bmpcc 4k - Jitter or not?

The place for questions about shooting with Blackmagic Cameras.
  • Author
  • Message
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

bmpcc 4k - Jitter or not?

PostFri Jan 28, 2022 7:36 pm

Hi there,
actually I'm looking for a timecode generator to sync my audio bag (zoom f8) to the bmpcc4k.
To get an idea of what I'm looking for I did some initial testing:

1. jam sync the bmpcc4k to the zoom f8 once and let both run for a couple of hours
2. continuously receiving timecode from the f8 via bnc to 3.5mm

Just jam sync every couple of hours does not work for music as expected. There is way to much drift. But even with syncing continuously there is a drift between the internal mic and my soundbag. It is always under one frame - sometimes 8ms ahead, sometimes 6ms behind. Taking a closer look to the images does not clearly reveal who is the one that jitters, the zoom f8 or the bmpcc4k.

Am I to picky? Do I have to live with it? Can this only be solved with genlock and timecode?

Thank you for your help,
Stephan
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos
Offline
User avatar

Uli Plank

  • Posts: 21280
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: bmpcc 4k - Jitter or not?

PostSat Jan 29, 2022 1:41 am

You may expect too much. After all, 6-8ms is only a fraction of a video frame, so who'll notice?
Perfect genlock needs classic broadcast cameras.
No, an iGPU is not enough, and you can't use HEVC 10 bit 4:2:2 in the free version.

Studio 18.6.5, MacOS 13.6.5
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G, iMac 2017, eGPU
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

Re: bmpcc 4k - Jitter or not?

PostSat Jan 29, 2022 6:00 am

Ok, that's good to hear. Sometimes you stumble across a problem you never noticed before and it starts to drive you crazy.

Up to the next question (but still related):
Right now I'm considering the following timecode generator:
- Ultrasync One
- Betszo TCX-2+
- Ambient Nanolockit
- Tentacle Sync E (not available right now)
- Denecke JB-1

Accuracy seems to be very decent for all (0.5ppm max), size is very close. The first two have continuous sync via RF (~865 MHz in Europe), three and four via bluetooth and the last does not sync wirelessly, but runs forever on battery (up to 6d).

How important do you consider permanent sync regarding jitter/drift? Do I need two devices, or could the zoom f8 be used as a master without getting into trouble? I'm talking about regular production days up to 8h with usually just shorts breaks for snacks where you do not want to fiddle with you gear.

Regarding wireless connection, I think both frequenzy bands are problematic for wireless audio (interference). Uhf might be better, as at least you can choose a channel to use. 2.4GHz is already packed with almost every cellphone, wlan, bluetooth device and what so ever.

What do you think? Am I too cautious again?

Thank you for your help,
Stephan
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: bmpcc 4k - Jitter or not?

PostSat Jan 29, 2022 11:35 am

The advantage of a full time clock is that it protects you against a loss of timecode if you power cycle for example and forget to re-jam. You could just run that without if you want the risk.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

Bromine 18

  • Posts: 145
  • Joined: Tue Feb 09, 2021 4:00 pm
  • Real Name: Aldous Barnes

Re: bmpcc 4k - Jitter or not?

PostSat Jan 29, 2022 4:34 pm

kontrakatze wrote:Right now I'm considering the following timecode generator...


I’m looking into these as well, in addition to the Track E and UltraSync BLUE.

If you go for one of them, let me know how it goes.

As far as Tentacle an UltraSync are concerned, I think one device should suffice for your purposes.
Offline

Ric Murray

  • Posts: 205
  • Joined: Tue Nov 07, 2017 8:45 pm
  • Location: North Kingstown, RI USA

Re: bmpcc 4k - Jitter or not?

PostSat Jan 29, 2022 4:46 pm

The Tentacle system is excellent, accurate, and easy to use with their bluetooth app on an iPhone. I have 3 units that I use on a Red A Cam, a BMPCC4K B Cam, and a Zoom F4. On long shoots you can check the sync several times a day on the phone app to be sure, but I have never had a drift. It's also excellent for monitoring battery status. In post, just select and "sync w timecode" and you are golden. In complicated edits where things might get bumped out of sync by accident you can always select the sound and visual in the timeline ( in Resolve or Premiere) and resync using TC. I would never work on a set without TC on all units ever again.
Creativity is the ability to accept ambiguity.
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

Re: bmpcc 4k - Jitter or not?

PostSat Jan 29, 2022 5:05 pm

Thats very interesting to hear. What about rf interference? Has anybody had any issues with rf and wireless audio? I‘m using lavaliers, like dpa or cos11 a lot. So going wired is very much not an option.
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos
Online
User avatar

rick.lang

  • Posts: 17175
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

bmpcc 4k - Jitter or not?

