something else, then try to come back again and select the segment I wanted to edit. I would sometimes have to click on a segment 3 or 4 times before it would let me edit it. This was driving me crazy because I'm not a very talented vocalist and I spend a LOT of time pitch-correcting my vocals to make them more "bearable" for others to listen to Once I reverted to DP 9.12, the problem went away. I mentioned this problem in this forum, but no one else had a solution or seemed to have this problem. I would love to upgrade to DP 9.52 if I didn't have this issue.

Doug

Ah…. DP 9.5 introduced ZTX and defaults to it.
https://motu.com/de/news/digital-perfo ... -shipping/

You need to use PureDSP instead which became an option but it’s still there and can be selected a number of ways including Preferences. I still use PureDSP when editing individual notes.

MOTU has released videos showing the differences and when to use what.

[Tidwells@aol.com]

I agree, Mike! In DP 9.52 I selected the PureDSP option, but the "segment selection bug" was driving me crazy. Since it's a problem no one else seems to have encountered, it may be specific to my particular Mac/OS setup. I may try upgrading again after I finish my current project, which I've been working on for over 3 years. It's a labor of love and my production philosophy has been that the 12 songs are not done until I can't find any way to improve them. Unfortunately, my DP editing skills have improved so much over the last 3 years that as I go back to the first songs to make the final mix I hear a lot of things I can improve, so I keep working to improve them. I comfort myself remembering that Tom Scholz (aka Boston) took 7 years to make his second album in his home studio and it was a massive hit. Once in awhile OCD actually works to produce something great....Other times it produces sterile, soul-less "polished turds"....

Merry Christmas Everyone!

Doug

[monkey man]

Agree 100% Doug!

[labman]

No advantage other than processing prints and mixing prints. Those prints have to be bounces. If you mix analog and record back, no advantage past 24 bit. Downside to 32? The extra 25% storage. 24 unprocessed in a 32float mix, those extra 8 bits are transitory and repeatable, not stored. 32 bit in a 32float mix, it’s changed and still 32. Nothing gained that I can sense.

Parallel: I’ve a lot of 20 bit ADAT import files. Too bad there’s no 20 bit setting, I’m wasting 20% saving those extra 4 bits of zero’s.

32 recording is great for remotes you can’t redo, in a semi-dummyproof recorder with no input gain controls. If you’ve gainstaged 24 bit even remotely reasonably, there’s plenty to work with which will get handled at 32 float.

Thank you for clarifying Doug et al. Very appreciated. The one problem we had with 24 bits, was that when the peeps who were in logics' files came in, they were registering on bit readers that they were only using 14-18 bits actually used, as they were cutting too conservatively since they didnt have outboard hardware racks to compress etc on way in. So they did not seem to be anywhere near as clear and alive when the DP cats cut with 32 bit float and still threw out 6-10 of the bits. Why might that be? Thanks !

[EMRR]

That would imply the tracks coming in as 32 had higher levels, like 0dBFS if translated as directly as possible back to 24. Could even be clipped? You’d have to look at the waveforms and find it since 32 float won’t tell you anything about input clipping. It’s really hard to see an apples to apples comparison. Especially since a lot of people are confused by all the portable 32 bit recorders and what is implied about headroom. Were the AD converters true 32 bit types? Or just 32 bit data paths printing to 32 float files? The MOTU AVB stuff is 24 bit though the converters send/receive 32. There’s a very large list of other things I’d be looking at first for differences in ‘clear and alive’. Reverting to older 16 bit multitracks for mixing doesn’t trigger any less feeling of quality here, in of itself. Rooms, mics, musician and engineer talent do show up first.

[labman]

That would imply the tracks coming in as 32 had higher levels, like 0dBFS if translated as directly as possible back to 24. <snip>. Reverting to older 16 bit multitracks for mixing doesn’t trigger any less feeling of quality here, in of itself. Rooms, mics, musician and engineer talent do show up first.

Thanks. And wholeheartedly concur with last sentence.