PostSat Jan 29, 2022 6:37 pm

I rely on the Tentacle Sync with the MixPre-6 II (usually wired connections) for XLR and also successfully used the Sennheiser wIreless AVX lavalier with XLR from the Receiver. I have tested the MixPre-6 II and it is reliable for many hours so I’ll jam the recorder with a Tentacle and then leave Tentacles mounted to my cameras as the BMD UM4.6K and BMPCC4K cameras aren’t quite as reliable to maintain sync for several hours. Best to resync everything anyway if your shoot includes a lunch break.

Other devices are also good, but I’m happy with the Tentacles as they’ve never failed to deliver. A charge lasts at least 24 hours.
Rick Lang
Offline
User avatar

Bromine 18

  • Posts: 145
  • Joined: Tue Feb 09, 2021 4:00 pm
  • Real Name: Aldous Barnes

Re: bmpcc 4k - Jitter or not?

PostSat Jan 29, 2022 7:34 pm

rick.lang wrote:BMD UM4.6K and BMPCC4K cameras aren’t quite as reliable to maintain sync for several hours.


Without the Tentacle, does your Pocket 4K drift even if you ensure it isn’t turned off?

It’d be useful to know the threshold beyond which the drift gets significantly perceptible, with a single jam sync and no powering down, neither the recorder nor the camera.

8 or 6 hours I can live with; just 2 hours would be concerning.
Online
User avatar

rick.lang

  • Posts: 17175
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: bmpcc 4k - Jitter or not?

PostSun Jan 30, 2022 4:53 am

You would not want to rely on the camera keeping sync after being jammed (in my opinion based on my older cameras) for more than a couple of hours. Honestly most shoots can find a spare minute to re-jam a couple of cameras and a recorder so that’s not a serious failing. And it might not be a problem if you had to wait a couple of hours longer. The tentacles can maintain sync for more than a day so that probably is the safest solution if you can’t re-jam for some reason. The MixPre-6 II can last a day’s shoot so that’s safe to jam and then leave the Tentacle on your camera.
Rick Lang
Offline
User avatar

Uli Plank

  • Posts: 21280
  • Joined: Fri Feb 08, 2013 2:48 am
  • Location: Germany and Indonesia

Re: bmpcc 4k - Jitter or not?

PostSun Jan 30, 2022 8:14 am

The P6K Pro is very precise, but that doesn’t apply to most older models.
No, an iGPU is not enough, and you can't use HEVC 10 bit 4:2:2 in the free version.

Studio 18.6.5, MacOS 13.6.5
MacBook M1 Pro, 16 GPU cores, 32 GB RAM and iPhone 15 Pro
Speed Editor, UltraStudio Monitor 3G, iMac 2017, eGPU
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

Re: bmpcc 4k - Jitter or not?

PostFri Feb 04, 2022 7:04 pm

So, I ordered a set of Ambient Nanolockits, because they have an internal rtc wich comes in handy and the wireless part can be disabled with an app if needed. I will do some testing when they arrive and if you are interested I can post some results.

Thank you for your help,
Stephan
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

Re: bmpcc 4k - Jitter or not?

PostTue Feb 08, 2022 6:55 am

Due to the actual situation, I had some sparetime for testing:

I recorded short clips of me using a slate over quite some time. Sound got recorded internally as well as to the zoom f8. As I can not tell for sure which one of these devices, if not both do drift, in the following graphs only the measured difference in ms (offset) between them is displayed. Both test were done twice with the same results (within common error range). Framerates were both set to 25, samplerate set to 48000Hz.

1. I took the zoom f8 as a master connected to the bmpcc4k via cable. No adapters needed, cable working fine, external sync indicated. See the repetitive pattern:
timcodetest-1.jpg
Timecode test 1
timcodetest-1.jpg (72.93 KiB) Viewed 1758 times


The maximum variance is 18ms, max difference about 25ms (~ 1/2 of a frame), visually both stay in sync with the picture. Visually I couldn't tell which one would be more accurate, slight echoes were audible.


2. I just jam synced the bmpcc4k from the zoom f8, than removed the cable. Beforehand both internal clocks (rtc) were set to the clock of my computer by cable, too:

timecode 2.jpg
timecode 2
timecode 2.jpg (32.96 KiB) Viewed 1758 times


Although they were instantly off by two frames, the maximum variance is only 6ms within the given time! I know, it was not that long of a period. Audio and video was visually off, of course.

So, what to do with these informations? Both stay in sync reasonably, although one with a more or less constant offset of two frames within the given time. But still, it is not the result I had expected. I expected a drift more smoothly over time, not with any jumps like above.

- Is the timecode clock from the f8 not stable enough?
- Is it a problem of the camera getting warmer over time?
- Is it a problem of missing genlock and two clocks (shutter and audio) running at different speed/pulses?

I still have to wait for the Nanolockits to arrive to compare, but will do so.

I'm curious about your opinions.
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: bmpcc 4k - Jitter or not?

PostTue Feb 08, 2022 12:19 pm

Few things to come around on.

Did you cut the camera and restart it over time ?

The timecode of ANY clip is grabbed at the start of recording, be it from an external source or internal clock. While recording it’s just a frame count.

Yes timecode does change with temperature. Some of the more accurate clocks are temperature regulated.

It’s also normal to have an offset for any given audio / TC procsss. As long as the offset doesn’t change over time.

If you want to really test for any TC drift you would do it over a lon period of time. Like 6-8 hours. You can record some sync sound at the beginning and then leave the clocks “powered” and check again in 6 or 8 hours.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline
User avatar

Bromine 18

  • Posts: 145
  • Joined: Tue Feb 09, 2021 4:00 pm
  • Real Name: Aldous Barnes

Re: bmpcc 4k - Jitter or not?

PostTue Feb 08, 2022 3:03 pm

kontrakatze wrote:I'm curious about your opinions.


Really appreciate the analyses.

My suggestion is also to retest, especially the second criterion.

8 hours of duration would be ideal; just one jam sync to examine the extremities in the drift, and ensure both the recorder and the camera are continuously powered – no shut down.

Maybe also do 3 replicates, one per day, and log the ambient temperature to confirm that the instantaneous 2-frame drift after a jam sync is indeed reproducible.

I think I’ll also do my own testing with my Zoom F6 and Pocket 4K.

Could you share your spreadsheets?
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

Re: bmpcc 4k - Jitter or not?

PostTue Feb 08, 2022 5:56 pm

@John:
No, the camera was powered on continuously by v-mount battery.

@Bromine: This is what I plan to do, maybe at the weekend. One longtime (8h) test with just a jam sync. I'm not sure if I can do it 3 times as this would take 3 days. I'm very interested in seeing the results with your Zoom f6. I think the f6 and f8 would probably have the same timecode clock.

The spreadsheet is a simple numbers document which can be converted to excel, but there is no way to upload it here.
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos
Offline

jhoepffner

  • Posts: 181
  • Joined: Wed Sep 05, 2018 9:43 am
  • Location: Paris
  • Real Name: Jacques Hoepffner

Re: bmpcc 4k - Jitter or not?

PostTue Feb 08, 2022 6:21 pm

I doesnt really understand your problem, there is a timecode generator included in Zoom F8. My technic is to output from the zoom a stereo output, left with TC output, right with sound premix. I send it to the main camera operated by myself (GH3 with 30' limit) with a simple mini-jack stereo cable and I use a Rode Wireless to go to send the same stereo signal to each other cameras (2 GH4, 1 BMPC 4K, 1 iPhone with Filmic Pro all without operators), one for each cameras and all is working perfectly.
Comming back to DVR:
– changing TC with audio for all clips
– suppress TC with clip attribute
– tagging angle for each camera and sound track
– make multicam clip based on TC
– outputing a continuity to transcipt text in Premiere and obtain subtitles
– prepare the edit with the .srt files
All is working perfectly. I edit at the moment two different projects, each 30 hours of rushes with the 5 cameras and all is working flawlessly, no need to other device.
W10 Ryzen7 3800X nVidia 3090 RAM 64 DVR-Fusion Studio 18.6.5 SSD 2To raid0 32To (HD) raid0 8To (SSD)
MSI G65 W10 / RAM 32 nVidia 1070 Intel Core i7-8750H DVR Studio 18.6.5
MBP 15" (2012) OS 10.14.6 RAM 16 nVidia 650M Intell i7 DVR Studio 17.4
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

Re: bmpcc 4k - Jitter or not?

PostTue Feb 08, 2022 7:34 pm

I'm glad to hear, that it works for you.

This might be working for talking heads when even a drift of one frame does not matter to much, but not for music. The more complex the music setup gets (orchestra and conductor), even slight offsets (~4-6ms) lead to comb filter effects no one wants to listen to. So if you take some audio internally, even if you don't move the camera and you are kind of toast. They do not sync up close enough. Just two choices left: move the tracks manually in post or delete the internal audio. And again: even with timecode from the zoom f8. As you can see above, the offset is not constant. Not nearly.

Same is true for story telling if you, lets say, record talking persons to the f8 and ambient to the camera (maybe because of a lack of tracks in the field). As long as you get some bleeding from the voices with the ambient tracks and they are not in sync to the ms, you get comb filtering. Only solution: keep the ambient mics far away from the voices, or move manually.

This is not to much of a bummer if the offset between the devices is kind of constant. You know what you have to deal with, take notes about the distances and you can work quite fast. It get's a pain if the offset changes with every take you do.

By the way, to keep things workable even wireless talking needs compensation within the f8, at least if the talking person is close to one of the other mics used. So, measure the offset you get from wireless, set it within the f8 and all is clean and nice in the soundbag. But take different shots at different times and you get different offsets to the camera if you use the timecode of the zoom f8.

I wouldn't send my soundbags outs to the camera wirelessly for many reasons.
But hey, there are a lot of ways to Rome.
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos
Offline

jhoepffner

  • Posts: 181
  • Joined: Wed Sep 05, 2018 9:43 am
  • Location: Paris
  • Real Name: Jacques Hoepffner

Re: bmpcc 4k - Jitter or not?

PostWed Feb 09, 2022 7:55 am

Hello, I do mainly dance with live music but I never use the sound of my cameras, only from the zoom or a computer with sound card if I need more tracks.
W10 Ryzen7 3800X nVidia 3090 RAM 64 DVR-Fusion Studio 18.6.5 SSD 2To raid0 32To (HD) raid0 8To (SSD)
MSI G65 W10 / RAM 32 nVidia 1070 Intel Core i7-8750H DVR Studio 18.6.5
MBP 15" (2012) OS 10.14.6 RAM 16 nVidia 650M Intell i7 DVR Studio 17.4
Offline

John Brawley

  • Posts: 4267
  • Joined: Tue Aug 21, 2012 7:57 am
  • Location: Los Angeles California

Re: bmpcc 4k - Jitter or not?

PostWed Feb 09, 2022 4:56 pm

kontrakatze wrote:@John:
No, the camera was powered on continuously by v-mount battery.


But did you CUT. Start recording, stop recording start recording etc.

JB
John Brawley ACS
Cinematographer
Currently - Los Angeles
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

Re: bmpcc 4k - Jitter or not?

PostWed Feb 09, 2022 5:07 pm

Ok, didn't understand your question correctly. Yes, of course. I took multiple shots. Every shot is about 30s in length starting a new clip about every 30min. Camera stayed on at all times, zoom f8, too.
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos
Offline

kontrakatze

  • Posts: 87
  • Joined: Fri Dec 04, 2020 3:55 pm
  • Location: Berlin
  • Real Name: Stephan Schoenharting

Re: bmpcc 4k - Jitter or not?

PostSun Feb 13, 2022 12:03 am

So, the Nanolockits arrived today and are quite easy to setup. And, yes, I did one initial test: exactly the same test I did with the zoom f8 before, but not as long in time.

The results are exactly the same, so there is absolutely no difference between the direct cable connection from the zoom to the bmpcc4k and using the Nanolickits with both devices (I mean, within reasonable margins):
Nanolickit.jpg
Nanolockit
Nanolickit.jpg (43.58 KiB) Viewed 1238 times


This seems to proof, that the changing offset measured is indeed related to the bmpcc4k. As the ambient temperatures were almost constant during all tests and the pattern is repetitive, I come to the conclusion, that the internal oscillator of the bmpcc4k is not stable enough and would need genlock (which it does not have) for real sync.

Is it accurate enough for syncing frames across different cameras? Yes, the offset was well within one frame
Is it accurate enough to use the internal audio in conjunction with a soundbag (here the zoom f8)? Definitely not without any manual corrections and there is no shortcut.

What does this mean for me as a one man band?
1. I'm going to use either or. Either the inbuilt audio recorder, or my soundbag (preferably the soundbag) and not both.
2. In case I need more devices to be in sync, I can rely on the zoom f8 for timecode quite nicely. I would take the zoom as a timecode master and lock the two Nanolockits as slaves. This would give me three devices with at least a starting point for audio sync and with visuals being still in sync.
3. If I'm in need of more audio channels, I would probably skip the backup recordings of the zoom (always a bit risky) for more channels available. Right now I use four channels for recording and four as a backup with lower levels.

And to answer the question of this topic "jitter or not?", yes, the bmpcc4k does jitter unpredictably within a little more than half of a frame.

Long hours of testing done - quite some things learned and a plan for the future: No more cameras without genlock.

Keep recording,
Stephan
Stephan Schoenharting
Berlin, Germany
https://www.youtube.com/channel/UC9HVF-6TcKDkavfcqu0lFMQ/videos

Return to Cinematography

Who is online

Users browsing this forum: Adam Langdon, devinpickering, evalsenola, Michel Rabe and 105 guests